Autonomous Navigation and Obstacle Avoidance for Drones

See Wiki for documentation.

Team

• Armand Behroozi <[email protected]>
• Ali de Jong <[email protected]>
• Sean Kirmani <[email protected]>
• Yuriy Minin <[email protected]>
• Josh Minor <[email protected]>
• Taylor Zhao <[email protected]>

Requirements

For stability and compatibility with open-source ROS packages, our system is optimized to work for ROS Kinetic on Ubuntu 16.04 LTS (Xenial Xerus).

It may be possible to build on other versions of ROS for different distributions, but do so at your own risk.

Installation

On Ubuntu 16.04:

git clone https://github.com/kirmani/senior_design.git
cd senior_design
bash setup.sh

This setup script will install ROS Indigo on your machine, upgrade to Gazebo 7 install the dependencies for the ROS Journey package, and build your environment.

Usage

Train Collision Avoidance Model

To run the simulation, open a terminal, and run the following commands:

journey
roslaunch journey simulation_randomization.launch

To start training, open terminal, and run the following:

journey
rosrun journey train_obstacle_avoidance.py

Sending navigation goals in simulation.

To run the simulation, open a terminal, and run the following commands:

journey
roslaunch journey journey_simulator.launch

To set a navigation goal with some delta x, delta y, at a fixed altitude z, just open a new terminal window and run the follow:

journey
ardrone_goal x y z

where x, y, and z are in meters.

Sending navigation goals on a real drone.

To run the simulation, open a terminal, and run the following commands:

journey
roslaunch journey journey.launch

To set a navigation goal with some delta x, delta y, at a fixed altitude z, just open a new terminal window and run the follow:

journey
ardrone_goal x y z

where x, y, and z are in meters.