APPLICATIONS OF AUTONOMOUS NAVIGATION IN NEXT-GENERATION MARS ROVERS

Abstract

This paper describes a module used to provide autonomous navigation and obstacle avoidance to a teleoperated prototype Mars rover designed to compete in the 2019 University Rover Challenge. For the competition’s Autonomous Traversal task, the rover must be capable of traversing difficult desert terrain in search of visual waypoints. Our design uses a custom Navigation Board (NavBoard), a mobile robotics computer, and a sensor capable of producing a dense point cloud. NavBoard provides quaternion-based orientation data, distance measurements from a 1D LiDAR system, and GPS data over ethernet to a mobile robotics computer. This computer derives a 3D point cloud from a three-headed collinear stereoscopic camera then processes that data along with the data from NavBoard to determine the correct action to navigate through sparsely mapped terrain.