Control design for power assisted bicycle trailer

Abstract

Abstract This thesis presents the design and implementation of a controller for an electric trailer attached to a bicycle, aiming to ensure that the cyclist does not experience any additional load arising from the trailer. Our presented controller’s objective is to minimize the force between the bicycle and the trailer, regardless of load variations, wind conditions, or road terrain. The studied trailer’s rigid mechanical construction results in a system that does not display any meaningful dynamics with respect to the applied wheel torque and the resultant force generated at the tow hitch. This static behavior enables the formulation of the controller as a static optimization problem. By applying Newtonian physics, we derive a motion model, leading to the development of a model-based controller capable of real-time control. Additionally, an online parameter estimation algorithm is developed, where the gradient of the model serves as a weight to penalize the relative change of each parameter. By assigning different convergence rates during runtime to the parameters — trailer mass, rolling resistance coefficient, and wind drag coefficient—the estimation process can more accurately reflect real-world conditions. The developed controller was evaluated against conventional control regulators, particularly proportional and integral controllers, using a set of initial parameters. The results are promising, indicating that the online parameter estimation method is effective with an model error of 12% when compared to the true sensor value.