Local Navigation and Docking of an Autonomous Robot Mower using Reinforcement Learning and Computer Vision

TL;DR

This paper presents a vision-only navigation and docking system for an autonomous mower using YOLO for object detection and Double DQN reinforcement learning, achieving precise docking without external sensors.

Contribution

It introduces a novel combination of YOLO and reinforcement learning for autonomous mower navigation using only a single camera, eliminating the need for GPS or external sensors.

Findings

Achieved centimeter-level docking accuracy from arbitrary positions.

Successfully trained in simulation and validated on real hardware.

System is inexpensive and simple for production.

Abstract

We demonstrate a successful navigation and docking control system for the John Deere Tango autonomous mower, using only a single camera as the input. This vision-only system is of interest because it is inexpensive, simple for production, and requires no external sensing. This is in contrast to existing systems that rely on integrated position sensors and global positioning system (GPS) technologies. To produce our system we combined a state-of-the-art object detection architecture, You Only Look Once (YOLO), with a reinforcement learning (RL) architecture, Double Deep QNetworks (Double DQN). The object detection network identifies features on the mower and passes its output to the RL network, providing it with a low-dimensional representation that enables rapid and robust training. Finally, the RL network learns how to navigate the machine to the desired spot in a custom simulation environment. When tested on mower hardware, the system is able to dock with centimeter-level accuracy from arbitrary initial locations and orientations.