Wheelset states estimation using unscented Kalman filter

Active primary suspension is a comprehensive solution to solve the design conflict between the stability and curving performance for the railway vehicle, and accurately obtain the state of a rail vehicle wheelset is a prerequisite for the implementation of the active guidance control. However, the state of the wheelset is difficult to be directly measured. In this paper, a nonlinear wheelset model, consist of LM wheel tread and R60 rail is established to study the application of Unscented Kalman Filter (UKF) algorithm for lateral wheel/rail displacement and angle of attack estimation of the wheelset. Computer simulation is used to verify the design and to assess its performance. This study reveals that the UKF algorithm can provides accurate estimation of the operating state of the wheelset and has good convergence in consideration of the wheel-rail contact nonlinearity, irregularity, sensor measurement noise and system state noise.