Reasearch On Algorithms Of Tyre-Road Friction Coefficient Estimation And Tyre-Force Distribution For Four-Wheel Independent Driven Electric Vehicles

In recent years, electric vehicles have developed quickly. However, the technologies of electric vehicle need to be improved further. Compared with traditional internal combustion engine, one of the advantages for electric vehicles is that four-wheel independent electric vehicles can be driven directly. Tyre force distribution is one of the key technologies for four-wheel independent driven electric vehicles. To improve the handling stability of four-wheel independent electric vehicle, the algorithms of friction coefficient estimation and tyre force distribution are studied.The friction coefficient is estimated by Extended Kalman filter based on the tire model for combined force generation. The friction coefficient is as the state variable. The longitudinal acceleration, lateral acceleration and yaw rate are as the measurement variables. Jacobian matrix is obtained based on the tire model for combined force generation to make the model linear. Then friction coefficient estimation is obtained by Extended Kalman filter.The controller of improving handling ability is designed based on hierarchical method, In the upper level, the total yaw moment and total lateral force is obtained by Sliding control to track the desired yaw rate and side slip angle. In the lower level, the minimization of sum of every tire utilization square as the cost function. With the restriction of total yaw moment, total lateral force and desired total longitudinal force, the unconstrained optimization problem is gained by multiplier penalty function. And analytical solution of the tyre force is obtained by derivation. If the tyre force is more than the maximum adhesive force, the tyre force is distributed again. And the longitudinal force is transformed into driving torque with the wheel dynamics. And the lateral force is transformed into wheel angle with PI controller.Simulation is conducted with Matlab/Simulink and Carsim to verify effectiveness of algorithms of friction coefficient estimation and driving force distribution. The single lane change and J-turn are selected as simulation conditions with different peak value of steering wheel angle, friction coefficient and initial velocity. Simulation results show tyre-road friction coefficient estimation converges to the real tyre-road friction coefficient. By the multiplier penalty function distributing tyre force, the electric vehicle handling ability is improved.