The Michigan VII developmental test environment or DTE for proof of concept (POC) of several applications officially started in November 2007. “However, prior to that, there was almost two years of work going on with the consortium in cooperation with the U.S. Department of Transportation,” Dave Henry, president, VII Consortium & senior manager, Chrysler LLC., noted.
Supported by nine major OEMs (General Motors, Ford, Chrysler, Mercedes-Benz, BMW, Nissan, Volkswagen, Honda, and Toyota), the test fleet consisted of 25 vehicles with four different makes. An On-Board Equipment (OBE) module, a 5.9 GHz Dedicated Short Range Communication (DSRC) radio with Linux drivers and an antenna were integrated in each vehicle with CAN access and a display. Prototype applications developed to exercise the network include probe data collection, off-board navigation, in-vehicle signage and payment applications. Figure 1 shows the VII concept.
“As we were developing the equipment and then integrating the equipment into the vehicles, the US DOT through its contractor Booz Allen, was also creating the RSUs or roadside units,” said Henry. “We now have equipped vehicles as well as what we call the DTE or the developmental test environment.” The DTE is basically a roadway in the Novi, MI area that is equipped with roadside units that cover an area of about 45 square miles and 75 miles of roadway. With the DTE in Michigan and vehicles with the integrated equipment, in November 2007, carmakers were ready to start proof-of-concept testing. Figure 2 shows the location of the RSE units northwest of Detroit.
The executive leadership team, the decision-making body of The VII National Coalition, made up of the VII consortium that represents the automakers, the state and local departments of transportation, as well as the DOT, met in May to discuss the results of the testing to date. With results and a plan to move forward, Henry presented the status of the VII testing at Telematics Update 2008. “I reported that the executive leadership team did agree to continue work on the VII concept, thus passing through the first milestone, which was the viability assessment decision,” he said. While requesting continued investment in VII, the team proposed an accelerated effort in describing the business deployment, security and governance alternatives for VII.
“I think we have kind of proven the technical feasibility, although we are not done, a lot of scalability work has to be done, but I think we are proving that the concept is a viable one to pursue,” said Henry. “One of the things that we have to begin a lot of good work on is the business model and infrastructure governance framework.” Next steps from the technology perspective will be more field operation testing, some of it collaborative, some by individual automakers.
As part of the preliminary conclusions, the cited observations and findings include successfully sending and receiving vehicle “heartbeat” messages among multiple vehicles. While early tests indicate that the system supports transmission of signal phase and timing and intersection maps, message priority testing has not yet been completed. However, the preliminary conclusion is that V2V safety applications can be supported. Also, pending the completion of tests to determine prioritization of safety messages, applications such as cooperative intersection collision avoidance systems can be supported.
Initial findings in specific applications include:
- Heartbeat messages can be exchanged between moving vehicles.
- Safety link messages can be wirelessly communicated to vehicles.
- Probe data can be wirelessly collected by DSRC and can be subscribed to through the network.
- Signage messages can be wirelessly pushed to the vehicles and displayed.
- Large files such as those required for Off Board Navigation can be communicated across the non- contiguous network.
- Electronic payments can be conducted at speed.
VII's leadership team came to the initial conclusion that DSRC wireless communication together with the POC network architecture can serve diverse applications required for safety, mobility and commercial applications. However, some wireless experts believe the protocol used in the current POC testing is inadequate for real-time safety — one of the major applications unique to DSRC.
While DSRC was being developed, a number of parallel efforts were being pursued by communications companies seeking to establish their piece of the navigation services and entertainment pie. “Purely from a technical perspective, the cell phone guys are not there to solve the safety problems,” said Subir Biswas, Associate Professor and the Director of the Networked Embedded and Wireless Systems Laboratory at Michigan State University. “If safety is the main concern with ultrafast message delivery, you need different protocols.” For safety-related systems, avoiding message collisions is key to avoiding vehicle collisions.
Medium Access Control (MAC) protocols generally can be classified in two broad categories. The first is random access protocols. “The Wi-Fi protocol, 802.11, fundamentally is a random access protocol,” said Biswas. With a random access protocol, the main problem is delay or latency. When a vehicle wants to send a message and when it actually sends the message is unpredictable. Message delivery depends on the behavior of others.
The other protocol is Time Division Multiple Access (TDMA), where time slots are allocated so each node gets a chance to communication in an assigned slot. This approach is predictable since they are allocated. In contrast, in the random access protocol, the unbounded delay can occur when all nodes are trying to send in an uncoordinated manner. “It is really chaos that can happen, which is also known as a collision storm,” explained Biswas.
An ideal solution does not exist today, so additional development work is required. In a 2007 IEEE paper, Biswas discusses a Vehicular Self-Organizing MAC (VeSOMAC) with an in-band signaling technique. As shown in Figure 3, in a cooperative collision avoidance (CCA) application, vehicle A sends a message alerting vehicles to slow down. The message should be forward by all vehicles across the platoon with minimum latency. The regular TDMA scheme in Figure 3b uses arbitrary slot allocation and takes three frames to reach vehicle D. In contrast, with the proposed VeSOMAC, the slots are allocated based on the vehicle's relative location, significantly reducing the delivery delay.
The concern goes beyond the academic world. The DSRC uses carrier sense multiple access (CSMA) but the latency of CSMA is undefined. “For a real-time safety system, the latency of that (CSMA) is too great,” said Milt Baker president and co-founder of Automotive Communications Systems, a company that has expertise in wireless communication systems. The company has a patent pending for a TDMA scheme that could solve the problem.
With all the testing performed in the VII test environment, the delay problem could have occurred but apparently did not. “They do not have a lot of cars running, so it is difficult to determine if there will be channel interference,” countered Baker. Simply turning the message rate up will not provide the confidence that the problem will not occur. Since each car has a different back-off time, two cars going faster does not simulate lots of cars.
On a busy highway with six to eight lanes of traffic, theoretically several hundred to a thousand cars could try to communicate with each other if they are all connected to the same access point. If there is an accident or traffic flow stoppage, all vehicles may want to transmit at the same time. Figure 4 shows how an out of range vehicle can receive timely input using ACS's TDMA architecture.
Besides the contention problem, there are two other problems according to Baker. VII wants to be able to resolve the position of each car within one foot. “Without differential GPS, no one knows how to do that today,” said Baker. “With different GPS there is a latency issue there, too, because you have to calculate the GPS equations and the car is moving down the road and you get positional inaccuracy.”
The third issue is the multi-path problem that can occur due to the conventional modulation in today's DSRC radios that use Orthogonal Frequency Divisional Multiplexing (OFDM). Baker noted that the telecommunications industry has essentially gone away from OFDM because it is subject to multipath. A new standard called Coded Orthogonal Frequency Division Multiplex (COFDM) is used in digital video broadcast to the handhelds in Europe, real-time video for surveillance and WiMAX.
ASC believes its patent solves the contention issue. “Once you solve the contention issue, we believe that you can add the one foot position accuracy by using time of flight from the cars and we believe that with the COFDM modulation, those three things together, you have a much better system and to our mind, it's lower in cost than the path the industry is going down today.”
Knowing how many cars are in the ring, strictly defines the latency in a TDMA protocol. The problem is a TDMA system requires a master to determine who is going to talk. Since a master does not exist in the automotive environment, a new approach is required. ACS's solution involves cars self-assigning time slots in a dynamic situation. The time slot is preserved so each car has its unique time to talk and the latency is defined. VII has a latency requirement of 100 ms. Baker insisted, “It looks to us like the system can be architected, so even with 400 cars or however many cars, you can still maintain 100 ms latency and give each car an access to the channel.”
The need for a time-critical bus has already been recognized by automakers for communication in onboard safety systems. CAN is not a deterministic bus, so automakers and tier one suppliers developed FlexRay, a deterministic bus developed initially for safety systems. However, to date the problem has not been observed in the VII testing.
According to Baker, the process to get from patent to POC involves simulation to make sure that they have a sound basis for proceeding. This activity is being pursued. The next step would be to create a field-programmable gate array (FPGA) to build hardware with the software to provide a drop-in replacement for the existing radio. Then comparison testing could be conducted by switching from the current to the proposed approach.
Encouraged by the architecture disclosed in ACS' patent, STMicroelectronics has already established a partnership with ACS. In doing so, STM supports the possibility of resolving some of the major challenges facing the VII initiative such as channel access, reliability of communications, and precision vehicle location at a much lower cost than current solutions. Based on the result of testing with an FPGA version, STMicroelectronics could investigate what it takes for a chip implementation using their standard cell library. Figure 5 shows the circuitry involved to implement the ACS concept, which is explained in greater detail in references[6,7].
With all the effort that VII participants invested in the current 802.11p, the need for a bounded TDMA protocol is certainly an unpopular message — one of contention. Since DSRC is quite far along in their process, the change from 802.11p to TDMA can be considered as a disruptive event, even though it has technical benefit insisted MSU's Biswas. Until there is sufficient testing to verify whether a bounded protocol is required or not, the current activity will continue. With the success experienced so far, if there is a latency problem, it will be eventually be uncovered.
“1-2-3 Red Light!,” Auto Electronics Nov/Dec 2007.
“Vehicle to Vehicle or Vehicle to Roadside Communications?,” Auto Electronics, Nov/Dec 2006.
“Making Vehicles and Highways in Intelligent,” Auto Electronics, Nov/Dec 2005.
David Henry, “VIIC Program Progress,” presentation at Telematics Update 2008, May 21, 2008.
Fan Yu and Subir Biswas, “Self-Configuring TDMA Protocols for Enhancing Vehicle Safety With DSRC Based Vehicle-to-Vehicle Communications,” IEEE Journal on Selected Areas in Communications, vol. 25, No. 8, October 2007.
International Publication No. WO 2007/133264 A2, “Integrated Vehicular Positioning and Integrations Scheme.”
Milt Baker and Lawrence Hill, “A Safety Communications Protocol for V2V and V2I,” Auto Electronics, July/August 2008.
Randy Frank is president of Randy Frank & Associates Ltd., a technical marketing consulting firm based in Scottsdale, AZ. He is an SAE and IEEE Fellow and has been involved in automotive electronics for more than 25 years. He can be reached at [email protected].