Abstract
Railway safety is of vital importance, researchers strive to improve the safety of railway vehicles using various approaches. This paper concerns the development of a sign detection system that detects abnormal conditions leading to the flange-climb derailment of a railway vehicle negotiating a sharp curve at a low speed. This research is the continuation of previous research completed at the University of Tokyo. In previous research, a basic detection algorithm was developed using 1/10-scale vehicle experiments. Following the development of a sign detection algorithm from the scaled model, verification of the sign detection algorithm was then completed using real-scale bogie experiments and Multi-body dynamics simulation. This paper focuses on additional improvements completed on the detection algorithm to shorten the detection time. The improved algorithm was verified by using past situational and experimental data published by Japan Transport Safety Board. Verification of the improved algorithm though conducting experiments using a real-scale bogie is also discussed in this paper.
