Use of Probe Vehicles to Increase Traffic Estimation Accuracy in Brisbane

Traffic congestion is an increasing problem with high costs in financial, social and personal terms. These costs include psychological and physiological stress, aggressivity and fatigue caused by lengthy delays, and increased likelihood of road crashes. Reliable and accurate traffic information is essential for the development of traffic control and management strategies. Traffic information is mostly gathered from in-road vehicle detectors such as induction loops. Traffic Message Chanel (TMC) service is popular service which wirelessly send traffic information to drivers. Traffic probes have been used in many cities to increase traffic information accuracy. A simulation to estimate the number of probe vehicles required to increase the accuracy of traffic information in Brisbane is proposed. A meso level traffic simulator has been developed to facilitate the identification of the optimal number of probe vehicles required to achieve an acceptable level of traffic reporting accuracy. Our approach to determine the optimal number of probe vehicles required to meet quality of service requirements, is to simulate runs with varying numbers of traffic probes. The simulated traffic represents Brisbane’s typical morning traffic. The road maps used in simulation are Brisbane’s TMC maps complete with speed limits and traffic lights. Experimental results show that that the optimal number of probe vehicles required for providing a useful supplement to TMC (induction loop) data lies between 0.5% and 2.5% of vehicles on the road. With less probes than 0.25%, little additional information is provided, while for more probes than 5%, there is only a negligible affect on accuracy for increasingly many probes on the road. Our findings are consistent with on-going research work on traffic probes, and show the effectiveness of using probe vehicles to supplement induction loops for accurate and timely traffic information.