Wireless Sensor Networks For Traffic Surveillance

University Of California-Berkeley, Berkeley CA

Investigators

Abstract

Title: Wireless sensor networks for traffic surveillance PI: Professor Pravin Varaiya, University of California, Berkeley Abstract: The proposal seeks to design, build and test low-cost wireless sensor networks for vehicle detection in two applications: (1) freeway and urban traffic measurement, and (2) 'smart' intersections to reduce collisions. The network must measure vehicle presence, speed, and it must classify vehicles (passenger cars, vans, buses, trucks). The technical challenges are to design: (1) communication protocols that consume so little power that network nodes can last several years with two AA batteries; (2) signal processing algorithms that can estimate traffic characteristics with accuracies better than and latencies lower than current technologies. Traffic surveillance systems provide the information that is used to improve transportation system productivity through more efficient operations (lower congestion, reduced accident rate) and better information to travelers (reduced travel time). Highway congestion costs $70 and accidents cost $200 billion annually. An extensive and reliable traffic surveillance system can significantly reduce these costs. The proposed research will also be useful in the design of sensor networks for applications in which measurements are made periodically and must be delivered with very strict delay guarantees.