This paper presents a simulation study of traffic in the Bay Area near San Francisco, CA, USA, to investigate the possibility of using Automatic Traffic Information Systems (ATIS), to produce an accurate, complete and reliable picture of dynamically varying traffic patterns. The simulation used a simultaneous land-use and transportation forecasting model, for simultaneous calculation of land-use patterns, mode choice and distribution. At the same time, an origin-destination matrix for car trips was estimated. The simulated system contained several probe cars, equipped with devices to transmit their locations and speeds. The time, needed to obtain an accurate picture of the actual status of the traffic system, depends on: (1) the number of probe cars; (2) the time intervals between their communications; and (3) the traffic densities on motorways and local roads. Probe cars send data via radio beacons to a central computer. The appropriate location of the beacons depends on the traffic flow densities. Studies are being made to determine the best type of communication. For example, the computer questions each probe car successively in a demand system, and probe cars transmit messages at random times in a supply system.
Abstract