Self-tuning of the Virtual-Bike Control for a Human-Powered Electric Bike with Series Architecture

Motivated by environmental awareness, electric bikes (e-Bikes) diffusion as a means of transport has significantly increased in cities, thanks to their low emission and footprint. Among the different alternatives, series-parallel e-Bike architectures are interesting because they merge the advantages of the most common parallel bikes and the series ones, which can be exploited to improve the user experience. When an e-Bike is operating in series mode, a specific control action is needed to handle the absence of a mechanical transmission and so the chain-less nature of series or series-parallel e-Bikes. To this aim, the virtual-chain control law has been proposed and recently extended to a virtual-bike approach, respectively aiming at emulating the experience of the chain or an entire bike, whose parameters are user-chosen. In this work, a self-tuning strategy for the control parameters in the virtual-bike approach is formulated, making it independent of the specific bike and rider. Experimental results showed the advantages and limitations of the proposed solution.