Reduced Infrastructure Cost in Transportation Systems through Intelligent Signal Processing

University Of California-Berkeley, Berkeley CA

Investigators

Abstract

Rapid advances in computing and communicating technologies, combined with inexpensive access to local area wireless networks, have the potential to revolutionize our nation's vehicular networks. The fundamental bottleneck to the timely deployment of these systems is the the infrastructure needed to support them. This proposal seeks to understand infrastructure requirements and explore architectures driven by considerations of cost, scalability, and robustness. In particular, we opportunistically utilize devices such as cellular phones, WiFi networks, and GPS receivers which, though originally deployed for unrelated applications, can significantly enhance the operation of vehicular networks. This proposal focuses on the systems theory needed to realize this, focusing on an interdisciplinary attack around intelligent distributed signal processing, communications, control and networking. The theoretical advances resulting from this work will in turn provide guidance on the design of new protocols for vehicular networks, particularly in rural areas and the developing world, where cost limits the growth of transportation systems. This proposal seeks to develop both the theoretical foundations and novel constructive designs for next-generation intelligent transportation systems that require minimum cost infrastructure. Motivated by the considerations of cost, scalability, robustness, and reliability, the proposed research thrusts include: (i) high-capacity communication based on multihop relaying and random linear network coding for data distribution in vehicular ad hoc networks; (ii) reliable estimation of traffic field information for congestion management based on parsimonious mobile sampling of probe vehicles, and distributed sensing, compressing, and aggregation of data based on accurate traffic-flow models such as METANET; and (iii) robust low-latency estimation for critical safety and collision avoidance based on investigation of concurrent, real-time distributed estimation of multiple dynamical systems over a shared multi-access wireless channel.