FPGA Implementation of Extended Kalman Filter for Parameters Estimation of Railway Wheelset

Abstract

It is necessary to know the status of adhesion conditions between wheel and rail for efficient accelerating and decelerating of railroad vehicle. The proper estimation of adhesion conditions and their real-time implementation is considered a challenge for scholars. In this paper, the development of simulation model of extended Kalman filter (EKF) in MATLAB/Simulink is presented to estimate various railway wheelset parameters in different contact conditions of track. Due to concurrent in nature, the Xilinx® System-on-Chip Zynq Field Programmable Gate Array (FPGA) device is chosen to check the onboard estimation of wheel-rail interaction parameters by using the National Instruments (NI) myRIO® development board. The NI myRIO® development board is flexible to deal with nonlinearities, uncertain changes, and fast-changing dynamics in real-time occurring in wheel-rail contact conditions during vehicle operation. The simulated dataset of the railway nonlinear wheelset model is tested on FPGA-based EKF with different track conditions and with accelerating and decelerating operations of the vehicle. The proposed model-based estimation of railway wheelset parameters is synthesized on FPGA and its simulation is carried out for functional verification on FPGA. The obtained simulation results are aligned with the simulation results obtained through MATLAB. To the best of our knowledge, this is the first time study that presents the implementation of a model-based estimation of railway wheelset parameters on FPGA and its functional verification. The functional behavior of the FPGA-based estimator shows that these results are the addition of current knowledge in the field of the railway.