Vision-Based Deep Reinforcement Learning of UAV Autonomous Navigation Using Privileged Information

TL;DR

This paper introduces DPRL, a deep reinforcement learning algorithm for UAV navigation that uses privileged information during training to improve obstacle avoidance and efficiency in complex environments.

Contribution

The paper presents a novel DPRL algorithm combining privileged learning and multi-agent exploration for improved UAV autonomous navigation under partial observability.

Findings

DPRL outperforms existing algorithms in simulation benchmarks.

The use of privileged information enhances perception and robustness.

Multi-agent exploration accelerates training convergence.

Abstract

The capability of UAVs for efficient autonomous navigation and obstacle avoidance in complex and unknown environments is critical for applications in agricultural irrigation, disaster relief and logistics. In this paper, we propose the DPRL (Distributed Privileged Reinforcement Learning) navigation algorithm, an end-to-end policy designed to address the challenge of high-speed autonomous UAV navigation under partially observable environmental conditions. Our approach combines deep reinforcement learning with privileged learning to overcome the impact of observation data corruption caused by partial observability. We leverage an asymmetric Actor-Critic architecture to provide the agent with privileged information during training, which enhances the model's perceptual capabilities. Additionally, we present a multi-agent exploration strategy across diverse environments to accelerate experience collection, which in turn expedites model convergence. We conducted extensive simulations across various scenarios, benchmarking our DPRL algorithm against the state-of-the-art navigation algorithms. The results consistently demonstrate the superior performance of our algorithm in terms of flight efficiency, robustness and overall success rate.