Robot navigation in crowd

• This project focus on make a simple baseline, minimal configuration, easy to train and test and simulation based on recent researches on crowd-aware robot navigation
• In this project, I use Q-learning with 80 actions (5 level of magnitude and 16 orientation),using RVO2 for generate human interaction based on ORCA algorithms. I use LSTM [3] to encode crowd information.

Result

Training Testing (updating)

How to use

• Training: python train.py
• Testing: python test.py(current test result don't meet the expectation)
• Simulation in Coppelisim: (updating)

Reference

• [1] https://github.com/vita-epfl/CrowdNav
• [2] Chen, Yu Fan, et al. "Decentralized non-communicating multiagent collision avoidance with deep reinforcement learning." 2017 IEEE international conference on robotics and automation (ICRA). IEEE, 2017.
• [3] Everett, Michael, Yu Fan Chen, and Jonathan P. How. "Motion planning among dynamic, decision-making agents with deep reinforcement learning." 2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). IEEE, 2018.
• [4] Chen, Changan, et al. "Crowd-robot interaction: Crowd-aware robot navigation with attention-based deep reinforcement learning." 2019 international conference on robotics and automation (ICRA). IEEE, 2019.