Implementation of an FPGA-based aided IMU on a low-cost autonomous outdoor robot

Abstract

This paper presents the implementation of real-time algorithms for an aided Inertial Navigation System (INS) on a fabric processor based field programmable gate array (FPGA). Such an approach allows for the development of a hard real-time computational architecture tailored to the specific INS requirements while still preserving flexibility. The development of a realistic simulator proved to be an important stepping stone in the deployment of the INS algorithms in addition to providing useful insight into overall system operation. The simulation supports modular code verification and validation for the embedded target. Specifically, we focus on the integration of an inertial navigation systems (INS) with a man-portable tracked MATILDA robot manufactured by Mesa Robotics. This platform has seen service in both Afghanistan and Iraq, and divides into three pieces for transportation where each piece weighs about 20 pounds.