Smart Return System: Optimizing Bike Return Experience Using Sensor Data and Trajectory Prediction
The proposed smart return system for shared‑mobility bikes uses accelerometer data and Hidden Markov Model trajectory prediction to distinguish temporary stops from actual parking, enabling 85% of returns to be pre‑judged, improving accuracy, speed, user experience, and operational costs.