Adaptive Fuzzy Systems for Predictive Diagnostics of Railway Facilities

The problem of the railway infrastructure predictive diagnostics is formulated and reviewed as semi-structured and weekly formalized problem. This paper presents an approach to build evolving fuzzy systems to solve such problems. Phase space granulation and fuzzy adaptation algorithm are suggested. Effectiveness of the proposed approach is demonstrated on adaptations to cascades of bifurcations and nonlinear behavior of the logistic mapping.