Visual Navigation in Real-World Indoor Environments Using End-to-End Deep Reinforcement Learning

TL;DR

This paper presents a novel deep reinforcement learning approach for real-world indoor visual navigation, enabling trained policies to be directly deployed on physical robots with high success rates.

Contribution

The authors introduce a new method with visual auxiliary tasks, a tailored reward scheme, and a specialized simulator to transfer DRL policies from simulation to real robots.

Findings

Achieved over 86.7% success rate in real-world navigation tasks

Developed a simulator facilitating domain randomization for real-world deployment

Fine-tuned policies on real images within approximately 30 hours

Abstract

Visual navigation is essential for many applications in robotics, from manipulation, through mobile robotics to automated driving. Deep reinforcement learning (DRL) provides an elegant map-free approach integrating image processing, localization, and planning in one module, which can be trained and therefore optimized for a given environment. However, to date, DRL-based visual navigation was validated exclusively in simulation, where the simulator provides information that is not available in the real world, e.g., the robot's position or image segmentation masks. This precludes the use of the learned policy on a real robot. Therefore, we propose a novel approach that enables a direct deployment of the trained policy on real robots. We have designed visual auxiliary tasks, a tailored reward scheme, and a new powerful simulator to facilitate domain randomization. The policy is fine-tuned on images collected from real-world environments. We have evaluated the method on a mobile robot in a real office environment. The training took ~30 hours on a single GPU. In 30 navigation experiments, the robot reached a 0.3-meter neighborhood of the goal in more than 86.7% of cases. This result makes the proposed method directly applicable to tasks like mobile manipulation.