Robot in a China Shop: Using Reinforcement Learning for Location-Specific Navigation Behaviour

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.

I. Introduction

Potential applications of modern robotics are becoming wider as the technology evolves. We are giving robots more tasks in more locations and allowing them to face more difficult challenges. Robotics has moved from vacuuming a room to delivering take-out food, from working the assembly line to managing an entire factory. As their functionality grows, so does the variety of their potential work environments. Currently, robots are designed to operate in a single class of work environment. It is feasible to treat a house as a single environment(although this is a simplification). However, it is no longer possible when the environment expands to an entire city. This limitation of the environment should be addressed promptly.