On-Site Reliable Wheel Size Measurement Based on Multisensor Data Fusion

Abstract

The real-time monitoring of wheel size is important technology in ensuring the normal and safe operation of railways. Meanwhile, online dynamic measurement equipment based on structured-light sensors is becoming increasingly available. However, such equipment only has an “artificial” measurement mode, collects minimal data, has low reliability, and cannot be used in an on-site complex environment. In this regard, we propose a measurement system based on multisensor data fusion. When a wheel to be measured passes through a rail, two groups of wheel cross-profile acquisition sensors are arranged. Then, two profiles of the measured wheel are reconstructed, aligned together, and segmentation parameterized for a complete profile. On the one hand, they can compensate each other to overcome the data missing in wheel profile. On the other hand, it improves measurement accuracy by resisting error noise data induces. With the complete wheel cross-profile, the reliable and accurate wheel size parameters are obtained. Furthermore, the feasibility of this method is verified by analyzing possible influencing factors. Real experiments demonstrate the robustness and dynamic measurement accuracy. In conclusion, this measurement system is simple and highly robust for an on-site complex environment and has a promising application value.