Development of a Fall Detection Algorithm for Powered Two Wheelers Application

Powered Two-Wheelers (PTW) control is more complex than any other motorized vehicle control, in particular during emergency events, such as panic braking or last second swerving. For standard PTW, a common cause of accident in these situations is the loss of stability due to braking maneuvers. It is worth noting that for PTW the loss of stability means a high probability of fall, especially while cornering. Accordingly, the aim of this study is to propose a fall detection algorithm for PTW performing maneuvers leading to potential instability. The algorithm is composed of a number of parameters, named RISKi, able to detect potential fall events, critical for PTW safety. This fall detection methodology was developed to alert an advanced riding assistance system in order to produce proper counteractions against the imminent fall. The parameters designed for the fall detection process take into account the vehicle destabilization due to the braking intensity and due to heavy oscillations of the vehicle body and the steering bar. The algorithm proposed was optimized by virtual data reproducing hard, but safe, braking maneuvers and emergency braking maneuvers leading the PTW to fall. In addition, the parameters were calibrated by experimental data extracted from naturalistic riding runs in urban traffic. The results of the tests demonstrated the applicability of the algorithm in the situations addressed and revealed good prediction accuracy. No false-positive detections were found in the simulations concerning not critical riding conditions and all the braking maneuvers leading the PTW to fall were detected as dangerous events.