LAUSR

Abstract The distance that an electrically powered bicycle can cover depends on factors such as the route’s elevation profile, the motor support selected, and the fitness of the cyclist. This fact requires the cyclist to estimate which motor support may be chosen to reach the goal with the battery’s current state of charge. For this reason, we propose a battery management control system based on a nonlinear model predictive controller (NMPC) for pedal-electric drive units (Pedelecs) that takes into account route information and cyclist fatigue. The goal is to guarantee a user-defined state of charge (SoC) at the end of the route while minimizing cyclist fatigue. The degree of support of the Pedelec is considered as the manipulated variable. In order to find an optimal level of assistance, the NMPC minimizes a quadratic cost function that is subject to three nonlinear distance-dependent models. The first two models describe the bicycle dynamics and the discharge behavior of the battery. To obtain an estimate of the maximum voluntary force that the cyclist can apply, the third model describes the cyclist’s fatigue. The identified models and the control strategy are validated with a trekking Pedelec on a 33 km test track. The proposed NMPC is able to guarantee a predefined target SoC at the end of the track while keeping the estimated cyclist’s fatigue low.