Data-driven model predictive control using road-based disturbance estimations in longitudinal driving of e-bike

Abstract

In this study, the electric bike is investigated under the data-driven model predictive control (MPC) model approach to develop road disturbance estimation. The focus of the research study is to develop a new control algorithm proposal for cargo e-bikes subjecting to uncertain road conditions in urban transportation. The developed model is proposed for model predictive driving option in e-bikes to provide optimal motor torque operation against predicted road disturbances. The driving profile is designed to simulate urban driving in different road types by considering pedal usage frequency and pedal load measurement in combined drive. Firstly, a datalogger is designed, and the microcontroller is programmed to measure pedal force independently in all road profiles. Road profiles are divided into two groups, pedal-assisted routes and the full electric driving mode, and all range capacities are recorded to make allover comparisons. Road types consist of asphalt and graded gravel at different slope levels. Pedal load statistical relationship based on each route is also indicated, and the road-based difference is also presented. The obtained data are used for data-driven model development in the MPC control model as measured disturbances. The control model then is integrated into the one-dimensional model to provide road estimation. The design of the developed model provides sensorless road disturbance estimation in urban driving. The proposed control model presents an innovative approach for system designing of micro-e-mobility vehicles in optimum longitudinal vehicle control considering the data-driven MPC method.