In this project, we will learn how to design a Deep Reinforcement Learning (DRL) agent to control a quadcopter for various tasks such as take-off, hovering, and landing. This blog will guide you through the installation, downloading the necessary files, developing your project, and submitting it for review.
Table of Contents
Install
This project requires the Robot Operating System (ROS) as the primary communication mechanism between your agent and the simulation. You can either perform a native installation on your own machine or use a Udacity-provided Virtual Machine (VM), which is recommended.
Using the ROS Virtual Machine
To get started with the ROS VM:
- Download the compressed VM disk image and unzip it: RoboVM_V2.1.0.zip
- Ensure that you have either VMWare or VirtualBox installed to run the VM.
Configure Your Virtual Machine
Open your VM player and import the unzipped VM disk image (.ova file). Allocate at least 2 processors and 4GB of RAM to your VM, then launch it and follow the prompts for one-time setup. The credentials are:
- Username: robond
- Password: robo-nd
Direct Native Installation
If you choose to install ROS on your machine, Ubuntu 16.04 LTS is recommended. Follow the instructions for installation here: ROS Installation. Please note that Udacity does not support this method. For assistance, consult ROS Answers.
Download
Start by creating a directory to hold your robot operating system code. Here’s how:
bash$ cd ~
bash$ mkdir catkin_ws
bash$ cd catkin_ws
bash$ mkdir src
bash$ cd src
bash$ git clone https://github.com/udacity/RL-Quadcopter.git
Your folder structure will be as follows:
~catkin_ws
└── src
└── RL-Quadcopter
└── quad_controller_rl
└── ...
This structure is designed to help organize your ROS-based projects. After setting up your directories, install the required Python packages:
bash$ sudo apt-get update
bash$ sudo apt-get -y install python3-pip
bash$ pip3 install -r requirements.txt
Also download the Udacity Quadcopter Simulator (DroneSim) appropriate for your host OS here.
Develop
Now that you have all the prerequisites, it’s time to develop your quadcopter control code.
Start by building your code in ROS:
bash$ cd ~catkin_ws
bash$ catkin_make
bash$ source devel/setup.bash
You will run your project using the following commands:
bash$ cd ~catkin_ws/src/RL-Quadcopter/quad_controller_rl/launch
bash$ roslaunch quad_controller_rl rl_controller.launch
You should see messages indicating the nodes are spinning up. The quadcopter will begin its training episodes. Keep in mind that you’ll need to repeat the build and run steps each time you wish to start the simulation.
Submit Your Project
After completing your project, ensure to document your code and results using the provided notebook. Save or export this notebook as a PDF/HTML report.
For submission, prepare a zip file containing:
- Your project notebook with completed sections.
- The PDF or HTML report.
- The quad_controller_rl Python package including your code implementation.
Troubleshooting
If you encounter issues, like errors in the configurations or networking problems between the VM and the Simulator, try these tips:
- Ensure VM network settings are configured correctly (NAT or Host-only).
- Check if both the Host and VM IP addresses in the ros_settings.txt file match their actual addresses.
- If using VirtualBox, ensure kernel extensions are correctly installed.
For a more technical community discussion, join in with other students on the Udacity Robotics Slack community in the #ros channel.
For more insights, updates, or to collaborate on AI development projects, stay connected with fxis.ai.
At fxis.ai, we believe that such advancements are crucial for the future of AI, as they enable more comprehensive and effective solutions. Our team is continually exploring new methodologies to push the envelope in artificial intelligence, ensuring that our clients benefit from the latest technological innovations.
