Multi-Robot Collision Avoidance using Reinforcement Learning
ROS Kinetic
Stage-ros
Python 2.7
tensorflow
numpy
Start ROS roscore
Run simple RL controller (will start simulation too): roslaunch collision_avoidance rl_controller.launch
Test a trained model roslaunch collision_avoidance test_trained_model
Start simulation alone: rosrun stage_ros stageros [worldfile]
Teleoperate robot: rosrun collision_avoidance teleop.py
Run super-simple autonomy: python collision_avoidance naive_controller.py