How To Train Your HERON

TL;DR

This paper demonstrates that a Deep Reinforcement Learning agent trained solely in simulation with domain randomization can successfully navigate real-world lakes and rivers using only a 2D laser scanner, outperforming traditional controllers.

Contribution

The work introduces a zero-shot transfer approach for navigation in natural environments using Deep RL and domain randomization, without real-world training.

Findings

Agent successfully navigates in real environments

Outperforms Model-Predictive-Controller in robustness and speed

Demonstrates effective zero-shot transfer from simulation to real world

Abstract

In this paper we apply Deep Reinforcement Learning (Deep RL) and Domain Randomization to solve a navigation task in a natural environment relying solely on a 2D laser scanner. We train a model-based RL agent in simulation to follow lake and river shores and apply it on a real Unmanned Surface Vehicle in a zero-shot setup. We demonstrate that even though the agent has not been trained in the real world, it can fulfill its task successfully and adapt to changes in the robot's environment and dynamics. Finally, we show that the RL agent is more robust, faster, and more accurate than a state-aware Model-Predictive-Controller.