无人机自主导航-基于强化学习实现的无人机自主导航-附项目源码-优质项目实战.zip

# Reinforcement Learning for Autonomous navigation of UAVs This repository contains the <b>simulation source code</b> for implementing reinforcement learning aglorithms for autonomous navigation of ardone in indoor environments. <b>Gazebo</b> is the simulated environment that is used here. ## Q-Learning.py Autonomous Navigation of UAV using Q-Learning (Reinforcement Learning). - State: <b>Discrete</b> - Action: <b>Discrete</b> - Action space: <b>5x5 grid space.</b> <b>Indoor Path Planning and Navigation</b> of an Unmanned Aerial Vehicle (UAV) based on <b>PID + Q-Learning algorithm (Reinforcement Learning)</b>. The quadrotor maneuvers towards the goal point, along the uniform grid distribution in the gazebo simulation environment(<strong>discrete action space</strong>) based on the specified reward policy, backed by the simple position based PID controller. <p align= "center"> <img src="Q-Learning/drone_qlearning.gif/"> </p> ### Dependencies & Packages: - <b><a href="http://releases.ubuntu.com/16.04/">Ubuntu 16.04</a></b> - <b><a href="http://wiki.ros.org/kinetic">ROS Kinetic</a></b> - <b><a href="http://gazebosim.org/">Gazebo 7</a></b> - <b><a href="https://github.com/AutonomyLab/ardrone_autonomy">ArDrone Autonomy ROS Package</a></b> - <b><a href="https://gym.openai.com/docs/">gym: 0.9.3</a></b> - <b><a href="https://www.tensorflow.org/install/">TensorFLow 1.1.0 (preferrable with GPU support)</a></b> - <b>Python: 2.7</b> ## DDPG.py Deep Deterministic Policy Gradient algorithm is used for autonomous navigation of UAV from start to goal position. This is applicable for continuous action-space domain. (Under development!)