Research Of Urban Expressway Traffic State Estimation And Control

Urban expressway is the road which has all or part of the grade separation andcontrol for the vehicle to driving with high speed. The construction of expresswaycontributes a lot to alleviate urban traffic congestion in a certain extent. With theincreasing amount of transportation, there is traffic congestion phenomenon onexpressway, and more and more serious. To solve the problem traffic congestion onexpressway, it is necessary to estimate the state of expressway traffic. Understandingthe expressway traffic situation can provide effective data-support for expresswaytraffic flow management and control.Expressway traffic state is using traffic flow parameters to measure theapproximate range of certain traffic state. So, estimating traffic flow parameters is oneway to estimate traffic state indirectly. Essentially speaking, traffic state parameterestimation is a nonlinear and non-Gaussian problem. Particle filter has the advantagein dealing with nonlinear and non-Gaussian problem. This paper combining the trafficflow model with particle filter algorithm to establish the traffic state parameterestimation model, and implemented on MATLAB platform for traffic state parameterestimation. The result shows that it is useful to estimate expressway traffic stateparameters with particle filter algorithm, and it has good applicability.In the applying process of particle filter algorithm, there will be a particledegradation phenomenon. The occurrence of particle degradation phenomenon, notonly makes the particle loss the diversity, but also concentrates the most of thecalculation to the particle which has little contribution to the result. In order to avoidparticle degradation phenomenon, this paper uses the immune particle swarmalgorithm to optimize a particle filter and verifies the improved particle filteralgorithm. Compared with basic particle filter algorithm, the result shows that theimproved particle filter algorithm can realize more accurate estimate to traffic stateparameter.Traffic state has the characteristics of fuzziness and uncertainty, and it is suitableto use fuzzy theory to divide traffic state. This paper uses fuzzy c-means clustering method to divide traffic state into three states: smooth state, mild crowded state andcrowded state. According to the result of the traffic state parameter estimation, clusterit to the corresponding state and receive the result of traffic state estimation.Based on the accurate result of traffic state estimation of expressway, achieve theeffective control of expressway mainline and on-ramp, to maximize the expresswayservice ability. Combining variable speed-limit control and ramp coordinated controland considering the characteristics of different traffic states to establish expresswaytraffic control model. This model has been solved with genetic algorithm, and carriesit on the instance validation and application. The result shows that the control strategycan better realize the effective control to the expressway.