Design, modelling and estimation of a novel modular multi-speed transmission system for electric vehicles

The efficiency of electric vehicles (EVs) should be improved to make them viable, especially in light of the current low energy-storage capacity of electric batteries. Research demonstrates that applying a multi-speed transmission (MST) in an EV can reduce the energy consumption of the vehicle through gear-shifting. However, for effective gear-shifting control in MSTs, first of all, the model of the transmission is required. Moreover, reliable methods should be employed for estimation of the unmeasurable loads and states of the system, under model-based control. This study establishes the mathematical model and estimation algorithms for a novel MST designed for EVs. The main advantages of the designed MST are simplicity and modularity. After devising the dynamics of our proposed transmission, the Kalman filter, the Luenberger obsever and neural networks (NNs) are used to estimate the states, the unknown arbitrary disturbance and the unknown clutch torque applied to the system. Simulation results demonstrate that the proposed approach is suitable for estimation purposes. Experiments were conducted using an in-house prototyped transmission testbed, to validate the simulation results and assess the estimation algorithms.