Research On The Train Operation Adjustment In Urban Rail Traffic Based On Fuzzy Neural Network

In the process of relieving urban traffic stress, urban rail traffic played a more visible role increasingly. It has advantages that other traffic tools don't have, such as large passenger turnover, safety, rare intervention, fast speed, In normal condition, the train operates strictly according to the planned train operation map. But because of existence of many random interventions, it is inevitable for the train to operate off the panned schedule. Especially in urban rail system conspicuously characterized by short train time internal, usually unpunctual operation of a train will impact normal operation of other trains. Sometimes it will even impact orderly operation of the entire urban rail traffic. Therefore study on operation adjustment of urban rail traffic is very imperative. Employment of intelligent automatic adjustment methods can overcome handicaps of manual adjustment such as big randomness and high requirement for dispatchers and make optimized adjustment projects in time, efficiently, comprehensively.If a number of ways are combined to realize intelligent simulation, the advantage will be very obvious. Neural network and fuzzy system have strong complementarity in manners of knowledge expression and access and computing capability. The common manners of distribution knowledge storage make them have the combination basis theoretically. Therefore this thesis presents a kind of automatic train operation adjustment algorithm based on fuzzy neural network. I finished related study and did some corresponding simulation test. The result of the test turned to be good.The composition of subway trains automatic train supervision (ATS) subsystem is described after a brief description of the development situation of research on train operation adjustment on this article, and the problems, which must be resolved in order to put the train operation adjustment function into practice, are analyzed. Subsequently, the definitions, principles and learning methods of fuzzy neural network are elaborated in detail based on two aspects, neural network and fuzzy control theory. By adjusting the goals and determining the constraint conditions, the fuzzy neural network model of subway train operation adjustment is established. A simulation testing environment is constructed on this article with the MATLAB simulation software according to the data of subway in xi'an city. Combined with run-time information, a simulation test in the case of train being late is implemented according to the base line data as well as the organizing principles of train operation. The simulation results show that the automatic train adjustment algorithm based on fuzzy neural network, which is raised on this article, is viable and effective for train operation adjustment.Since the trains running is affected by many factors which are interrelated to one another, there is no clear criteria with which the quality of the algorithm can be evaluated. Therefore, this paper just initially attempt to adjusting the subway train operation with a method of fuzzy neural network, and which shall have practical significance on the subject research on the automatic train operation adjustment function in ATS subsystem.