Researchers in both academia and industry are increasingly interested in vehicle-to-vehicle (V2V) and vehicle-to-infrastructure (V2I) communication. An article published in Pervasive Computing, IEEE 2006, solicited input from experts; Wieland Holfelder, vice president and chief technology officer of DaimlerChrysler Research and Technology North America, and Jean-Pierre Hubaux, a professor of communication systems at the École Polytechnique Fédérale de Lausanne. In addition, the article includes short articles from Timo Kosch and Markus Strassberger of BMW Group Research and Technology and from Ken Laberteaux, Lorenzo Caminiti, Derek Caveney, and Hideki Hada of Toyota.
Here is a summary of this article.
AN INTERVIEW WITH DAIMLERCHRYSLER’S WIELAND HOLFELDER (Mercedes-Benz & Chrysler)
Vehicle-to-vehicle and vehicle-to-infrastructure communication seems to be an up-and-coming topic. What is this technology’s potential?
Dedicated Short Range Communication, or DSRC, has the potential to significantly reduce accidents and improve traffic flow. … DSRC is the next logical step in vehicle safety, because it lets vehicles communicate with each other and with the infrastructure. It gives the vehicle systems and therefore the drivers a much better awareness of their surroundings so they can avoid dangerous situations altogether.
Developing a new communication system exclusively for automotive use sounds like an expensive and difficult undertaking. Are other technologies such as wireless local area networks a valid alternative?
DSRC is actually based on wireless LAN technology and piggy backs on IEEE 802.11a, which operates in the adjacent frequency band at 5.8 GHz. However, DSRC will require some modifications for automotive use. IEEE 802.11p—also called Wireless Access for the Vehicular Environment (WAVE)—provides the required standardization for these modifications. For example, WAVE standardizes intelligent power management of the DSRC radio to achieve larger distances between vehicles while still being able to scale in a dense environment (such as at a toll plaza with hundreds of vehicles in a small area). WAVE also addresses priority management, so safety messages will always have the highest possible priority. DSRC is unique in its ability to support safety applications, but it will also support a broad range of other vehicle-to-vehicle and vehicle-to- infrastructure communications needs.
What are some of the most promising communication-based safety applications?
One promising example of an infrastructure-to-vehicle communication application is what we call traffic signal violation warning. This application attempts to reduce the number of vehicles running red lights by informing the driver of changes in the signal phase, warning the driver if the vehicle dynamics suggest it will enter the intersection on a red light. This application directly addresses about 2,300 fatalities a year in the US and is the first step in addressing the general problem of intersection safety. An interesting example in the area of vehicle-to-vehicle communication applications is the extended electronic brake-light application. This application would send a message from a hard-braking vehicle to all following vehicles so they can warn their drivers to avoid collisions.
AN INTERVIEW WITH EPFL’S JEAN-PIERRE HUBAUX (Swiss Federal Institute of Technology, Lausanne)
What are the major challenges for vehicular networking?
Vehicular communication is complicated because of the vehicle’s speed, harsh radio channel conditions, and so forth. The interaction between the computers and drivers is very complicated to design, as is the security. …
What will be the most promising vehicular computing applications?
Safety-enabled applications are really the ones that deserve all the attention and justify this effort. You have to remember that in Europe, for example, approximately 60,000 people are killed on the road each year, and 1.5 million people are injured. The most compelling applications relate to sensing and warnings about hazardous conditions, which can help drivers at night, in the fog, or during bad weather in the winter. Besides safety, there is a wealth of other possible applications, notably in e-commerce. They can be useful to bootstrap the whole business of vehicular communications.
Can you talk about interesting research and engineering work in designing and building vehicular networks?
Yes, particularly in networking because you can’t just take an ad hoc protocol and use it for vehicle communication. Geocasting will probably be very useful in such a scenario. The vehicle’s speed can vary from zero to about 150 KM an hour, if not more, so all this has to be considered when designing the network. There are papers that address the physical, MAC, and network layers, but more work is expected for the transport and application layers.
HARNESSING MOBILE NETWORKS FOR DRIVER ASSISTANCE (BMW)
Timo Kosch and Markus Strassberger, BMW Group Research and Technology
In the future, vehicles will feature wireless connections between them and to roadside infrastructures, leading to huge mobile, ad hoc, connected sensor networks. They will be able to harvest data available in the immediate surroundings and from the road ahead that could provide high precision knowledge about relevant traffic situations if interpreted and reasoned upon correctly. … Using wireless communication, vehicles will be able to manage traffic cooperatively in a self-organizing fashion.
Transmitting critical messages
Active and preventive safety applications must be reliable and work correctly under all circumstances. That means that critical messages must be securely transmitted with high reliability and very short latency. Even with IEEE 802.11p and IEEE 1609 standardization under way, from a manufacturer’s perspective, a variety of important questions have not yet been answered in a satisfactory manner— such as those on congestion control or on reliable and trustworthy low-latency transmission in an order of magnitude of 10 to 100 ms. The requirements on the timeliness of the information stems from the necessity that driver-assistance systems only have accordingly small time windows to decide on a driver warning. Considering that vehicles typically travel at speeds of up to 15-20 meters per second in inner city areas, it’s necessary to react within a few milliseconds in order not to lose too much valuable distance in critical situations. With many dozens of vehicles transmitting information in intervals of around 100ms in certain situations (such as at intersections) and the intrinsic hidden station problem, effective congestion control is crucial, too.
Data integrity
Another important area where a viable and feasible solution still hasn’t fully taken shape, despite many promising concepts, is ensuring the transmitted data’s integrity and trustworthiness while protecting the data provider’s (or networked vehicle’s) privacy. We need to ensure full privacy under any circumstances, meaning no one can track our customers, derive their current condition, or disclose their identities. This is difficult, because receivers must be able to fully rely on a received beacon containing another vehicle’s position— intruders shouldn’t be able to fake such messages. At the same time, the sender usually has no interest in disclosing his or her identity, avoiding, for example, road-side network sniffers that can infer route patterns or driving habits.
PERVASIVE VEHICULAR NETWORKS FOR SAFETY (Toyota)
Ken Laberteaux, Lorenzo Caminiti, Derek Caveney, and Hideki Hada, Toyota Technical Center
On an average day in the US, vehicular collisions kill 116 people and injure 7,900. Governments and automotive companies are responding by making the reduction of vehicular fatalities a top priority. In particular, Toyota actively participates in several consortiums that explore safety applications and related technologies. In particular, we’re working on two promising vehicular-safety applications: the Emergency Brake Warning (EBW) and Intersection Violation Warning (IVW) applications.
Emergency Brake Warning
The EBW application alerts the driver when a preceding vehicle performs a severe braking maneuver (see Figure 2 below). Each vehicle running EBW broadcasts a heart-beat message (HBM) at 3 Hz. Each HBM contains the vehicle’s current position, speed, direction, acceleration, and brake-applied status. Each HBM represents these values with short, minimally precise representations, which minimizes channel loading. A vehicle (say, vehicle 1, or the EBW sender) that is braking above a deceleration threshold sends an EBW message at 5 Hz. The EBW message includes a path history (in the form of bread crumbs) over the past several seconds. Furthermore, the EBW contains higher-precision versions of the data elements contained in the HBM. A vehicle that receives the EBW message (such as vehicle 3, or the EBW receiver) must determine if the sender is “relevant” and only warn its driver if the EBW sender lies in the forward path of the EBW receiver (using a proprietary algorithm that uses the bread crumbs of the EBW sender).
Intersection Violation Warning
With the IVW application, a roadside unit, located at a traffic-light intersection, will broadcast 3Hz information regarding the traffic light, including its location, light status, time until color change, intersection dimensions, and so forth. A vehicle (such as vehicle A) approaching the intersection will use its state to predict a 4-second trajectory. Vehicle A compares its predicted trajectory to determine if it will likely be in the intersection during a red light. If so, the driver of the potentially violating vehicle is alerted. In addition, the violating vehicle broadcasts a message with its speed, direction, and so forth, indicating the likely violation. Recipients of this warning message— for example, the traffic light and surrounding vehicles—can then use this warning notification to take appropriate countermeasures (see figure 3).
Challenges
The two applications we’ve described require only road-level (3 to 5 m) accuracy of position measurements. This is currently satisfied using differential GPS with a clear view of the sky. However, long periods of GPS blackout are likely in many driving environments, especially in urban settings. Researchers are developing techniques to withstand GPS blackouts, such as fusing vehicle data (speed, acceleration, steering-wheel angle, and so on) or using terrestrial triangulation techniques, but these approaches aren’t yet mature. Furthermore, lane-level (1 to 3 m) or sublane (less than 1 m) accuracy would be necessary to perform vehicle-to-vehicle collision warnings and other attractive safety applications. Human-machine interfaces pose yet another challenge. If not well presented to a driver, additional safety information could overload or distract the driver, potentially creating a less-safe condition. Other non-safety applications, if not well designed and executed, also have the potential to dangerously distract the driver.
Source:Vehicular Communication [Link]
Farkas, K.; Ittode, L.; Heidemann, J.; T. Kosch; M. Strassberger; K. Laberteaux; L. Caminiti; D. Caveney; H. Hada;
Pervasive Computing, IEEE
Volume 5, Issue 4, Oct.-Dec. 2006 Page(s):55 - 62
Technologies for Road Advanced Cooperative Knowledge Sharing Sensors. TRACKSS, a EU funded project (2007), aims to develop systems for cooperative sensing and prediction of flows, infrastructure and environmental conditions with a view to improve safety and efficiency of road transport operations.
