The following subsections describe each of the advanced VDR technologies (or "options") in detail, and provide an overview of the technology, implementation issues, infrastructure requirements, and costs. The optional or advanced technologies reviewed include:

• Section 3.1: Supplemental internal memory storage (e.g., EEPROM, flash memory)
• Section 3.2: Removable storage media (e.g., magnetic, optical, solid-state memory)
• Section 3.3: Onboard vehicle network communication and downloading (e.g., CAN, IDB, serial)
• Section 3.4: Vehicle location, direction of travel, and absolute time e.g., GPS)
• Section 3.5: Digital imaging (e.g., video)
• Section 3.6: Sensors for determining the relative location of nearby vehicles (e.g. radar, ultrasonic)
• Section 3.7: Short-range wireless communications (e.g., infrared, Bluetooth, WiFi 802.11)
• Section 3.8: Long-range wireless communications (e.g., satellite, cellular)
• Section 3.8: Driver performance (e.g., attentive driver monitoring, drowsy driver warning)
• Section 3.10: Tractor-to-trailer communications

The following questions are addressed for each technology or feature (where applicable):

• Technology Overview
• How does the system function?
• What components are required?
• Are there different levels of implementation for this technology (i.e., is there an advanced or full-featured version? Is there a basic version?)?
• Where is this technology currently used?
• What information will it provide to the VDR? How could this information be used for accident reconstruction, operational management, driver training, or emergency personnel?
• What are the current and near-term commercially available systems?
• Are any long-term development projects underway that might impact the integration of the technology with a VDR?
• VDR Implementation Issues
• How might this technology be implemented into an EDR on a heavy-duty vehicle?
• What additional hardware will be required on the vehicle or VDR to use this technology (e.g., additional sensors, receiving and/or transmitting antennas, or databuses required)?
• What are some of the disadvantages of using this technology over other similar technologies (e.g., removable media versus short-range wireless versus long-range wireless)?
• Infrastructure Requirements (if applicable)
• What kind of infrastructure is necessary for this technology to operate (e.g., cellular service, satellites, road markings, WiFi "hot spots")?
• Is this infrastructure currently available? What is the timeframe for development?
• Who is likely to develop this infrastructure?
• Technology Implementation Cost
• What is the range of costs, per vehicle, for the hardware required to implement the technology?
• How might these costs decrease as the technology becomes more popular?

Supplemental Internal Memory Storage

Technology Overview

One of the key components of any onboard data recording device is memory. The size and type of memory affects how much data can be stored, how quickly the data is saved, the durability of the recorder, and how long the data will remain intact. Typically, there are two main types of memory-volatile and non-volatile. Volatile memory requires a constant source of power for the memory contents to remain intact and is typically used for temporary, high-speed, data storage. Examples of volatile memory in computer electronics include random access memory (RAM) and cache (high-speed memory located in or near the CPU). In contrast, the contents of non-volatile memory remain intact when power is disconnected. Non-volatile memory is typically used for long-term storage of data and generally has slower read and write speeds. Examples of non-volatile memory include read-only memory (ROM); programmable read-only memory (PROM), which has many variants (e.g., erasable PROM or EPROM, electronically erasable PROM or EEPROM, and flash memory); and even magnetic storage disks (i.e., hard disk drives).

Most electronic devices, including VDRs, use both volatile and non-volatile memory. Program software is typically stored in non-volatile memory (usually ROM). When the device is turned on, that program is loaded into volatile memory (RAM and cache) and the CPU. Data can then be retrieved from another non-volatile memory source, typically one that allows for both read and write provisions (EEPROM, flash memory, or hard disk drive).

In a VDR, the size and speed of the volatile memory often determines how much data is stored in a continuous buffer before an event triggers the data to be saved permanently. Although this function could be accomplished using non-volatile memory, it is often significantly slower (especially when the non-volatile memory is an EPROM or EEPROM). While this is very important to the design of the VDR electronics, volatile memory is often not a limiting factor as the size required is typically small.

The non-volatile memory used for permanent data storage significantly impacts VDR design as it is a limiting factor for how much event or operational data a VDR records. Non-volatile memory is available in a range of capacity-from a small capacity to store only a few seconds of critical data to a large capacity able to store multiple data element streams for days or weeks.

The most basic non-volatile memory is PROM, a non-volatile integrated circuit memory chip on which data can be written only once. Once a program has been written onto a PROM, it remains there forever. PROMs retain their contents when the computer is turned off. To write data onto a PROM chip, you need a special device called a PROM programmer or PROM burner.

An EPROM is a special type of PROM that can be erased by exposing it to high-intensity ultraviolet (UV) light. Once it is erased, it can be reprogrammed. To accomplish this, EPROM chips normally contain UV-permeable quartz windows exposing their internals. The UV light clears the contents, making it possible to reprogram the memory. To write to an EPROM, you can use the same device as a PROM. Generally, EPROMs are used to store program code and static data, and are widely used in personal computers because they enable the manufacturer to change the contents of the PROM before the computer is actually shipped.

Largely, PROMs and EPROMs are used in computer systems to store program code generated by the manufacturer of an embedded computer system. Because these devices require additional hardware to record and erase data, they generally are not used to store data. A common alternative to the PROM and EPROM typically used in embedded computer system applications to store data is EEPROM, a special type of PROM that can be erased and written electrically. EEPROM maintains its contents without power backup and is frequently used to retain program data. EEPROMs have slow read and write times because they require data to be written or erased one byte at a time. EEPROM chips generally range in capacity from 128 bits to 512 KB, with write times in the 5 to 15 millisecond range.

Flash memory is a type of EEPROM that can be erased and reprogrammed in blocks of data instead of one byte at a time. Flash memory is also popular because it enables hardware manufacturers to support new protocols and technologies as they become standardized by simply installing updated program code. Flash memory is typically available in three formats, disk-on-chip, flash disk drive, and removable CompactFlash (see Section 3.2 - Removable Storage Media). With disk-on-chip, flash memory is contained in a single chip that can be integrated into a printed circuit board of an embedded computer directly. Disk-on-chip flash memory chips typically range from 1 MB to 32 MB, but can reach up to 300 MB. Flash drives are flash memory chips that are integrated in a package similar to a hard disk drive. Flash drives generally operate using similar computer protocols to that of hard disk drives, but are noiseless, rugged, and light weight. Flash drives are generally available to replace hard disk drives up to 1 GB in size.

Hard disk storage is another non-volatile memory alternative and is commonly used in personal computers. Hard disks are typically used when there is a large memory requirement, and can be sized up to 300 to 400 GB. While hard disks are robust enough for portable computers, they are susceptible to temperature, moisture, and vibration, which can reduce their long-term durability and reliability.

VDR Implementation Issues

EEPROMs and flash memory are uniquely suited to store data in a VDR-not only because they are non-volatile, but also because they are robust and can be integrated into the VDR or could be removable in the event of an accident (see Section 3.2 - Removable Storage Media).

Hard disks are susceptible to temperature, moisture, and vibration, which can reduce their long-term durability and reliability. Additionally, the high forces of an accident event could cause damage to a hard disk drive, thus preventing data from being recorded. Therefore, it is likely that hard disk drives would not be used for VDRs designed for accident event recording. VDRs designed for operational data recording would avoid the use of hard disks where possible.

Memory type and size selection for a VDR is important and depends largely on the intended application. For instance, a VDR might require the same memory capacity to record 5 high-resolution, high-frequency inputs for 2 seconds as a VDR designed to record 50 low-resolution, low-frequency inputs for 2 weeks. Exhibit 3.1 provides an order-of-magnitude estimate for the amount of memory required for various applications.

Exhibit 3.1 - Example Memory Requirements

Exhibit 3.1 illustrates how much raw memory space would likely be required to record different amounts and types of data. For example, it shows that recording three high-resolution accelerometer inputs during a 30-second event at 100 Hz (100 times/sec) would require approximately 70 KB of raw (excluding any necessary memory overhead) memory space. This could easily be accomplished using a moderately sized EEPROM or a small flash memory chip. Similarly, to record GPS location and time for one hour at once per second would also require the same 70 KB of raw memory.

The third and fourth examples in Exhibit 3.1 illustrate the typical memory requirements to record accident event data and operational data respectively. Example 3 shows what a typical accident record might resemble-five high-resolution, high-frequency inputs (accelerometers, yaw, tilt) for 5 seconds; five low-resolution, low-frequency inputs (e.g., vehicle speed, engine speed, brake pressure) for 20 seconds; and five digital (on/off) inputs for 20 sec. Memory requirements for Example 3 would likely be modest.

Example 4 illustrates the memory requirements for a day's worth of operational data and GPS location and time. Typical operational data could include 10 low-resolution inputs (e.g., vehicle speed, engine RPM, engine temperature) recorded every minute, 10 digital inputs (e.g., gear, headlight activation) recorded every minute, and GPS location and time recorded every minute. Memory requirements for Example 4 (approximately 98 KB) would be more than for the accident data in Example 3.

Technology Implementation Costs

EEPROM chips vary in cost, and depend significantly on size and quantity purchased. Exhibit 3.2 shows typical costs for various size EEPROMs with a moderate purchase quantity.

Exhibit 3.2 - Typical EEPROM Chip Costs

Both disk-on-chip and flash disk drive memory also vary in size and cost. Exhibit 3.3 shows typical costs for disk-on-chip and disk drive flash memory.

Exhibit 3.3 - Typical Flash Memory Costs

Exhibits 3.2 and 3.3 show typical costs for EEPROM and flash memory chips, based on moderate purchase quantities. Of course, these prices are intended to illustrate the relative cost differences between memory types and sizes. It should be recognized that there are many other design factors that affect the price of memory (e.g., access time, memory organization, packaging, power supply and usage, and temperature range).

Removable Storage Media

Technology Overview

Removable storage media for a VDR could provide a relatively straightforward approach for retrieving data. There are many forms of removable storage available, but most would not be suitable for the vehicle environment. The most common types of removable storage media include:

• Removable Magnetic Storage
• 3.5" floppy disks
• Tapes
• Zip disks
• Super disks
• Removable hard disks
• Removable Optical Storage
• CD-R (write once)
• CD-RW (re-writable)
• DVD-R/DVD+R (write once)
• DVD-RW (re-writable)
• Removable Solid State Memory
• CompactFlash
• SmartMedia
• Sony Memory Stick
• SD Memory
• XD Memory cards

Due to the automotive environment, removable magnetic storage (e.g., 3.5" floppy, Zip, and Super disks) is not practical because of its susceptibility to moisture, temperature, and vibration affecting both the long-term durability and the sensitive write heads. Optical storage (i.e., CDs and DVDs) have similar vibration issues that affect the recording ability and accuracy of the laser heads. In addition, these technologies have multiple mechanical components that are generally not very durable (limited application in vehicles with 100,000+ mile lifespan). Tape drives are very slow, have similar durability and environmental concerns, and are generally expensive. Removable hard disk storage, while more reliable than diskettes and CDs and DVDs, would likely still not be suited for automotive applications where vibration and moisture could cause data loss and premature failure of the disk drive. Hard disk drives have the advantage of very high-speed access and large capacity, and are a very mature technology in the mobile computer industry, but, for the reasons described above, have not been widely used in the automotive industry.

The most promising technology is solid-state memory. The trend toward mobile computing (e.g., handheld PC, cellular phones) and digital cameras has led to the mainstreaming and continued evolution of solid-state, non-volatile memory. Solid-state memory is a type of EEPROM on which data can be electronically written and electronically rewritten without any mechanical components. In addition, solid-state memory (unlike a computer's RAM) does not require a continuous source of power to maintain data. The advantages of solid-state memory are numerous:

• Faster access speeds than hard disks
• Compact and lightweight
• Noiseless
• No mechanical parts
• Low susceptibility to environmental conditions and vibration with the appropriate packaging

The major disadvantage is cost. While the price continues to decrease as demand increases, solid-state memory still costs more than hard disk drives, CDs and DVDs, and diskettes.

There are a multitude of removable solid-state memory technologies and brands available, including:

• CompactFlash
• SmartMedia
• SD Memory
• XD Memory
• Memory Stick

Originally developed by SanDisk in 1994, CompactFlash is the oldest technology, and is common in embedded system applications. It is the largest (35.4x42.8x3.3/5.5 mm) and heaviest (11.4 g) of the technologies, but also the most affordable. CompactFlash cards can range from 4 MB to 2 GB in size and have a write time of 300 to 2,000 KB/sec.

SmartMedia, originally developed by Toshiba and called the solid-state floppy-disk card, ranges in capacity from 4 to 128 MB. SmartMedia has a very simple architecture and is one of the smallest (37x45x0.76 mm) and lightest (2.0 g) of the technologies. It has a capacity of up to 128 MB and a write time of 500 to 1,000 KB/sec. In general, however, for personal applications, it has been found to be less rugged than the other solid-state memory devices (i.e., it is generally very fragile) due to the packaging and electrical contacts.

SD Memory, developed as a "next-generation" memory card by both SanDisk and Toshiba along with Panasonic, is very small in size (24x32x2.1 mm) and lightweight (2.0 g). It supports up to 512 MB with a write speed of up to 10 MB/sec. Unlike both CompactFlash and SmartMedia, SD Memory has both copy protection and write protection built in; enabling data to be encrypted and protected from overwrites.

XD Memory is a similar "next-generation" technology with future capacities of up to 8 GB on a single tiny (only 20x24.9x1.8 mm) XD card. It is currently commercially available up to 256 GB in size.

Memory Stick is a proprietary Sony memory format. It is larger (50x21.5x2.8 mm) and heavier (4.0 g) than SD and XD memory, and is largely only found in Sony products. It is beginning to make headway in other manufacturer's equipment, but because it is proprietary, it is still slightly more expensive than newer and smaller technologies. It has a maximum capacity of 128 MB and a write time of 2.45 MB/sec. Its proprietary design means that, until recently, it was not often used in embedded applications.

VDR Implementation Issues

Integrating either magnetic or optical storage technologies into a VDR presents some significant challenges as these technologies would have to be proven to be reliable to withstand both the everyday environment and vibration of heavy-duty commercial vehicles and the impact force of an accident. The VDR's reserve backup power capabilities would have to be sufficient to complete all data acquisition and recording on the magnetic or optical storage after an event. It is not likely that either magnetic or optical storage systems would be suitable for an extended VDR application.

Again, solid-state memory appears to have a significant advantage in VDR integration as solid-state memory devices continue to be integrated into portable electronics and personnel computers. Solid-state memory devices are easily integrated into embedded systems because their interface is relatively straightforward. Technologies such as SmartMedia, SD, and XD use simplistic serial I/O to read and write data that could be integrated into the VDR's microcontroller. CompactFlash uses the common ATA/IDE standards for communication with disk drives.

Infrastructure Requirements

Little to no infrastructure requirements are required to support removable storage media. Specialized diagnostic tools with the appropriate memory slots or custom programming of existing equipment will be necessary to read data from removable solid-state memory cards, but this technology is readily available and should be straightforward to implement. Personal computers, laptops, and PDAs could readily be programmed to read and analyze data from a VDR's removable storage media. Adapter cards are currently available to read solid-state memory in a laptop's PCMCIA slot (used mainly for network cards), and many new computers are have dedicated memory card slots built-in.

Technology Implementation Costs

Integrating magnetic or optical storage into a VDR would likely be prohibitive as these technologies would require a significant amount of research and development in order to ensure they would withstand the environmental conditions and the impact of an accident. However, both magnetic (specifically removable hard disk drives) and optical memory are very inexpensive, typically less than $0.02 per MB. Solid-state removable memory would be significantly less costly to implement. Typical solid-state memory costs per MB are approximately:

• CompactFlash: $0.70 -$0.86 per MB
• SmartMedia: $0.74 -$1.14 per MB
• SD Memory: $1.94 per MB
• XD Memory: $0.95 per MB
• Sony Memory Stick: $1.49 - $2.65 per MB

Higher per MB prices can be expected for recently developed cards with larger MB capacity.

Onboard Vehicle Network Communication and Downloading

Technology Overview

Hardwired onboard communication (i.e., communication that uses physical electrical wiring as the medium) is widespread in the automotive industry. Applications are continuing to increase as manufacturers look for ways to reduce costs and simultaneously add features to vehicles. Both the light- and heavy-duty markets have seen the introduction of onboard vehicle networks to allow communications between different electronic control units on the vehicle. On the light-duty side, this was largely precipitated by the onboard diagnostics (OBD) regulations requiring emissions control system monitoring. Since light-duty vehicle subsystems historically have largely been manufactured and integrated by the OEM or in close cooperation with its suppliers, a multitude of communications protocols have evolved, with the OBD regulations supporting many of them. In the heavy-duty industry, where common subsystems are provided by various suppliers (e.g., engines, transmissions, and brake systems) and integrated by vehicle OEMs, a need for common communications standards has facilitated only a few prominent network protocols.

SAE, in particular, has defined three distinct protocol classifications for both light- and heavy-duty applications: Class A, Class B, and Class C. Class A is the first SAE classification and maintains the lowest data rate, a rate that peaks as high as 10 Kb/s. Class A devices typically support convenience operations like actuators and "smart" sensors. The implementation of Class A has significantly reduced the bulk of automotive wiring harnesses. The second SAE classification is the Class B protocol. Class B supports data rates as high as 100 Kb/s and typically supports intermodule, non-real-time control and communications, including current OBD protocols. The use of Class B can eliminate redundant sensors and other system elements by providing a means to transfer data (e.g., parametric data values) between nodes. Class C is the last of these three classifications, and readily supports performance as high as 1 Mb/s. Because of this level of performance, Class C is typically used for critical, real-time control. Class C facilitates distributed control via high-data-rate signals typically associated with real-time control systems.

Light-duty protocols are largely focused around satisfying OBD regulations, but have been used more frequently for intermodule communications in recent years. Light-duty protocols include:

• SAE J1850 Class B Data Communications Network Interface
• ISO 9141 Road Vehicles: Diagnostic Systems Requirements for Interchange of Digital Information
• ISO 14230 Road Vehicles: Diagnostic Systems Keyword Protocol 2000
• SAE J2284 High-Speed CAN (HSC) for Vehicle Applications at 125, 250, and 500 Kbps

In addition, three other communications standards are gaining popularity for light-duty, non-OBD applications; Local Interconnect Network ( LIN); Media Oriented Systems Transport ( MOST); and Automotive Multimedia Interface Collaborations ITS Data Bus ( IDB). LIN is designed as a low-speed network for use on doors, roofs, seats, dashboards, and steering wheels to connect switches and relays to actuators-reducing wiring costs. It is a single-wire network. MOST is a plastic fiber-optic digital media network for sound systems and video entertainment systems. It has largely been developed by European automakers and is in production on high-end European vehicles. IDB was designed to allow aftermarket suppliers to connect equipment using a standardized connector and have access to vehicle information and systems without disturbing the vehicle control systems or OBD. To date, IBD has not largely been adopted due to low-transmission rate and lack of industry interest in a standard media protocol.

In the heavy-duty commercial vehicle market, a single OEM might offer vehicles with multiple types and manufacturers of engines, transmissions, ABS, and other vehicle subsystems. This necessitated a common method for communications between subsystems so that any engine could theoretically communicate with any transmission or ABS. In addition, there was a need for a common diagnostic interface to minimize cost and facilitate troubleshooting of vehicles with multiple subsystems. In North America, two main protocols have become standard, SAE J1708/J1587 and SAE J1939 for Class B diagnostics and a high-speed Controller Area Network ( CAN) for Class A communications. In addition, in Europe, the ISO 11898 CAN standard is the most popular along with ISO 11992 for tractor-to-trailer communications.

SAE J1708, Serial Data Communications Between Microcomputer Systems in Heavy-Duty Vehicle Applications, defines a recommended practice for implementing a bi-directional, Class B serial communication link among vehicle ECUs. SAE J1708 defines the hardware and basic software compatibility such as the interface requirements, system protocol, and message format of the serial communications link. SAE J1708 does not define actual data to be transmitted by particular ECUs, which is an important aspect of communications compatibility, and is left to other standards and individual manufacturers to define. SAE J1587, Joint SAE/TMC Electronic Data Interchange Between Microcomputer Systems in Heavy-Duty Vehicle Applications, builds on J1708 to define the diagnostic serial communications link between onboard ECUs and off-board diagnostic equipment. It defines the format of messages and data to be transmitted including field descriptions, size, scale, internal data representation, and position within a message.

SAE J1939, Recommended Practice for a Serial Control and Communications Vehicle Network, is intended for light-, medium-, and heavy-duty vehicles used on or off road as well as appropriate stationary applications that use vehicle-derived components (e.g., generator sets). Vehicles of interest include, but are not limited to, on and off highway trucks and their trailers; construction equipment; and agricultural equipment and implements. The purpose of J1939 is to provide an open interconnect system for electronic systems. J1939 is designed to allow ECUs to communicate with each other by providing a standard architecture.

ISO 11898, Road Vehicles: Interchange of Digital Information Controller Area Network (CAN) for High-Speed Communication, describes the general architecture of CAN in terms of hierarchical layers according to the ISO reference model for Open System Interconnection (OSI), which is a standard for worldwide communications that defines a networking framework for implementing protocols in seven layers. The standard contains detailed specifications of aspects of CAN belonging to the physical layer and the data link layer. It specifies characteristics of setting up an interchange of digital information between electron control units of road vehicles equipped with CAN at transmission rates above 125 Kbit/s up to 1 Mbit/s. The CAN is a serial communication protocol that supports distributed real-time control and multiplexing.

ISO 11992, Road Vehicles: Interchange of Digital Information on Electrical Connections Between Towing and Towed Vehicles, describes the protocol used for the interchange of digital information between road vehicles with a maximum authorized total mass greater than 3,500 kg, and towed vehicles, including communication between towed vehicles in terms of parameters and requirements of the physical and data link layer of the electrical connection used to connect the electrical and electronic systems. It also includes conformance tests of the physical layer.

VDR Implementation Issues

Wired onboard vehicle networks can be used to both collect data from vehicle subsystems and download VDR data to an off-board reader. Current heavy-duty commercial vehicles have both J1708/J1587 and J1939 networks onboard, and a significant amount of data is currently being broadcast across these networks, including:

• Vehicle, engine, and wheel speeds
• Engine and emissions control information
• Transmission status and gear
• Brake system and ABS information
• Safety system status (e.g., seatbelt, airbag, and door lock)
• Subsystem fault codes

Connecting a VDR to these networks would be straightforward and would not only allow data collection but also downloading the diagnostic ports. The technology is commonplace in the heavy-duty vehicle industry and could be readily implemented.

Light-duty and ISO standard protocols would require slightly more development to implement in heavy-duty North American vehicle applications, but the technology is readily available and could be easily integrated. Most heavy-duty vehicle systems in North America do not currently use light-duty and ISO networks onboard, so data would likely not be available to be recorded over these networks in a VDR. While it is technically possible, it is unlikely that any of the light-duty networks will be used in the heavy-duty market in the future.

Infrastructure Requirements

Little to no infrastructure requirements are required to support vehicle-based wired downloading and data collection. Specialized diagnostic tools or custom programming of existing equipment will be necessary to download and read the VDR data, but this technology is readily available and should be straightforward to modify.

Technology Implementation Costs

The cost to connect a VDR to an onboard vehicle network and provide downloading of VDR data over the vehicle's diagnostic connectors would be minimal. For CAN network standards, microcontroller chipsets (CAN protocol controllers) can cost around $5.00 with the additional transceiver interface (to interface between the microcontroller and the physical bus) costing around $1.50. J1850 transceivers are priced similarly around $2.00. LIN microcontrollers can cost approximately $2.50 with transceivers costing less than $1.00.

Vehicle Location/Time/Direction of Travel

Technology Overview

The Global Positioning System (GPS) has become the most popular and efficient means for determining location and direction of travel in mobile applications. The relative compactness, low cost, and commercial availability of the technology are key factors in its attractiveness as an accurate measure of location and direction of travel. In addition to providing location information, the design of the technology allows for the accurate measurement of time.

Originally developed by the U.S. military, GPS is a worldwide radio-navigation system formed from a constellation of 24 satellites and multiple land-based stations. GPS uses these satellites as reference points to calculate positions accurate up to a couple of meters. Advanced forms of GPS, such as Differential Global Positioning System (DGPS), can have accuracies close to a centimeter. Modern GPS hardware has been miniaturized into a few integrated circuits, yielding an economical and compact system for determining precise location, direction of travel, and time data for a variety of automotive, marine, aviation, construction, and computer applications. In addition, GPS hardware can be combined with Geographic Information Systems (GIS) data to

overlay vehicle locations on maps to reference them to geographic features, streets, and buildings.

GPS uses distances from multiple satellites to triangulate a location, using the travel time of radio signals. GPS satellites transmit RF signals, and the signal's travel time is measured by the receiver. GPS receivers have an almanac programmed into them, which has the position of all the satellites. The U.S. military continuously monitors each satellite's position, and corrections are transmitted to the receivers within each timing signal. Using the timing signal from four satellites and knowing the location of each satellite, the receiver can calculate its location in 3-dimensional space.

DGPS supplements these satellite signals with a signal from a land-based, fixed location receiver. This fixed receiver knows what the actual signal travel time should be and compares that with what it receives. It then transmits this difference as a correction factor to nearby DGPS receivers over radio frequencies. The correction factors could also be post-processed if precise positioning is not needed immediately.

This GPS technology is not only used for tracking and mapping but also to distribute precise time and time intervals. GPS satellites carry highly accurate atomic clocks. In turn, ground receivers synchronize to these clocks in order to precisely calculate distances. This leads to GPS receivers having accuracy close to that of an atomic clock.

GPS technology can be implemented at various levels, from standalone chipsets to complete GPS and GIS systems and DGPS. At the lowest level, GPS chipsets, similar to those developed

for the FCC's E911 regulations governing cellular phones, are available that receive the GPS satellite timing signals but leave the distance calculations and triangulation to post-processing, often referred to as Assisted GPS (A-GPS). The chipsets are compact and require relatively little power for operation, but leave all of the calculations and analysis to be done externally from the chips either off-board or using separate processors. GPS receivers, similar to those in handheld applications, can receive timing signals; process distances using an almanac of known satellite locations; and provide calculated latitude, longitude, and elevation information along with speed, direction of travel, and time. GPS receivers can also be integrated with GIS to place this latitude, longitude, and elevation data on maps of terrain, roads, buildings, and other landmarks-similar to those systems for automotive and marine navigation systems.

DGPS systems are also available with built-in radio frequency receivers to pick up the correction factors from nearby land-based DGPS broadcast centers. The U.S. Coast Guard, military and government organizations, and many commercial companies have DGPS broadcast centers located throughout the U.S. and globally. DGPS is not currently available in all areas, but planned DGPS sites will cover most of the continental United States in the future. (DGPS receivers can operate on standard GPS if the correction factors broadcast is not available.)

Many commercial manufactures offer handheld GPS receivers for mobile and stationary applications. Leading manufacturers include Trimble, Garmin, and Magellan. E911 chipsets are also available and are currently being integrated into cellular phones for major U.S. telecommunications providers. Manufacturers of these chipsets include SiRF, Motorola, and RF Micro Devices. Additionally, many commercial heavy-duty vehicle products are available that use GPS technology, mostly for fleet management and tracking, and their manufacturers include Terion, Vetronix, and Qualcomm. Most major light-duty automobile manufacturers also offer GPS-guided navigation systems that incorporate GIS data on CDs and DVDs.

VDR Implementation Issues

GPS technology for recording vehicle location, direction of travel, and time can be integrated into a VDR in a variety of levels of sophistication. At the lowest level, E911 chipsets could be modified (if not used directly) for integration into a VDR. The E911 chipsets could receive GPS timing signals and the raw timing data (not location calculations) could be stored in the VDR. Post-processing of the data could then be performed either externally at a later date or by the VDR's microprocessor. This would result in a very economical means of recording location. Correction factors, similar to those used for DGPS, could also be taken into account at a later date (pending availability from a nearby DGPS fixed receiver) to provide a very accurate account of the vehicle's location for crash investigators or fleet management personnel.

A more advanced (and potentially more costly) implementation of geographic position could involve integrating the necessary GPS hardware to perform the distance calculations and location triangulation, similar to what a handheld GPS unit might have. This would provide the VDR with location, direction of travel, and time data directly and would minimize the post-processing requirements necessary. While this would require some additional cost (above the cost for E911-like hardware), systems and hardware are currently available and being manufactured by a variety of companies for commercial use.

Since VDRs are primarily targeted at accident reconstruction and operational efficiency improvements, integrating GIS data along with the geographic position information in a VDR would likely not prove overly beneficial. GIS maps of roadways, buildings, and terrain would provide useful information, specifically for accident reconstruction and causation analysis, but could be more efficiently overlaid during data analysis post event/trip. If, however, a VDR was to provide real-time data to the driver, then the integration of GIS data on top of the raw geographic position data would be relatively straightforward as many automotive, marine, and personnel navigation systems already use this technique.

This technology can easily be implemented into a VDR. Many of the products on the market have the capability to have additions, such as sensors, integrated into the product itself. Some systems, such as Qualcomm, are systems from which sensors and event triggers could be added. Other time and location technologies could be integrated into already existing VDRs through either a RS232 or a USB port. These connections could facilitate the storage and manipulation of location, time, and direction of travel data.

The most difficult challenge of integrating geographic position into VDRs is antenna placement. GPS receivers do require antennas to acquire the GPS signals, and since four GPS satellite signals are required, antenna placement can be important. Recent advances in GPS technology have allowed these antennas to become very small, specifically with the E911 mandate as they must fit into the compact space of a cellular phone. If a GPS antenna is integrated into the VDR package, then the VDR itself would have to be located in a specific location on the vehicle where it is able to receive GPS signals. An external GPS antenna is also an option, and most automotive (including commercial motor vehicles) navigation and tracking systems use an external GPS antenna mounted on top of the vehicle.

Additionally, it is likely that geographic position data will not be available at all times. If the vehicle is traveling through tunnels or under overpasses, the satellite signal could be interrupted. This will likely result in a temporary loss of location and direction of travel data if alternative algorithms are not employed to interpolate the location based on speed and direction of travel.

Infrastructure Requirements

The GPS satellite infrastructure is already in place. The U.S. military has completed the necessary satellites, land-based stations, monitoring equipment, and timing and synchronization equipment to provide GPS services globally. DGPS transmitters are in place globally, but DGPS coverage is still not universal. DGPS covers most of the continental United States, as well as Alaska, Hawaii, and bordering parts of Canada and Mexico, with future DGPS sites planned to fill in the remaining gaps (largely located in Idaho, Nevada, Oregon, Texas, West Virginia, Kentucky, and western Virginia and North Carolina).

Two other government-sponsored satellite navigation systems are available or in development, similar to that of GPS: (1) the Russian Global Navigation Satellite System (GLONASS), and (2) the European Galileo Global Navigation Satellite System (GNSS). Similar to GPS, these systems will provide global navigation coverage; however, it is still unclear as to what form of commercially available receiving hardware will be developed that uses this infrastructure.

Technology Implementation Costs

The cost to implement geographic position on a VDR is dependent on the level of integration necessary. Low-cost E911 chipsets are becoming increasingly more available as companies strive to meet the FCC mandate. These chipsets are small and compact, and in large quantities would likely cost around $25. GPS receiver boards that provide complete location, direction of travel, and time data (designed for integration into an automotive navigation system), such as those sold by RF Micro Devices, could cost as little as $40 for quantities over 10,000 per year, but likely would be initially in the range of $50 to $100. Commercial handheld and portable GPS units run from around $100 to $500, with units with advanced GIS topping out at $1,000. Complete automotive GPS navigation and/or tracking systems, some with built-in GIS or cellular location transmitting capability can range up to and over $2,000.

Digital Imaging

Technology Overview

Video imaging provides a unique approach for recording and analyzing vehicle accident events as it can often be difficult to understand crash dynamics by only interpreting data from sensors. Companies are beginning to look to video as a way of supplementing or even replacing the need for recording onboard sensors. Digital video differs from standard video in that video sequences are stored in memory as binary (1/0) digits instead of analog signals. With the development of low-cost and reliable digital video cameras and developments in video compression technology, it is becoming economical to install digital imaging and recording systems onboard vehicles. Cameras can be mounted on the front, sides, or rear of the vehicle, or even facing the driver to record all aspects of an event. Video information collected on the environment and the driver can play a significant role in determining the causes and effects of accidents. Showing actual footage of an accident or other event along with the other sensor information is one of the most effective ways to review causal factors and outcomes of accidents and other safety-critical events.

There are many aspects of digital video that effect how the image is perceived by the human eye, how much memory the video occupies, and how much it costs to produce and replay. Video can be decomposed into a succession of images (called frames) shown in rapid succession (called frame rate in Hz). Each frame is composed of tiny colored dots (called pixels). The most common format for video is 720x480 pixels per frame at 30 Hz (frames/sec). Assuming that the video is in 16 bits color (i.e., each pixel's color is represented in binary by 16 bits), one second of video would require over 16 Mbits (16 bits/pixel x 720 pixels x 480 pixels x 30 frames/sec = 160 Mbits) or 20 MBytes of memory. Therefore, a minute of video would require 120 MB, and a 90-minute movie requires 100 GB. These memory requirements for digital video are often termed as bit rate or Mbits per sec (Mbps).

The development of advanced video compression techniques facilitates the development of digital video systems, particularly in applications where minimal storage memory is available. Video compression is a method for reducing the memory necessary to store a video segment by reducing resolution, removing parts of the video that are not visible to the human eye, and eliminating duplicate elements of the video sequence that do not change from frame to frame.

The most common technique for video compression is the ISO standard MPEG, of which there are three variants-MPEG-1, MPEG-2, and MPEG-4. MPEG-1 is a low-resolution (typically 320 x 240, with a maximum of 352 x 288), low-bit rate (1.5 Mbps) standard designed for quality similar to VHS cassettes. MPEG-2 is a high-resolution (720 x 480, with a maximum of 720 x 576), higher bit rate (7.5 Mbps) standard. MPEG-4 uses a more sophisticated technique to deliver scalable high- to low-bit rate video from 10 Kbps to 10 Mbps. MPEG-4 is the current state of the art for video compression. Another technique for digital video compression is H.261 and H.263, but these are mainly designed for video conference applications where minimal bandwidth is available. Exhibit 3.4 provides a comparison of video compression.

Exhibit 3.4 - Video Compression Technique Comparison

The main components of a digital video recording system are cameras, a digital video recording input card, compression software (see previous discussion), and storage memory (see Section 3.1).

Cameras are readily available and have been used for many years in surveillance equipment. Both black-and-white and color cameras are available in a variety of configurations-domes, balls, mini-pinhole cameras. Also, many styles of cameras come in a variety of levels of weatherproofing and casing, making them ideal for vehicle applications.

The most sophisticated hardware is the digital video recording (DVR) hardware. The DVR hardware is basically an analog-to-digital converter that converts the analog video signal from the camera into digital video. The hardware usually also includes the necessary software to perform the video compression. While this hardware could be incorporated directly onto the VDR's circuit board, it is usually in the form of an expansion card plugged into an expansion card slot (e.g., PCI, PC104, PCMCIA slot) in the VDR's motherboard. DVR cards vary in complexity and cost based on the number of video inputs (usually 2 to 16), frame rate, resolution, and compression technology.

There are digital video recording products and packages currently on the market for vehicle applications. The following are brief descriptions of some of the products that are commercially available.

• Assistware Technologies has developed a digital imaging product that analyzes and "grades" a driver's performance in maintaining position within a lane. The camera is mounted in the front windshield, and processes the white lines of the road ahead. This white line information is digitized and presented on a digital display in front of the driver. A numeric number represents the performance of the driver based on driving within the white lines of the road. This number represents the driver's alertness and is available through an RS232 connection. In addition, the forward video image is an RGB video signal and can be recorded prior to, during, and after the accident.
• SeaView Technologies is offering Power Line Carrier (PLC) video technology to the transportation industry. The technology will allow video communication via the existing power wires; thus, cameras can be incorporated into the lamps of the vehicle. For the commercial market with tractor/trailers, the existing SAE J560 connector can be used without an additional cable between the tractor/trailer. Furthermore, the PLC video technology can coexist with the FMVSS-121 PLC communications for the ABS warning. The technology will allow 360-degree visibility such as passenger-side visibility during right turn activation and trailer-rear visibility during backup lamp activation. This video image is displayed on a CRT/LCD display via an RSG video signal, which will be available to VDR systems.
• Loss Management Systems has developed a product called the MACBox (Mobile Accident Camera). This product not only provides a driver's eye view of an accident, but also provides vital information such as speed, brakes, acceleration in three axes, safety belt status, and distance to the vehicle in front. MACBox provides digital imaging before, during, and after an incident to provide the maximum amount of data.
• Safety Vision has a fully operational VDR system with its Road Recorder 5000. This product not only has video before, during, and after the event, but also has the data needed to further reconstruct an accident. It receives this data from an SAE J1708 interface and also through multiple sensor inputs. The Road Recorder 5000 also includes a GPS interface, which records longitude, latitude, speed, and direction of travel.
• DriveCam's Video Event Data Recorder is a comprehensive system that is either triggered by an event or manually. This system includes a real-time clock and provides information regarding the brakes, acceleration, and "Lateral G-Forces". This information not only could be used to reconstruct an accident, but also to monitor a driver's erratic driving, which could prevent accidents in the future.

VDR Implementation Issues

Implementing digital video recording in a VDR is straightforward and can be done with commercial-off-the-self (COTS) hardware. The surveillance industry has been moving toward digital video and has commercialized much of the hardware. Storage memory size is an issue, but with the latest MPEG-4 compression technology, video can be stored at varying levels of resolution.

In most, if not all, applications, video recording would be based on a triggered event. Since it is difficult to design software to discern objects or events in video directly, a sensor (e.g., accelerometer and brake pedal) would have to be integrated with the recording system to trigger the video capture.

Technology Implementation Costs

The cost of digital video recording varies with the level and complexity of the system. Black-and-white cameras can range from as little as $50 to $400 for a weatherproof unit. Color cameras start around $100 to $500. Smaller, circuit board-mounted cameras are around $100 for black and white or $150 for color. It should be noted that not all camera types and equipment can operate effectively in the commercial vehicle environment. Current camera systems can be susceptible to:

• Damage and degradation of quality from extensive vibration and shock
• Failure in extreme hot and cold climate conditions
• Image degradation in viewing conditions due to poor weather and lighting
• Electronic failure due to complex parts and powered components
• Failures related to camera positioning, field of view, distance, and obstructions in viewing area

It should be recognized that additional costs may be required to mitigate these issues and integrate a camera into a VDR designed for the commercial vehicle environment.

Digital video recorder cards can range from $80 for an economy single channel card to over $4,000 depending on the level of complexity. Typically, PCI-based cards with average performance are $100 for single channel, $150 for 2-channel, $300 for 4-channel, $700 for 8-channel, and $1,300 for 16-channel.

To better understand the cost and memory requirements of various levels of digital video recording, Exhibit 3.5 shows some typical operating scenarios for a digital video system in a VDR. The exhibit is not intended to be an exact estimation of the costs, but rather to provide a basis for how levels of digital video integration affect cost and storage.

Exhibit 3.5 - Example Digital Video Requirements

Sensors for Determining the Relative Location of Nearby Vehicles

Technology Overview

VDRs, specifically those designed to record accident event data, could benefit from an understanding of the relative location, speed, and direction of travel of nearby vehicles. This information would be particularly useful in accident reconstruction and crash causation determination.

Two main technologies have become popular for measuring the location of surrounding vehicles- radar and ultrasound. Radar and ultrasound use different media (radio waves and sound waves, respectively) in a similar manor to determine the speed and range of surrounding vehicles. A transmitter sends out a short, high-intensity pulse of high-frequency radio or sound waves. A receiver then listens for an echo and measures the time it takes for the echo to arrive and, with radar, measures the Doppler shift (change in frequency) of the echo.

Ultrasonic sensors use sound wave to detect objects and can be located in multiple locations around the vehicle to provide a 360-degree view. Ultrasonic sensors have long been used for parking aids and for detecting the range of objects-they are generally not used for detecting object speed and direction of travel. Ultrasound technology is mature and sensors are one of the lowest-cost technologies on the market. The major disadvantages of ultrasound are its limited range (usually less than 15 feet), inability to detect object speed and size, and inability to detect objects that absorb sound. One example of an ultrasonic sensor for commercial vehicles is Transportation Safety Technology's Eagle Eye system, which can utilize up to seven sensors and has a dash-mounted visual and audio display. It is intended to assist the driver during maneuvers such as backing up, but can also be used to assist during lane-change maneuvers by detecting objects in the vehicle's blind spot.

Radar sensors use high-frequency radio waves to detect objects' range, range rate, and azimuth. Typical radars use both monopulse (to determine range and azimuth) and Doppler radar (to determine range rate). Radar has become popular for use in forward-looking collision warning systems due to its long range (>350 feet), good range and speed resolution, and ability to estimate object size. The major disadvantage of radar is cost. The most common radar sensor for commercial vehicles is Eaton's VORAD Collision Warning System. VORAD uses both monopulse and Doppler radar in the 24 GHz band to measure up to 12 objects in a 12-degree radar beam. It has a range of 350 feet, and has an integrated yaw sensor to determine when objects are directly in front of the vehicle during turns. It is intended for use as a forward collision warning system, but also can be equipped with up to two side sensors for detecting objects located in the vehicle's blind spots. In addition, the VORAD system can also be equipped with an adaptive cruise control feature that allows a driver to set and maintain a desired following interval to the object directly in front rather than a particular speed setting.

A second forward-looking radar-based system is being introduced to the commercial vehicle market by Meritor WABCO. The system includes collision warning, adaptive cruise control, and collision mitigation functions. The collision warning and adaptive cruise control features operate in a similar manor to that of the Eaton VORAD system. Collision Mitigation, as Meritor WABCO has termed, initiates active braking for impending collisions both when this system is in adaptive cruise control mode and when the system in only in collision warning mode. This feature controls vehicle speed and braking by controlling the engine speed/torque, the engine brake, and applying the service/foundation brakes. Currently this system does not include side-mounted radar.

VDR Implementation Issues

It is unlikely, especially in the near term, that radar or ultrasonic sensors would be integrated directly with the VDR, but rather the VDR would record data from a separate radar or ultrasonic sensor package (e.g., Eagle Eye or VORAD). Transmitting data from a radar or ultrasonic sensor to a VDR could be straightforward if the object or target data was available on the vehicle's databus (J1939/J1587). If the particular sensor package was not connected to the databus, it would be necessary to have additional analog or digital inputs to the VDR from the sensor. While it is technically feasible, integrating a radar or ultrasonic sensor with a VDR would be very costly as these systems require a great deal of processing and electronics to properly identify objects and determine their range and speed. It is more likely that, to record the location of nearby vehicles, a VDR would communicate with a separate radar or ultrasonic system or package over one of the vehicle databuses.

Infrastructure Requirements

No infrastructure requirements would be necessary to support radar or ultrasonic sensors. However, it may be necessary to work with the industry and the FCC to further refine the operating frequencies allowed for these systems to prevent them from interfering with other devices and to prevent other devices using radio or sound waves from causing false alarms. For instance, typical radar-based systems currently use the 24 GHz band, which is also shared by other devices (namely police K-band radar). The FCC has allocated frequency spectrum in the 76-77 GHz band for vehicle radar-based systems to avoid these interference issues. The Eaton VORAD system has been operating on a waiver from the FCC to operate its system in the 24 GHz band. It is anticipated that future generations of vehicle radar systems would operate in the 76-77 GHz band.

Technology Implementation Costs

Radar and ultrasonic sensor packages are currently commercially available and, while not commonplace, have gained acceptance in the industry for use in rear-end and side-collision avoidance and parking assistance. As previously discussed, it is likely that data from these packages will be used in VDRs-at least in the near term. The cost of these systems, specifically radar-based systems, will likely mean that additional product features (e.g., collision warning, adaptive cruise control) will be necessary to justify the cost. The Eagle Eye ultrasonic sensors cost approximately $750 per vehicle and have seven sensors that can be located around the vehicle. Currently, Eaton VORAD radar-based collision warning systems cost approximately $2,000 per vehicle for a basic system. The optional side sensors and adaptive cruise control (called SmartCruise) increase the cost. Exhibit 3.6 shows an estimated cost breakdown for the VORAD components.

Exhibit 3.6 - Estimated Eaton VORAD Component Costs

* Based on estimated retail service parts cost

Short-Range Communications

Technology Overview

Short-range wireless communication has been increasing in popularity for computer and network systems, and is just being introduced into the automotive market. Short-range wireless communications devices work by attaching a transmitter to a data source (such as an EDR) and "shooting" the data to a receiver using various transmitting media and protocols. The receiver operates over the same medium and protocol, collects the data, and formats it to be used in various applications. The two main short-range communication media are infrared (IR) and radio-frequency (RF). The most popular IR standard, IrDA, was developed by the Infrared Data Association and is commonplace in most PCs, PDAs, printers, and other computer systems. Two major RF protocols have gained widespread popularity:

• WiFi - Developed by the Institute of Electrical and Electronics Engineers (IEEE) and commonplace in wireless computer networks.
• Bluetooth - Developed by the Bluetooth Special Interest Group comprised of leading manufactures in telecommunications, computing, and networking, and designed for improving personnel connectivity between portable computers, mobile phones, PDAs, and other handheld devices.

A third RF standard, Dedicated Short-Range Communication ( DSRC), designed specifically for vehicle-to-infrastructure and vehicle-to-vehicle communications is currently in development and leverages the WiFi standards framework.

IR light communication has been standard in many commercial products such as laptop computers, printers, digital cameras, and PDAs for many years. The Infrared Data Association's IrDA standard is comprised of two segments, Data and Control. The IrDA standard defines an interoperable universal two-way cordless IR data port used primarily for device-device data transfer. IrDA Control allows cordless peripherals (e.g., keyboards, mice, joysticks) to interact with intelligent host devices (e.g., PCs or PDAs). IrDA Data, the most applicable standard for VDRs, is used for high-speed, short-range, line-of-sight, and point-to-point data transfer, and requires continuous operation within 1 meter (20 cm for low power consumption devices). IrDA data operates from a transmission rate of 9600 bit/sec up to 4 megabits/sec.

The most popular RF technology, WiFi, is spreading rapidly across multiple markets, from computer networking to automotive communications. There are three main WiFi standards-802.11a, 802.11b, and 802.11g, with 802.11b being the most popular. Finalized four years ago, 802.11a works in the 5- to 6-GHz band at speeds of up to 54 Mbps, but with a limited range of about 60 feet. The 802.11b RF standard operating in the 2.4-GHz band delivers speeds of up to 11 Mbps with a range of about 330 feet. The third WiFi standard, 802.11g, operates in the same 2.4-GHz band as 802.11b, but provides speeds of up to 54 Mbps, with a similar 330-foot range, and is completely interoperable with 802.11b.

The DSRC standards currently in development by the ASTM and IEEE working groups leverage the WiFi 802.11a standards and are designed to operate in the 5.9 GHz range (designated by the FCC as ITS Radio Service). The DSRC frequency band is subdivided into seven channels dedicated to specific functions:

• One 10-MHz channel dedicated specifically to vehicle-to-vehicle communications
• Four 10-MHz service channels for public safety and private communications (two 10-MHz channels can be combined to create a 20-MHz channel)
• One 10-MHz control channel for announcements and warnings
• One 10-MHz channel for high-power public safety, intersection, and emergency vehicle communications

DSRC is intended to support both public and private applications including:

• Emergency and transit vehicle signal preemption
• Roadway condition warning
• Vehicle-to-vehicle collision warning
• Intersection collision warning
• Cooperative vehicle systems (including platooning)
• Private access control and payment
• Data transfer and infotainment
• Fleet management
• Vehicle safety inspections
• Onboard safety and emissions control data transfer

The range of DSRC depends on the application and channel allocation, but can be up to 3,000 feet for emergency vehicle services and approximately 1,000 feet for data transfer and messaging services. DSRC hardware will be very similar to that of 802.11a since the lower layer standards (physical and medium access control) are based on 802.11. Field testing of the first DSRC system is just starting. The Crash Avoidance Metrics Partnership, a light-duty OEM venture to facilitate collaboration on crash avoidance research, is currently working to define communication parameters for a variety of safety applications.

It is also worth noting that a second ITS Radio Band in the 902-928 MHz band has also been designated by the FCC. It is currently used for toll pass applications and emergency vehicles. Its limited range (300 feet), limited data rate (0.5 Mbps), interference potential (i.e., 900 MHz phones), and implementation cost make it an unpopular option for future development, especially with the 5.9 GHz DSRC standard almost complete.

The popular RF wireless communications standard Bluetooth was designed primarily for interconnection of portable and handheld devices. Bluetooth uses the 2.4-GHz band that is based on the IEEE 802.11 standard. However, Bluetooth requires little power and is intended for transmitting small amounts of data over short distances (up to 10 meters). Bluetooth operates at a speeds between one to two Mbps, which is significantly slower than WiFi. However, Bluetooth chips and components are cheaper and more compact than those needed for WiFi, and have lower power consumption. Bluetooth is still an emerging technology and is, therefore, less mature than WiFi. The low cost, compactness, and low power consumption of Bluetooth hardware make it ideal for portable applications such as mobile phones and PDAs, where WiFi would not be viable.

Wireless infrared IrDA products are available from many commercial manufacturers, including Acer, Agilent, Intel, Maxim, National Semiconductor, and Texas Instruments among others. In addition, many manufacturers of computer components offer completely integrated IrDA products and add-on IrDA transceiver peripherals. Detroit Diesel Corporation (DDC), a subsidiary of DaimlerChrysler Corp., markets and distributes an aftermarket IrDA product called IRIS for the heavy-duty vehicle industry. This product interfaces to the SAE J1708 databus and retrieves information useful to mechanics and fleet managers, and that could be used for investigating accidents. The IRIS product has the capability of bi-directional communications at 9,600 baud and costs about $300.

WiFi products operating on the 802.11b standard are readily available in multiple levels of integration. New products based on the 802.11g standard are just emerging in the marketplace, while 802.11a products (released over the last two to three years) have not found a large acceptance due to their limited range and the new compatibility of 802.11g and 802.11b systems, but might begin to reemerge in the automotive industry as DSRC continues to develop and the first DSRC systems begin to be tested. Many commercial manufacturers offer a wide range of WiFi products, from chipsets to consumer WiFi transceivers. WiFi chipset manufacturers include Atheros Comm., 3Com, Intel, Texas Instruments, and Intersil Corp. Manufacturers of consumer WiFi transceivers include Cisco, Apple, D-Link, Linksys, NetGear, Microsoft, Vivato, Proxim, and Dell.

Bluetooth technology is still emerging, but many commercial companies are currently offering Bluetooth hardware modules (complete chipsets), including 3Com, Ericsson, IBM, LG Innotek, Lucent, Motorola, Siemens, Philips, and others. Many commercial products are also available with Bluetooth technology. Nokia, Ericsson, Motorola, and LG manufacture cellular phones with Bluetooth technology. IBM, Motorola, Xircom, National Semiconductor, Hewlett Packard, and 3Com manufacture Bluetooth PC cards for wireless connectivity of laptop PCs. Visteon, Nokia, Kenwood, Johnson Controls, DiamlerChrysler, and several other manufacturers have Bluetooth-enabled automotive products largely focused around hands-free functionality of cellular phones.

VDR Implementation Issues

Integration of short-range wireless communication using IrDA, WiFi, or Bluetooth standards into a vehicle data recorder would be relatively straightforward as these technologies (especially IrDA and WiFi) have been commercialized heavily and have had a high level of acceptance in the industry. For each of these technologies, there are two possible approaches. The first approach would be to integrate the wireless chipsets, transceivers, antennas, and software into the VDR hardware directly. This would likely reduce costs, reduce power consumption, and make the VDR more compact. A second approach would be to use a commercially available IrDA, WiFi, or Bluetooth card that could be integrated into the VDR through a PCI, PCMCIA, or other bus slot. This would likely increase the flexibility of the VDR as hardware could be switched between either of the three protocols.

The most difficult challenge of integrating short-range wireless communication into the EDR is antenna and transceiver placement. With infrared, direct line of sight is necessary between the VDR and an external transceiver. If the transceiver is a handheld unit, then placement of the VDR infrared transmitter could be internal to the vehicle cab or in another protected location. However, if the external transceiver is in a fixed location, the infrared VDR transmitter will have to be located on the vehicle such that the two align to transfer data (e.g., perhaps the vehicle passed by a gate reader). With WiFi or Bluetooth technology, line of sight is not necessary, but the sheet metal body of a heavy-duty vehicle would likely limit the range of these technologies, especially if the antenna is located within the metal structure of the vehicle (e.g., cab, engine compartment, storage compartment). Therefore, it is likely that an external antenna will be needed to obtain the necessary range to transmit data from a VDR to a stationary external receiver. As with the infrared technology, if a handheld receiver is used, the placement of the antenna would be less critical.

Infrastructure Requirements

The infrastructure requirements are similar for each of the three short-range communications protocols. As these technologies are intended for local broadcasting of data, receivers would likely need to be installed at locations where VDR data is needed (e.g., maintenance garage, inspection and weigh station, or fueling station), or handheld or mobile devices would be needed. Receivers for WiFi and Bluetooth technologies would have to be installed so that they are within effective range of the VDR, while infrared would require direct line of sight to the VDR transmitter. Stationary infrared, WiFi, and Bluetooth receivers and base stations are readily available and could easily be integrated into a PC software package to collect, filter, and store data from a VDR that enters its range. In addition, portable receiver hardware is also readily available and would only need to be integrated into a handheld or mobile device.

The commercial infrastructure for the WiFi technology is continuing to expand. Access points called "Hot Spots" are being created for a variety of applications, mostly centered around wireless high-speed Internet connectivity. As WiFi proliferates, it will most likely serve large populations in concentrated areas, such as downtown districts, universities, and business centers. Companies such as Cometa Networks intend to build a 20,000-node nationwide WiFi network before 2007, while AT&T, IBM, and Intel will deploy a nationwide network of public wireless access points, named Project Rainbow. SiriCOMM Inc. is also planning a nationwide wireless access point network that will be based in full-service truck stops. It remains to be seen how these access points will integrate with DSRC systems since they are mostly 802.11b or 802.11g and operate in a different frequency band.

Intel, Nokia, Proxim, and other companies have launched WIMax, a non-profit group formed to certify and promote the wireless broadband standard 802.16. This standard will apply to communications equipment that will connect 802.11 access points to the Internet and provide a wireless extension to cable and DSL for last-mile broadband access (between local hard-wired access points and wireless customers). The 802.16 standard provides for up to 31 miles of linear service area range and allows users connectivity without a direct line of sight to a base station. The technology also provides data rates of up to 70 Mbps with enough bandwidth to simultaneously support more than 60 businesses (or heavy-use modes) with high-speed connectivity. Hot Spots connected to the Internet through 802.16 will offer wireless access to citywide areas bringing WiFi closer to cellular network levels of availability.

Due to the limited range of both IrDA and Bluetooth, commercial infrastructure is limited (almost nonexistent for IrDA due to its requirement for line of sight). It is not likely that wide-scale Bluetooth or IrDA access points will be available nationally.

Technology Implementation Costs

Bluetooth chipsets that could be integrated directly into the embedded hardware of a VDR or handheld receiver can cost as low as $5 to $10, but would require significant development and integration costs. Commercially available Bluetooth adaptor cards designed for PCI or PCMCIA buses range from $150 to $250 and require less development time and cost, but would also be less compact and flexible, and consume more power.

WiFi chipset prices have been falling steadily as the market grows and more companies begin manufacturing 802.11-compliant hardware. In 2002, the average price of 802.11b chipsets was $16, while in February 2003 the price reached $6. In June of 2004, WiFi 802.11b chipsets cost between $4 and $6, while 802.11g chipsets cost $12 to $13. Manufacturers are now supplying combination 802.11b/g chips for $14 to $20. WiFi adaptor cards cost between $30 and $150 and are commercially available. It is likely that DSRC equipment will be priced in a similar range.

IrDA chipsets cost between $1.50 and $2.50 for the IrDA controller. An additional infrared receiver/transmitter will be necessary at a cost of around $2.50 in production quantities. Commercially available IrDA adaptor cards for PCI or PCMCIA buses are also available and would cost between $50 and $150.

Long-Range Communications

Technology Overview

Advanced long-range wireless communications devices were introduced into the transportation industry in the 1980s. Long-range communications technology, paired with a VDR (or integrated with VDR-like functionality), would provide another dimension to monitoring vehicle and event data. Long-range vehicle communications has been used by fleets for several years, mainly for vehicle tracking purposes. Recently, however, fleets and service providers have begun implementing recording and vehicle health monitoring capabilities within the long-range communications packages. There are two main avenues for information to be exchanged using long-range communication products-satellite-based technologies and land-based technologies.

There are two main satellite-based technologies available to North American users-geosynchronous satellites and low earth orbiting satellites (LEO):

• Geosynchronous means orbiting at the same rate and position relative to the earth, which is the reason, from Earth, a satellite in geosynchronous orbit appears to remain stationary over one spot on the Equator. This type of satellite is not motionless. It is in a very high orbit where it circles the Earth once a day, matching the Earth's rotation on its axis. This allows a receiving dish to spot the satellite at one point in the sky-and it does not have to track its movements. There are many satellites in this type of orbit that are needed for important and timely information such as weather satellite pictures, satellite television, or communications. Geosynchronous satellites are reliable enough for the National Aeronautics and Space Administration (NASA) to use them to relay communications and data between spacecraft and control centers on earth. Qualcomm, a supplier of commercial vehicle long-range communications equipment and services, uses satellites in geosynchronous orbit.
• Low orbiting satellites or LEO satellites are the second type of satellite technology used for long-range communication. An LEO satellite orbits generally 500 to 2,000 km above the surface of the earth, and is typically grouped with several other LEO satellites forming a constellation that achieves wide-area coverage with lower power requirements and shorter propagation delays (delays between when a signal is transmitted from a satellite and received by a receiver on the ground, or vice-versa). Two distinct LEO groups have emerged-Little LEO for data-communications satellites and Big LEO for data-and-voice communications, each using pre-assigned frequency ranges. The LEO satellite system works by employing a large constellation of satellites, each in a circular orbit at a constant altitude of a few hundred miles. The orbits of the satellites take them nearly over the geographic poles. Each revolution takes anywhere from about an hour and a half to a few hours. The satellite groups are arranged in such a way that, from any point on the surface at any time, at least one satellite is in line of sight of a receiver. LEO systems operate in a manner similar to the way a cellular phone works. Unlike geosynchronous satellites, LEO satellite systems require directional antennas whose orientation must be constantly adjusted to follow the satellite's path across the sky. Orbcomm, a long-range communications provider for commercial vehicles, uses satellites in an LEO constellation.

Land-based technologies are also available to support long-range communications. Most land-based systems use cellular phone technology as a way of distributing and collecting information. Most land-based systems are very reliable in urban settings where a cellular system is readily available and working, but if these products move to areas where cellular towers are either non-existent or sparse, the service quality diminishes. Over the last 10 years, the cellular coverage area has increased dramatically, particularly in urban areas and along major transportation corridors.

Mobitex is similar to the digital cellular technology. The network consists of interconnected cells, each of which is served by a radio base station that provides wireless access to the network for mobile users. The base stations and other network nodes are connected together by fixed links. Despite these basic similarities, Mobitex is unique in several key respects. It is a narrowband data-only network that uses packet switching, and it is based on an open international standard. A packet-switched network breaks the data stream up into small packets, each of which can be sent across the network individually. Unlike voice cellular networks, dedicated connections are not needed, and network access is virtually instantaneous because each data packet contains the destination address and can be routed dynamically as network conditions change. For subscribers, the benefits are that they are always connected and online, yet pay only for the number of packets sent-not connection time.

Long-range communications technologies will continue to play a major role in the VDRs for the transportation industry. Many fleets use this long-range technology for transmitting fleet and truck data. Dispatchers know when and where the vehicle is, as well as other information they need to maintain the vehicle and maximize the productivity of the fleet. Current long-range systems give the capability of sending large amounts of data in very short periods of time, sometimes just a few seconds. It would allow accident data to be sent to the fleet, police, and emergency crews in a matter of a few seconds. These systems could also send information needed to properly determine the reason for the crash. The benefits implemented into long-range systems are the reason many vehicles already use such equipment. This technology improves the performance related to trucking and, more importantly, will give a clearer picture of the cause of certain events.

VDR Implementation Issues

Satellite-based communications systems are self-contained and integrated into the trucks onboard databuses.

Infrastructure Requirements

The majority of the infrastructure and Qualcomm and Orbcomm satellites are already in place. New satellites are sent up periodically to replace old and worn-out equipment.

Cellular networks for land-based systems have been in place for the past decade and are growing on a daily basis. Older analog cellular systems are beginning to be replaced by digital networks that provide more reliable and higher speed data transmission. Also, with the advent of General Packet Radio Service (GPRS) technology, the amount of area in the United States covered by cellular networks is rapidly approaching 99 percent. However, there are still numerous gaps in the cellular infrastructure.

Mobitex networks consist of the Mobitex base stations, routers, and the network control centers. Mobitex base stations are typically 13"x17"x7"; router/switch racks are 29"x24"x31"; and the network control center consists of servers, client computer, and terminals with LAN connections to the routers and switches. A Mobitex network might consist of a single network control center connected to several router and switches, each connected to multiple base stations covering an area from a few hundred feet to nationwide. Mobitex networks can be configured in different ways, from a large public nationwide network to a small, privately owned network serving a single region or company. The networks can be configured to match a particular service mix and requirements and re-configured as service uptake and demand changes over time.

Technology Implementation Costs

There are many commercially available products on the market today. The following are several products that incorporate different types of long-range communication. Prices range from a few hundred dollars to a few thousand dollars, with the land-based technologies the less costly of the two.

• Qualcomm is the leader in the long-range communication market. It has approximately 400,000 users in the trucking industry, using geosynchronous satellite-based infrastructure to provide geographic location information and packet data bi-directional communications. Qualcomm uses a free-form data packet communication format via SAE J1708. The Qualcomm free-form data communications would be a means to collect VDR information and send it to a home base for analysis. This free-form packet data is similar to a zip file and can only be decoded by the appropriate and approved system. The Qualcomm system, OmniTRACS, varies in cost depending on the number of vehicles to be equipped, from about $1,000 per vehicle to around $3,000 per vehicle for the communication equipment and software. Qualcomm also requires the purchase of monthly service plans that range in price from $35 to $50 that may include the frequency and type of data transfers a vehicle is permitted to communicate. The $35 plan is basically a connection fee that only includes one GPS position poll per hour. The $50 plan also includes one GPS position poll per hour, but also allows for data transfers of up to 180 messages or 18,000 characters. As an alternative to its satellite communication system, Qualcomm offers a terrestrial-based communication system using the Sprint PCS cellular network called OmniExpress. The OmniExpress system can communicate the same amount and type of data as the OmniTRACS system, and can do so at a much lower cost. The tradeoff with this lower cost, however, is that the OmniExpress system can only transmit or receive data while the vehicle is in range of a cellular tower.
• Orbcomm is an example of an LEO application. The company owns and operates a network consisting of 30 LEO satellites and terrestrial gateways deployed around the world. Small, low-power, and commercially proven subscriber communicators can connect to private and public networks via these satellites and gateways. Through this network, Orbcomm delivers information to and from virtually anywhere in the world on a nearly real-time basis. Volvo Trucks North America has selected the Orbcomm system to provide two-way data communications for a tracking and monitoring for use on heavy trucks. Orbcomm can be interfaced to a VDR system via RS-232.
• Terion uses a land-based system for trailer tracking and security. Its FleetView product is used to track trailers throughout the United States. Data is communicated using a cellular-based infrastructure. FleetView has many benefits such as reducing the trailer-to-tractor ratio, eliminating yard checks, minimizing loss from theft, and allowing fleets to be more responsive to their customers. The FleetView unit mounts on the front inside wall of the trailer, while the antenna is enclosed within a marker light. FleetView comes with a standard movement sensor to record the beginning and end of a drive. The system can also report when the door is opened or closed or when the trailer is loaded or unloaded with added sensors.
• XATA's Mobile Asset Management solutions integrate enterprise software, onboard computing, global positioning, and wireless communications to provide an integrated asset management system. XATA systems support multiple onboard computing platforms and have their own proprietary onboard computer, called the XATA Application Manager (XAM). The XAM incorporates a microprocessor, dual-mode wireless communications, and a 12-channel GPS receiver, all housed in an aluminum alloy base with a UV-resistant plastic dome. The XAM is equipped with both an Orbcomm satellite modem and a CDMA cellular terrestrial modem to communicate wirelessly with a host server. The CDMA cellular modem is the primary communication system and sends and receives data over the Sprint PCS wireless network. The satellite modem is used to communicate data only if the vehicle is not within range of a cellular tower. GPS location satellites ping the XAM every 30 minutes to track the vehicle's location and provide location data for use in state fuel tax calculations and geo-fencing applications. All other data collected by the XAM, such as vehicle diagnostics and driver data, is compressed into a zip file and transmitted once a day (or more frequently if necessary) to a host server managed by Agility Web Hosting. The data can then be used in fleet management applications and remain on the host server, or it can be transferred to the fleet's back office computer system via Microsoft Web Services. XATA's application packages encompass multiple applications, including:
• Asset tracking
• Operations profiling
• State fuel tax calculating
• Vehicle call back
• Daily mileage reporting
• Diagnostic warnings reporting
• Accident analysis
• DOT driver logs
• XATANET can connect to a range of display devices including driver terminals, in-cab computers, printers, handheld computers, and other peripherals. XATANET can be coupled with a variety of wireless services such as WiFi, cellular, or satellite. However, nearly all XATANET systems produced today are equipped with an Orbcomm satellite modem and a CDMA cellular modem. The cost to incorporate the XATANET system consists of the initial purchase of the equipment and a monthly service plan that is similar in structure to a cellular phone plan. The monthly plans are billed as a flat rate, not by number of transmissions, and vary in price based on the amount of features included in the plan. To install XATANET on a fleet of 400 trucks, for example, the cost of purchasing the XATANET hardware is approximately $1,700 per vehicle. The monthly services charges vary by plan type and are listed in Exhibit 3.7.

Exhibit 3.7 - Typical XATANET Monthly Service Charges

The optional premium support package shown in the exhibit includes a rapid replacement parts service, an extended warranty on equipment (two to five years), 24/7 technical support, and automatic software updates sent wirelessly.

• Tripmaster is another manufacturer of asset management solutions. Its product line includes long-range and short-range communications, GPS, route scheduling, maps, handheld input terminals, time clock, driver performance monitoring, and fuel tax calculators. The base of the Tripmaster system is the DT-240/GPS ECU, which incorporates a GPS receiver, I/O ports, and J1708 connectivity. There are several input terminal options, from portable to hard-mounted. Tripmaster's wireless LAN provides a short-range wireless connection solution to a terminal. The long-range communication option, Mobilecom, uses cellular GPRS technology to send and receive data. The Mobilecom unit costs approximately $2,500 per vehicle and requires a monthly service plan that costs roughly $45. The DriveRight package allows fleets to set speed, acceleration, deceleration, and other limits and then records when those limits are exceeded. The DriveRight Trip 500AL hardware provides a driver performance score, driver ID, and location; stores data from over 500 independent trips; and maintains 5 accident logs. The unit also contains a tamper indicator as notification of a disconnection or if an unauthorized person attempts to change limits or other settings.
• Trimble's Telvisant Mobile Resource Management (MRM) platform includes a line of intelligent vehicle tracking units that determine and report vehicle location and other data without driver involvement. Trimble currently offers two "smart" mobile units that mount in the vehicle and an optional in-cab messaging terminal. The CrossCheck GPRS 1900 Mobile Unit consists of a GPS receiver, GPRS wireless communication platform, and computer processor-all integrated into a single compact enclosure. The unit can be used as a standalone device to report vehicle location and speed data, or it can be interfaced with external sensors to monitor vehicle status and activate security systems such as ignition lockout and audible alarms. The CrossCheck GPRS can also be interfaced with the EchoLDX, which is an in-cab messaging terminal that is offered as an optional accessory. The EchoLDX allows the driver and fleet dispatcher to communicate via text messages by displaying and storing messages received from the dispatcher, and enabling the driver to send a predefined response back to the dispatcher. The CrossCheck GPRS 1900 Mobile unit costs approximately $550 per unit and requires a monthly service charge of roughly $40. The EchoLDX messaging terminal costs $235 and requires either the CrossCheck GPRS or the CrossCheck CDPD mobile units to operate. The CrossCheck CDPD (cellular digital packet data) mobile unit is nearly identical to the CrossCheck GPRS mobile unit, except it communicates over a CDPD wireless platform instead of a GPRS wireless platform. Trimble is currently developing another intelligent vehicle tracking unit that communicates using CDMA technology.
• PeopleNet offers a wireless fleet management system that communicates between the vehicle and the PeopleNet Network Operations Center (NOC) using cellular technology. PeopleNet has partnered with over 180 cellular service providers to provide its system with a wide operating area. The PeopleNet system can be customized to meet the specific needs of a fleet vehicle, but will typically include a base computer unit, message display or PDA type device, voice handset (hardwired cell phone), PerformX engine diagnostics cable, and a keyboard. The system is also equipped with a GPS receiver that is pinged every 15 minutes on long-haul vehicles or as often as every 5 minutes on local route vehicles. The PerformX tool monitors the vehicle's engine performance by connecting directly to the vehicle's J1708 databus to record data such as speed, distance, fuel efficiency, and idle time. The PerformX tool does not currently report J1708 fault codes, but PeopleNet is developing this capability. PeopleNet is also developing the capability to enable the driver to use printers and scanners, and to access the Internet from inside the cab.
• There are also products implemented by light-duty OEMs in the market today. One such product is the OnStar product offered by General Motors. This is a land-based system that automatically communicates to the emergency officials once an airbag has been activated, and is also used for driver communication.

These are a few examples of commercial products available on the market today. There are other products that use similar technologies, but these perform similar functions as those mentioned before. Overall, long-range communication products are and will continue to be beneficial to fleet logistics management, and could provide a conduit for VDR data transmission.

Driver Performance

Technology Overview

Recording driver distraction and drowsiness levels could provide important driver training and monitoring information to a fleet's dispatchers and management. In a survey conducted for NHTSA in 2002 by The Gallup Organization, 3.5 percent of all drivers (light- and heavy-duty) involved in a crash in the past 5 years attribute it to being distracted-equating to an estimated 6.0 to 8.3 million drivers. The same survey showed that 37 percent of drivers have nodded off for at least one moment or fallen asleep while driving at least once in their driving career-8 percent have done so in the last 6 months. While this survey did not particularly call out commercial drivers, it does illustrate the dangers of distracted and drowsy drivers. FHWA performed a commercial vehicle driver fatigue and alertness survey in 1995. Of the 511 interviews conducted, 14 percent of drivers admitted to dozing or falling asleep at the wheel one to two times within the past month-9 percent admitted to dozing or falling asleep three to six times.

There are several commercially available systems that can detect drowsy or distracted drivers, most use one of two methods: (1) lane position monitoring, or (2) driver eye monitoring.

Lane position monitoring systems use video cameras and digital video processing to determine the location of the vehicle relative to the lane. In some systems, this data is supplemented by forward-looking radar. The digital video processing tracks road features (e.g., lane markings) to determine the relative location of the truck. It then provides alert indicators (i.e., audible and/or visual) when the vehicle begins to drift out of the lane. Lane position monitoring to detect weaving and erratic steering behavior can provide an indication of driver fatigue or inattentiveness. Since these are video-based systems, they often have difficulty tracking during adverse conditions (e.g., white-out snow storm, heavy and nighttime rain, or glare). There are several companies offering lane guidance systems, including Visteon, Delphi, and Assistware Technology.

Assistware Technology has developed the SafeTRAC lane guidance system targeted toward commercial vehicles. The system uses a small video camera to monitor the road and digital image processing to track road markings and determine the vehicle's location in the lane. It then provides an audible and visual indication (on the dash-mounted display) during an un-signaled lane change or roadway departure. The side-to-side sensitivity of the system can be adjusted between EARLIEST, EARLIER, EARLY, AT EDGE, LATE, LATER, LATEST, and NEVER. A later sensitivity setting will allow the vehicle to depart the lane further before triggering a warning. The default threshold, AT EDGE, will trigger an alert at the moment that the outside tire crosses the lane boundary. Setting the threshold to EARLY, EARLIER, or EARLIEST will result in the triggering of an alert at successively earlier locations, approximately 8 inches for the EARLY setting, 13 inches for the EARLIER setting, and EARLIEST being approximately 1.5 feet before the tire reaches the lane boundary. Setting the threshold to LATE, LATER, or LATEST will result in the triggering of an alert at successively later locations after the tire has passed the lane boundary, approximately 10 inches for the LATE setting, 20 inches for the LATER setting, and approximately 2.5 feet past the boundary for the LATEST setting. Setting the drift alert threshold to NEVER turns drift alerts off. In addition, the unit could be pre-programmed (by the manufacturer or the operator when purchased with the serial communications link option) to an explicit drift alert threshold from 3.25 feet before to 13 feet after the point where the tire crosses the lane boundary. Additionally, the system provides an alertness index (AI) that scores a driver's performance. This score is also shown in the dash-mounted display. The AI factors in how well you maintain the vehicle's position in the lane and how quickly and consistently you react when the vehicle begins to drift towards the edge of the lane. If driving becomes inconsistent or erratic, the score will drop, alerting the driver to pay more attention. The AI ranges from 0 to 99, the higher the AI, the better the driver's steering and the less drowsy they appear to be. Typically, an AI score from 70-90 can be considered good steering performance, while an AI score below 50 may indicate some form of driver impairment (e.g., drowsiness). A visual and audible "Get Rest" indicator will be given if the score drops to an unsafe level; the default setting is 50 for approximately 10 minutes.

SafeTRAC also has an optional fleet management and driver monitoring feature that continuously logs performance data to be retrieved by fleet management. Driver performance data is time stamped and recorded including un-signaled lane changes, roadway departures, alertness index, and vehicle position in the lane. Fleet management can customize the sensitivity, visual and audio indicators, and limit operator access.

A second option with the SafeTRAC system is a video recording feature. Called the "black-box" option, the feature stores 10 seconds of video immediately preceding an accident. It uses an internal accelerometer to determine when an accident occurs, then stores the previous 10 seconds of video in non-volatile memory. In addition to the video, the "black-box" feature, coupled with the fleet management feature, also records data on the vehicle's position in the roadway and any lane departures over several miles prior to the incident. This data can then be downloaded by fleet management for accident reconstruction.

Iteris, Inc. manufactures another lane-guidance system called AutoVue. The system, available for both light- and heavy-duty applications, is a small integrated camera, processor, and software that attaches to the windshield. It uses image recognition software to determine the lane markings. When an un-signaled lane change or roadway departure occurs, the system emits a left or right directional audio signal (similar to the noise from "rumble strips"). The heavy-duty system communicates over the J1939 network to acquire the vehicle speed (it only operates above 35 mph) and outputs a message over the network when an un-signaled lane change or roadway departure occurs. The AutoVue system is currently available on several Freightliner and International vehicles.

Monitoring the behavior of a driver's eye and eyelid is another method for drowsy-driver detection. A metric of drowsiness that has gained popularity, particularly with NHTSA, is PERCLOS. PERCLOS is the percentage of eyelid closure over the pupil in a given period of time, and reflects slow eyelid closures rather than blinks. The PERCLOS metric, established in a 1994 NHTSA study on drowsy-driver detection algorithms, is the portion of time in a minute that the eyes are at least 80 percent closed. The PERCLOS metric provides an accurate measurement of driver drowsiness. However, accurately measuring PERCLOS in a vehicle during operation has proved difficult. Typically, PERCLOS systems use small cameras and an infrared illumination source to pickup and discern a driver's eye. Daytime sunlight can cause the infrared camera signal to become "washed-out." These systems operate best during nighttime. Additionally, driver's glasses and sunglasses can cause interference. Technology has been developing, and NHTSA is currently conducting with the Virginia Tech Transportation Institute a field operational test of a PERCLOS system developed by Attention Technologies.

The Attention Technologies system uses a single camera and two different frequency infrared emitters to capture two images of a pupil. There is an LED display that has three amber LEDs for moderate drowsiness levels, and three red LEDs for severe drowsiness levels. However, the system is still in development and is not yet commercially available.

A third type of driving performance monitoring technology is currently under development at George Washington University that uses steering-wheel monitoring. Algorithms are under development that use artificial neural networks to analyze steering-wheel motion. This type of technology is still under development; no commercially available product is yet available.

VDR Implementation Issues

Integrating these systems with a VDR could be relatively straightforward. The output of a drowsy and distracted driver detection system could be connected to a VDR input, similarly to how other external systems connect (e.g., forward radar and GPS). This process can be simplified if the output is available on a vehicle network databus (e.g., J1939 or J1587). The Iteris AutoVue system for heavy-duty vehicles already transmits a message over the J1939 bus when an un-signaled lane change or roadway departure occurs. Additionally, the AutoVue system has a discrete output that could be connected to a VDR via an analog or digital input.

Technology Implementation Costs

Most drowsy and distracted driver monitoring systems are still in development as this is an emerging technology and pricing data is therefore difficult to obtain. However, Assistware Technology's SafeTRAC is currently available for purchase by fleets and individuals. The base SafeTRAC system, without the fleet management or "black-box" features, costs around $1,500. The fleet management and "black-box" options are customizable, but start around $500 each. The system is customizable to record various amounts of data and can use multiple data extraction protocols at additional costs. Accelerometers with different resolutions also add to the cost.

The Iteris system has been available on Mercedes trucks in Europe for several years and recently has become available on Freightliner Century, Columbia, and Argosy trucks in North America. The system has a list price in Freightliner's data book for $2,492, but can be obtained for between $1,000 and $1,100 with various Freightliner incentives. The retrofit OEM cost is approximately $1,100 to $1,250.

Tractor-Trailer Communications

Technology Overview

There is currently only one electrical interface between highway tractors and trailers, and that is the SAE J560 7-pin connector. The SAE J560 connector was designed to provide a standard connection for trailer power, lighting, and signals. It provides six switch-controlled 12-volt signals (i.e., clearance, stop, left turn, right turn, tail, auxiliary) and a return ground. The J560 connector has no available (i.e., unassigned) pins for data, and the quality of the contacts is only capable of supporting the basic light and power loads and should not be considered for any future data network usage. The J560 connector has a typical lifetime of 6 months to a year because it is subject to physical damage from improper disconnection, stowage, and the possibility of being open to the elements when a tractor is not connected to a trailer.

There are a number of reasons that the trucking industry has resisted installing any new connectors (or a second connector)-most notably, initial cost and maintenance concerns. The industry is at a stalemate in its ability to satisfy the long-term needs for high-quality, reliable data communications between the tractor and trailer.

Other parties (i.e., tractor and trailer OEMs) have strongly resisted most potential solutions primarily on the basis of the increased costs they anticipate. Adding a second connector is viewed as doubling these replacement costs and the downtime resulting from these failures. Adding a backwards compatible connector that would provide a connection suitable for reasonable data networking would add considerable costs, not just for the initial installation, but also for the frequent replacements that are inevitable in the CMV environment.

Experience has shown that any new connector must be implemented with the plan of keeping it in use for a period of 20 to 30 years. The problem that the industry faces in defining a connector is that no one can predict the kinds of communication needs that will exist over that long period of time.

Several possible solutions have been proposed over the past decade, but all have been rejected largely due to cost, backward compatibility, long-term maintenance cost, and future expandability issues. These proposed solutions include:

• Adoption of a second high-speed connection (similar to the European ISO 11898)
• Modification of the pins in the current J560 connector to install a multiplexing network (a 1998 NHTSA study called TRUCKMUX)
• Development of a new backward compatible connector system with additional pins
• A "piggy-back" carrier approach called Power Line Carrier (PLC) that uses the auxiliary power connection and some low-cost chips to communicate
• An infrared approach using the IRDA standards
• A radio-frequency approach

Currently, the PLC approach is being used for illuminating the trailer ABS warning light on the dashboard. The SAE J2497 standard, often referred to as PLC4TRUCKS, superimposes a spread-spectrum frequency carrier on the auxiliary power connection in the J560 connector. The standard uses the same electronic data interchange standard, SAE J1587, as the standard SAE J1708 network, so interoperability is streamlined. The PLC network connection is limited in bandwidth (9600 baud) and dependent on the carrier power signal for reliability. The diagnostic data could possibly be transmitted over the PLC connection to a tractor-based VDR, but the PLC connection is likely not suitable for the high-speed, high-reliability, low-latency communications that would be necessary for advanced safety systems (i.e., rollover control) that would need to interface with the trailer. To date, the PLC solution is the only solution for tractor-trailer communications that has gained acceptance by the industry. It is being implemented on new trailer ABS systems due to regulations requiring that a warning light be lit on the dashboard when there is a problem with the trailer ABS system. To accomplish this, ABS tractor and ABS trailer ECUs are equipped with this PLC transceiver. In the event of a trailer ABS fault, the trailer ABS ECU sends a message to the tractor ABS ECU. The tractor ABS ECU then communicates with the vehicle dash display (either over J1587, J1939, or a discrete signal) to illuminate a fault indicator.

Currently, there is a SAE J1587 message standard (PID 204) under development that will allow the tractor and trailer ABS ECUs to act as a bridge between the J1587 networks and PLC. This message standard will allow ECUs connected to the tractor J1587 network to communicate with ECUs connected on the trailer J1587 network to communicate with each other by passing messages through the ABS ECUs onto the PLC network. This will allow a tractor-based VDR to record data from any electronic ECU located on the trailer.

VDR Implementation Issues

Changes in tractor-trailer communications will have an important impact on the development of VDRs since the amount and type of data to be recorded is directly impacted by this communications link. While the use of a separate trailer-based VDR would solve the above communications issues, it would likely prove too costly, and properly synchronizing tractor and trailer data may be difficult to implement.

In a situation where the PLC connection's bandwidth and reliability are not suitable, the connection of a VDR to a more appropriate connection will be necessary, but it remains to be determined as to which standard or technology the industry will eventually adopt.

Technology Implementation Costs

On new tractors and trailers manufactured after January 1, 2004, PLC communications between tractor and trailer ABS ECUs is standard. Therefore, there would likely be minimal cost associated with PLC-based tractor-trailer communications (specifically with the standards development of PID 204 if the VDR is connected to the tractor J1587 network).

Foot Notes

• International Consultative Committee for Radio (CCIR).
• In 1996, the FCC mandated that all new cellular phones be capable of sending location information to emergency 911 operators by the year 2006. Phase I of the FCC Wireless E911 mandate requires that a dialable number accompany each 911 call and the location of the cellular site that received the call as rough indication of the caller's location. Phase II for the E911 mandate requires that carriers provide automatic location identification using either handset- or network-based technology. Handset technology requires new, modified or upgraded handset with technologies such as GPS to determine location. This location is then broadcast along with the 911 call to alert emergency personnel.
• Qualcomm's wireless communications "messaging link" is a Ku band satellite-based mobile communications system. This system uses geosynchronous satellites orbiting the earth approximately 25,000 miles above the equator at a rate that synchronizes them with the rotation of the earth. Qualcomm on-board systems monitor tractor location using either GPS or QASPER. QASPER is a Qualcomm proprietary satellite-based location determination system similar to GPS, but not typically as accurate as GPS.
• OBD has been required on light-duty passenger vehicles in some form since 1988 and was intended to identify emissions control sensors, actuators, and other emissions control systems which are no longer functioning properly. The most recent light-duty OBD regulations, OBD-II, developed a set of monitors/checks for various emissions related components, as well as, specific diagnostic trouble codes, allowable communications protocols, and a specific data link connector for external diagnostic tools. OBD-II defined the way diagnostic data was made available on light-duty vehicles.
• Society of Automotive Engineers, Glossary of Vehicle Networks for Multiplexing and Data Communications - SAE J1213/1, SAE Vehicle Network for Multiplexing and Data Communications Standards Committee, June 1991.
• Federal Communications Commission, Amendment of Parts 2, 15, and 97 of the Commission's Rules to Permit Use of Radio Frequencies Above 40 GHz for New Radio Applications, Federal Communications Commission, ET Docket No. 94-124 RM-8308, Washington, DC, 1995.
• Mobitex Technology AB, http://www.mobitex.co, SE-417 56 Gothenburg, Sweden.
• Qualcomm Fleet Management Solutions, http://www.qualcomm.com/qwbs/solutions/flt_mgt_overview.shtml , 5775 Morehouse Drive, San Diego, CA 92121.
• Orbcomm, http://www.orbcomm.com, 21700 Atlantic Boulevard , Dulles, VA 20166 .
• Terion, http://www.terion.com, 6505 Windcrest Drive, Suite 200, Plano, TX 75024.
• XATA Corporation, http://www.xata.com/, 151 E. Cliff Road, Suite 10, Burnsville, MN 55337.
• Tripmaster Corporation, http://www.tripmaster.com, 805 Avenue H, Suite 501, Arlington, TX 76011.
• Trimble Navigation Limited , http://www.trimble.com/vt_telvisant.shtml, 749 North Mary Ave, Sunnyvale, CA 94085.
• PeopleNet Communications Corporation, http://www.peoplenetonline.com/, 1107 Hazeltine Blvd, Suite 350, Chaska, MN 55318.
• National Highway Traffic Safety Administration, National Survey of Distracted and Drowsy Driving Attitudes and Behaviors - Volume I Findings Report, The Gallup Organization, Washington, DC, 2003.
• Abrams C., Shultz, T., & Wylie, C.D. Commercial Truck Driver Fatigue, Alertness, and Countermeasures Survey, Federal Highway Administration, FHWA-MC-97-002, Washington, DC, 1997.
• Wierwille et al., Research on vehicle-based driver status/performance monitoring: development, validation, and refinement of algorithms for detection of driver drowsiness, National Highway Traffic Safety Administration, DOT HS80824, Washington, DC, 1994.
• National Highway Transportation Safety Administration, Development, Evaluation, and Demonstration of a Tractor Trailer Intelligent Communication and Power Link. DOT HS 808685, Washington, DC, 1998.