UAV Path And Trajectory Follower

This repository contains all the code for running this package on ROS. It is composed of two cpp-classes.

Generator is in charge of generating a trajectory based on the ordered lists of 4D waypoints received. The generated trajectory is a much more dense list of 3D waypoints, which can be approximated to the continuous curve. It also generates an interpolated list of the initial times that matches the amount of the more dense list of waypoints. Both interpolated list are necessary for the presented framework to follow a 4D trajectory.
Follower receives the desired trajectory defined as a list of 4D waypoints. It may receive the look-ahead distance and the maximum velocity. Setting these parameters properly is key to get a good performance, depending on the desired 4D trajectory. A much more dense list of waypoints is required to apply the trajectory following method efficiently. Hence, it uses the trajectory generator to get a discrete curve from the ordered list of waypoints. Then, continuously, it receives the UAS pose and the actual time to generate the velocity commands.

Installation Instructions - Ubuntu 16.04 with ROS Kinetic

Follow the instructions for install and build the GRVC UAV abstraction layer on release 3.0.
Clone this repository in your workspace.

$ cd ~/catkin_ws/src
$ git clone https://github.com/hecperleo/upat_follower.git

Install ecl_geometry.

$ sudo apt-get install ros-kinetic-ecl-geometry

Build and source your workspace.

$ cd ~/catkin_ws
$ catkin build

Run tests to check everything. For this step, you need two terminals.

terminal 1: $ roscore
terminal 2: $ catkin run_tests

Note: After this step, you can use just one terminal to run tests without roscore using this command.
$ rostest upat_follower tests_run.tests.

How to use

Running this command in a terminal is enough to launch the simulation.

$ roslaunch upat_follower sim_empty.launch

sim_empty creates a node per UAV that communicates with UAL. This launch is an example of the whole project generating and following a trajectory. By default, you will see one UAV, if you want to see more UAVs you can turn on the flag multi, you also can turn off the flag trajectory to see how this project works with a path instead of a trajectory. If you want to use ROS interface instead of cpp-classes, you can turn off the flag use_class and see how this project works using ROS services and topics.

$ roslaunch upat_follower sim_empty.launch multi:=true trajectory:=false use_class:=false

Note: Check ual_communication to see an example.

C++ class interface

The Follower class is defined in follower.h. You can create one object in your code and use its public methods:

updatePose(const geometry_msgs::PoseStamped &_ual_pose)
prepareTrajectory(nav_msgs::Path &_init_path, std::vector<double> &_times, int _generator_mode = 0, double _look_ahead = 1.0)
preparePath(nav_msgs::Path &_init_path, int _generator_mode = 0, double _look_ahead = 1.0, double _cruising_speed = 1.0)
updateTrajectory(nav_msgs::Path &_new_target_path, nav_msgs::Path &_new_target_vel_path)
updatePath(nav_msgs::Path &_new_target_path)
updateTrajectory(nav_msgs::Path &_new_target_path)
getVelocity()

The Generator class is defined in generator.h. You can create one object in your code and use its public methods:

generateTrajectory(nav_msgs::Path &_init_path, std::vector<double> &_times, int _generator_mode = 0)
generatePath(nav_msgs::Path &_init_path, int _generator_mode = 0)

ROS interface

You can interact with Follower and Generator classes using a ROS interface.

Follower:

PreparePath.srv
PrepareTrajectory.srv
UpdatePath.srv
UpdateTrajectory.srv

Generator:

GeneratePath.srv
GenerateTrajectory.srv

Each service will interact with the corresponding cpp method. Create a client of these services with each corresponding requests and you will be able to interact with it and receive exactly the same response as using the cpp class interface.

Generator and Follower Modes