Autonomous Navigation with ROS for a Mobile Robot in Agricultural Fields

Abstract

Autonomous monitoring of agricultural farms and fields has recently become feasible due to continuing advances in robotics technology, but many notable challenges remain. In this paper, we describe the state of ongoing work to create a fully autonomous ground rover platform for monitoring and intervention tasks on modern farms that is built using inexpensive and off the shelf hardware and Robot Operating System (ROS) software so as to be affordable to farmers. The hardware and software architectures used in this rover are described along with challenges and solutions in odometry and localization, object recognition and mapping, and path planning algorithms under the constraints of the current hardware. Results obtained from laboratory and field testing show both the key challenges to be overcome, and the current successes in applying a low-cost rover platform to the task of autonomously navigating the outdoor farming environment.