Technology Rationale:
Maturing ad hoc wireless communications and distributed computing technology enable novel distributed automotive sensing and control systems with compelling applications, such as preventing traffic accidents and improving traffic flow.  Automotive vehicle accidents still account for approximately 40,000 fatalities in the US and are the leading cause of death for people aged 5-45 years.   Accidents are also a primary contributor to congestion, which overall results in 5.7 billion person-hours of wasted travel time annually in the US (2002 estimate). Current technology seeks to address these challenges through in-vehicle sensing, control systems (e.g., electronic vehicle stability control) and road side infrastructure (e.g., dynamic traffic signs, ramp metering).

Significant additional potential can be realized by combining these systems through wireless communications into distributed sensing and control systems. Applications of such systems include cars on collision course warning each other and coordinating an evasive maneuver or cars can communicate their sensor readings to following vehicles to notify them of hazards (e.g., slippery road conditions) and congestion levels. These compelling applications have recently led the US Federal Communications Commission (FCC) to approve radio spectrum for Dedicated Short Range Communications, a networking standard for inter-vehicle and vehicle to roadside communication and innovative uses, from cordless phones and wireless LANs to meter readers and home entertainment products. These diverse services will need to coexist with the emergence of a wide variety of wireless devices, ranging from low bit rate sensors to high resolution full motion video cameras. The combination of increasing data rates and the proliferation of devices could easily lead to inefficiency in the use of unlicensed spectrum due to a combination of overuse and failure to develop mechanisms for efficient sharing of this resource.

Technical Approach:
Realizing this vision requires protocols for highly reliable communication between vehicles in severe fading channels, protocols that scale to very high node densities, and addressing privacy concerns. To this end, we have developed the GeoMAC protocol, conducted scalability evaluations, and designed the virtual trip line concept, which are described below:

GeoMac:
GeoMAC exploits spatial diversity in highly mobile wireless networks. It aims to achieve low latency and high reliability, goals that are intrinsic to the success of many envisioned vehicular safety applications. Conventional MAC layers address reliability through ARQ mechanisms that retransmit messages from the source, if earlier transmissions were not acknowledged. These schemes essentially exploit temporal diversity since retransmissions are only likely to succeed if the channel has improved. GeoMAC exploits spatial diversity, by allowing other nearby nodes to opportunistically forward and retransmit messages. Through a geo-backoff mechanism it uses geographic distance to the destination as a heuristic to select the forwarder most likely to succeed. This mechanism does not require nodes to monitor channel state or position of their neighbors, except for approximate node density, thus enabling their use in highly mobile networks with low coordination overhead. To date we have evaluted the performance of GeoMAC using trace-driven ns2 simulation using packet error measurements from a freeway environment. GeoMac leads to lower delay jitter combined with up to 50% packet delivery rate gains, compared to the AODV and GPSR routing protocols, which also take advantage of nearby nodes for packet forwarding. Spatial diversity is also shown to better utilize available channel opportunities than ARQ mechanisms.

Scalability:
Dedicated Short Range Communications (DSRC)- based communications enable novel automotive safety applications such as an Extended Electronic Brake Light or Intersection Collision Avoidance. These applications require reliable wireless communications even in scenarios with very high vehicle density, where these networks are primarily interference-limited. Given the uncertainties associated with current simulation models, particularly their interference models, it is critical to experimentally validate network performance for such scenarios.

Towards this goal, we present a systematic, large-scale experimental study of packet delivery rates in a dense environment of 802.11 transmitters. We show that even with 100 transmitters in communication range with a frame size of 128 bytes and a bit-rate of 6Mbps, (a) most receivers can decode over 1500 pps in a saturated network, which corresponds to a packet delivery rate of 45% and (b) the mean packet delivery rate, for 10 pps per node workload that emulates vehicular safety applications, is about 95%. These results demonstrate that a COTS 802.11 implementation can correctly decode many packets under collision due to physical layer capture and can serve as a reference scenario for validation of network simulators.

Virtual Trip Lines:

Automotive traffic monitoring using probe vehicles with Global Positioning System receivers promises significant improvements in cost, coverage, and accuracy. Current approaches, however, raise privacy concerns because they require participants to reveal their positions to an external traffic monitoring server. To address this challenge, we propose a system based on virtual trip lines and an associated cloaking technique. Virtual trip lines are geographic markers that indicate where vehicles should provide location updates. 

These markers can be placed to avoid particularly privacy sensitive locations. They also allow aggregating and cloaking several location updates based on trip line identifiers, without knowing the actual geographic locations of these trip lines. Thus they facilitate the design of a distributed architecture, where no single entity has a complete knowledge of probe identities and fine-grained location information. We have implemented the system with GPS smartphone clients and conducted a controlled experiment with 20 phone-equipped drivers circling a highway segment. Results show that even with this low number of probe vehicles, travel time estimates can be provided with less than 15% error, and applying the cloaking techniques reduces travel time estimation accuracy by less than 5% compared to a standard periodic sampling approach.     

Results To Date and Future Work Plan:

We are continuing the development of GeoMAC and plan to test it on the vehicular ORBIT outdoor testbed. We are also designing protocols for steerable antenna array systems on commuter vehicles to efficiently communicate with roadside access points.