Implementation of extended Kalman Filter on FPGA

The purpose of this paper is to explore the concepts and consequences of implementing the Extended Kalman Filter (EKF) on the FPGA. The methods of Runge-Kutta and Taylor-Heun are applied to approximate the continuous time update. The methods of Rugge-Kutta and Gauss-Legendre are used to solve the Riccati equation in order to update the discrete time measurement. The simulation on Matlab Simulink and implementation on the hardware in-loop are completed. Tradeoff between clock frequency, hardware resources and design accuracy are analyzed in the design. Recommended works describe the limitation of the design and give the suggestion on how to save the hardware resources and increase the accuracy.