UCLA

Autonomous driving based systems will improve safety and enhance vehicular traffic flow.

A fully-autonomous highway system must make effective use of a reliable and robust communication system. We develop methods for the design of data networking mechanisms that provide for low-latency dissemination of critical messages, as well as enable high system data throughput capacity levels that are used to accommodate the transport of other message flows and sensor data streams.

The data networking mechanisms presented in this dissertation encompass vehicle-to-vehicle (V2V) and/or infrastructure-aided data communication. The latter employs vehicle-to-infrastructure and infrastructure-to-vehicle (V2I) communications. We develop novel networking protocols by considering mobile systems that employ sub-6 GHz spectral resources as well as emerging systems that make use of millimeter wave (mmWave) frequency bands. Data transmissions across sub-6 GHz bands experience lower channel propagation degradations than those incurred across mmWave bands. In turn, mmWave communications channels provide for vastly wider spectral resources, and thus yielding much higher data rates and lower message latencies.

For regions whose transportation networks are not supported by a dense communications infrastructure, we show that an effective use of V2V networking systems can be well realized. In turn, we show that when a proper infrastructure system, which consists of interconnected road side units (RSUs), is available, highly upgraded networking operations can be realized. We show that such a system, when properly designed, can be configured to yield very low critical data message dissemination delays while assuring ultra high throughput rates for other message classes. In setting the system schemes and cross layer parameters to induce desired delay-throughput performance behavior, we examine a multitude of scheduling schemes, and properly set the underlying cross layer parameters, including spatial reuse factors, modulation/coding schemes and data rates, and the underlying transmit power levels. We also involve the following system parameters: antenna gains, vehicular formations and speeds, density of the RSU backbone.

The schemes and techniques presented in the dissertation provide system designers with guidelines, protocols and performance evaluation methods to synthesize a network system for the autonomous highway that will guarantee enhanced data networking performance.