Maxime

This thesis concerns the control of the effort of a cyclist by an intelligent electric bicycle. The main idea is to use the electrical assistance of an e-bike in order to increase or lower the mechanical load perceived by the cyclist. The effort level is chosen based on the physiological characteristics of the rider formulated in terms of respiratory gas exchange. In order to develop such solution novel modelling, estimation and control strategies were proposed. First, a simulation model including a mechanical, biomechanical and physiological description of the bike and cyclist system was proposed in order generate realistic trajectories for the pedaling force, pedaling speed and respiratory gas exchange variables of the cyclist during exercise. Also, state estimations strategies based on an Explicit Error Bound Set-Membership Observer and on a Robust Proportional Integral Observer were proposed in order to estimation the respiratory gas exchange variables of the cyclist during exercise. Finally, control strategies based on a Proportional Integral controller and on a robust Linear Quadratic Controller for the electrical assistance of the bike were developed in simulation and validated experimentally on an e-bike prototype. Using these new tools, the exercise intensity of a cyclist was regulated at a desired value during an outdoor cycling session in an uncontrolled environment.