robotics-rl-navigation-env

A model based RL approach for robot navigation.

Table of contents

• General Info
• Technologies
• Setup

General Info

Navigation RL environment for turtlebot3 and innok heros skidsteer robot. This code is adapted from Open AI Gym and ROBOTIS.

For testing this env, I used DQN [] for robot mapless navigation. The implementation is in Pytorch with Replay Buffer and Target network.

Technologies

UML Class Diagramm

Dependencies:

• ROS noetic

• Gazebo >9.x

• Turtlebot description

• innoks heros description

Setup

Tutorial:

• run gazebo simulation env: turtlebot/innok
• run agent node which use dqn. Can resume from trained episode