Hardware Architecture of the EKF Prediction Stage applied to mobile robot localization

Abstract

This work presents an FPGA-based Hardware Architecture to implement the Prediction Stage of the Extended Kalman Filter (EKF) applied to the localization problem in mobile robotics. The algorithm has been implemented and run on an Altera Cyclone IV FPGA with a Nios II processor, being adapted and applied to the mobile platform Pioneer 3AT (P3AT). The prediction stage was based on a dead-reckoning system model and its architecture was designed for floating-point representation. In this project the complete EKF was also implemented considering an estimation stage hardware architecture previously developed using a Laser Range Finder (LRF) sensor, producing an overall balanced implementation. Finally, it was evaluated the system performance and suitability, measuring FPGA resources consumption and comparing execution time with a software solution.