Research On Cognitive-Mechanism-based Urban Traffic Regional LOS Evaluation

Urban traffic LOS (Level Of Service) is an indicator measuring the traffic operating conditions as well as the quality of service which the driver and passengers feel. It is an important prerequisite to realize real-time regional LOS evaluation for improving the road network operational performance. Topology is an important characteristic of regional road network, and subjective cognitive is the core of the LOS' definition, so both of the topology and the cognitive mechanism should be necessary consideration in evaluating the regional LOS.In the existing methods of evaluating regional LOS, the regional traffic state was obtained by simple integrating of the basic sections' traffic data. The regional dynamic topology characteristic hasn't been taken into account.In regard to the existing problems, this paper introduces a function of network topology-degree distribution, which belongs to the field of complex network theory, and proposes a regional LOS evaluation method which degree distribution'scaling exponent will be put forward as an input indicator and the ANFIS (Adaptive-Network-based Fuzzy Inference System) model will be applied. Putting road network' average speed, average LOS and the scaling exponent index value as input can comprehensive extract basic sections' traffic flow characteristics and the dynamic topology characteristics of road network. The ANFIS which is based on fuzzy set theory not only can simulate human' cognitive reasoning mechanism, but also set up non-linear approximation functions between the values of input indicators and regional LOS, so that the urban traffic regional LOS can be real-time evaluated easily.This paper discusses the advantages and disadvantages of two regional LOS evaluation methods:ANFIS and subjective-perception. The former is proved to be effective in describing the regional traffic state by the example calculation using real topology and traffic data of Beijing Zhongguancun region. Reasonable and accurate evaluating result can help people understand the regional traffic condition and its trend. Furthermore, it is significant for urban traffic safety management and network planning.