Combined heuristic knowledge and limited measurement based fuzzy logic antiskid control for railway applications

In modern railway applications, the prevention of wheel skid is very important. This is because wheel skid can lead to an increase in noise and vibration from wheels with nat points, as well as an increased braking distance. However, conventional antiskid control has problems because the train wheel adhesion and skid characteristics are difficult and time consuming to accurately model. In addition, adequate measured numerical data describing wheel skid is difficult and expensive to obtain from actual railway systems. Therefore, a fuzzy logic based antiskid controller was implemented, where both linguistic and numerical system information could be used. In this paper, the design and implementation of the fuzzy logic controller is described. Results show that the antiskid controller has a very good performance, and performs better than a conventional controller. The described controller is currently running in Mitsubishi Electric railway brake sets in both Japan and overseas.