Modularized and hierarchical state estimation strategy of 4WIDEV based on global-local dynamics model decoupling

This study introduces a hierarchically structured estimation strategy for four-wheel independent drive electric vehicles (4WIDEV), employing global-local dynamic model decoupling to achieve high-fidelity state estimation while avoiding reliance on empirical tire models.An electric drive wheel model (EDWM) was established and an unknown input observation equation was constructed using a proportional integral observer (PIO) method, achieving the decoupling between longitudinal tire force and model state variables. Through the local decoupling of the vehicle dynamics model, observation equations for lateral tire forces acting on the front and rear axles were formulated. An unknown input observer (UIO) was subsequently designed to estimate lateral tire forces, accompanied by rigorous stability proofs for the observer's asymptotic convergence.Then, a hierarchical state estimation strategy was proposed based on the strong tracking extended Kalman filter algorithm (STEKF). Simulation analyses and experimental validations collectively demonstrate that the proposed strategy can effectively improve the estimation accuracy of vehicle states.