An open-source low-power wheel detector node for urban rail monitoring.

Wheel-passage monitoring is relevant to non-safety-critical railway applications such as temporary traffic observation, maintenance support measurements, and open-hardware field experimentation. This study presents an open-source, low-power wheel detector developed as a monitoring-oriented railway sensing node for operation under real field conditions. The proposed hardware combines a mechanically preloaded pedal, a reed-switch-based triggering arrangement, and an ultra-low-power electronic subsystem built around a prototype control platform. In deep-sleep mode, the prototype consumes approximately 10 µA. The mechanical behavior of the detector was analysed using MATLAB/Simscape, focusing on wheel-pedal contact, preload selection, transient displacement response, and contact-pressure evolution with wheel speed. Preload optimization based on an overshoot criterion showed that 24 mm is the minimum preload required to eliminate secondary penetration. At 90 km/h, the model predicted full damping within 17.076 ms, remaining below the 30 ms digital filtering window and thereby supporting stable single-event generation. A damping sensitivity analysis showed that this timing margin is preserved under reasonable variation of the assumed structural damping. Within the adopted reduced-order 2D plane-strain model with isotropic hardening, the cyclic elastoplastic assessment predicted a shakedown-like response, with residual vertical indentation approaching 0.70 mm, well below the 5 mm triggering allowance of the reed-switch mechanism. A field-installed prototype was evaluated over the 0-90 km/h speed range. A total of 500 monitored wheel-passage events were assessed, with no missed detections and no visible wheel-flange deformation after testing. The results support the feasibility of a low-cost, ultra-low-power, open-source wheel-passage detector for non-safety-critical railway monitoring.