Dynamic Simulation of Regenerative Braking in an Urban Electric Vehicle

Abstract

This study integrates an indirect estimate of the regenerative torque-speed curve, obtained from deceleration tests, with the dynamic simulation of an L6(b) urban electric vehicle in Matlab/Simulink. High-resolution GPS records and measured vehicle parameters were used to infer the Te(n) curve, which was implemented as a lookup table in a Newton-based model. On the same trajectory used to construct Te(n), calibration of the equivalent moment of inertia by 2.56 % yielded a correlation coefficient of 0.999756 and a MAPE of 2.050297 %. Because curve construction and model fitting use the same record, these metrics represent calibration and internal verification rather than independent validation. The model provides a basis for future testing with trajectories not used during fitting.