Brake Detection for Electric Bicycles using Inertial Measurement Units

The traffic situation of today's streets is changing through an increase in light electrical vehicular transportation systems. One of those systems is the electronically power assisted cycle (EPAC). As this vehicle group is regarded as strongly exposed, the improvement of cycling safety and the analysis of the rider-behavior is gaining importance. By integrating inertial measurement based brake detection systems, both goals are simultaneously addressed. The approach presented in this paper estimates the brake magnitude of electric bicycles by combining the state-of-the-art brake detection methods and a model-based longitudinal dynamics system to achieve not only a fast response time, but also a reliable detection of persistent braking situations. The experimental results show good accuracy and the influence of further pre-estimated parameters is evaluated through a sensitivity analysis.