Robot in a China Shop: Using Reinforcement Learning for Location-Specific Navigation Behaviour
This work addresses the need for robots to deploy environment-specific behaviors efficiently, though it appears incremental as it builds on existing reinforcement learning and multi-task learning approaches.
The paper tackles the problem of enabling robots to adapt navigation behavior to different environments by treating navigation as a multi-task learning problem, resulting in a 26% reduction in training time and increased accuracy in simulated and real-world evaluations.
Robots need to be able to work in multiple different environments. Even when performing similar tasks, different behaviour should be deployed to best fit the current environment. In this paper, We propose a new approach to navigation, where it is treated as a multi-task learning problem. This enables the robot to learn to behave differently in visual navigation tasks for different environments while also learning shared expertise across environments. We evaluated our approach in both simulated environments as well as real-world data. Our method allows our system to converge with a 26% reduction in training time, while also increasing accuracy.