Human-Aware Robot Navigation via Reinforcement Learning with Hindsight Experience Replay and Curriculum Learning
Authors:
- 27 December 2021
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 346-351
Abstract
In recent years, the growing demand for more intelligent service robots is pushing the development of mobile robot navigation algorithms to allow safe and efficient operation in a dense crowd. Reinforcement learning (RL) approaches have shown superior ability in solving sequential decision making problems, and recent work has explored its potential to learn navigation polices in a socially compliant manner. However, the expert demonstration data used in existing methods is usually expensive and difficult to obtain. In this work, we consider the task of training an RL agent without employing the demonstration data, to achieve efficient and collision-free navigation in a crowded environment. To address the sparse reward navigation problem, we propose to incorporate the hindsight experience replay (HER) and curriculum learning (CL) techniques with RL to efficiently learn the optimal navigation policy in the dense crowd. The effectiveness of our method is validated in a simulated crowd-robot coexisting environment. The results demonstrate that our method can effectively learn human-aware navigation without requiring additional demonstration data.Keywords
All Related Versions
Funding Information
- Health
This publication has 12 references indexed in Scilit:
- Robust Deep Reinforcement Learning for Security and Safety in Autonomous Vehicle Systems2018
- Motion Planning Among Dynamic, Decision-Making Agents with Deep Reinforcement Learning2018
- Socially Compliant Navigation Through Raw Depth Inputs with Generative Adversarial Imitation Learning2018
- Passenger discomfort map for autonomous navigation in a robotic wheelchairRobotics and Autonomous Systems, 2018
- Socially aware motion planning with deep reinforcement learning2017
- Decentralized non-communicating multiagent collision avoidance with deep reinforcement learning2017
- Social-aware robot navigation in urban environments2013
- Reciprocal n-Body Collision Avoidance2011
- Unfreezing the robot: Navigation in dense, interacting crowds2010
- Social force model for pedestrian dynamicsPhysical Review E, 1995