Automatic Identification Algorithm Based On Fuzzy Comprehensive Evaluation Of Urban Road Traffic State Study

Since the remarkable economy development results in the eye-catching enlargement of the cities and the increasing number of the private cars all over the cities, the capacity of the urban roads has been decreased sharply, and in consequence brought to traffic congestion. Traffic congestion is a serious problem in almost every metropolitan all over the world. Because of the huge project costs and limited rights-of-way space, it is impractical for the road management and government to count on the new roadway construction or widen the existing road nets to increase the capacity of the whole transportation system. Instead of that, the method that people have begun to pay attention to is on developing Intelligent Transportation Systems (ITS), which are capable of better providing the car drivers or the road users with optimal and efficient travel routes more efficiently. As a crucial part of traffic deduce, traffic parameter could not be separable indeed, which presents how to analyze the traffic parameter and detective the occurrence of automatic incident and congestion automatically. According to the theory of fuzzy comprehensive evaluation relating to Automatic Congestion Identification at both home and abroad, this paper makes a study about how to imply this theory into current situations of China.This thesis studies on the parameters of the traffic flows, the definition and the standard of the traffic congestion firstly. The spatial distribution and the extension of congestion are studied when urban roads have traffic congestion during the peak time. The location of the loop detectors are also analyzed particularly. Then, the interval time of statistic is set. According to the classic algorithms of automatic congestion identification before, also associating with the traffic parameters people always use, the three important parameters(a case in point, traffic stream, average velocity and average time occupancy) which are applied to the fuzzy comprehensive evaluation are of important selected because of the economy, the service life, the cost and so forth. Finally, the paper analyzes the attribute and the reasons that results to the incorrect identification, and proposes the method to improve the previous model based on fuzzy algorithms.First of all, this thesis analyzes the characteristics of the traffic stream distribution on the urban roads changes over time based on time and spatial relativity, which could not match with the expert weight relating to the fuzzy comprehensive evaluation if the urban roads have less traffic when the vehicles are running normally. As a result, a method of giving the maximum value to the sectional average velocity is proposed to cope with the incorrect identification during the sub-period of time. Secondly, owing to the basic data pre-processing selected by different parameters leads to both smoothing effect and countermeasure effect of the processed data with the cluster center, the paper proposes a multi-parameter fusion data to pre-process the basic data in order to modify the cluster center. Objective to compare the influence with the previous one, we use the mean error and the maximum absolute error to assess the data which is based on multi-parameter fusion algorithm is more approximate to the actual data. At last, the real data is simulated experimentally to reform the enhanced model and algorithms that show the fusion algorithm could fit the traffic flow identification from the rush hour to gradual disappearance.