Mobile Robotics Assignment GoTo

Installation

Place the contents of this repo inside $MR_DIR/ws02/src/mr_goto
git clone git@github.com:VincentKen/MR-GoTo.git $MR_DIR/ws02/src/mr_goto

replace ekf_node.cpp and ekf_node.hpp with the ones in MR_EKF also in your ekf.cpp add reset_ = Reset::INTI_POSE; to set_init_pose void EKF::set_init_pose(const Pose2D &p) { if (!init_pose_) init_pose_ = make_shared<Pose2D>(p); else *init_pose_ = p; reset_ = Reset::INTI_POSE; }

same for pf replace particle_filter_node.cpp and particle_filter_node.hpp with the ones in MR_PF also in your particle.cpp add reset_ = Reset::INTI_POSE; to set_init_pose void ParticleFilter::set_init_pose(const Pose2D &p) { if (!init_pose_) init_pose_ = make_shared<Pose2D>(p); else *init_pose_ = p; reset_ = Reset::INTI_POSE; }

tuw::Pose2DPtr ParticleFilter::compute_estimated_pose() { if (!samples_.empty()){ std::sort(samples_.begin(), samples_.end(), Sample::greater); estimated_pose_ = samples_.at(0); } return estimated_pose_; }

Build

Build using colcon build

Run

Run using ros2 run mr_goto mr_goto

Run with map

We can also choose the map file we want mr_goto to run with.

ros2 run mr_goto mr_goto --ros-args -p map_file:=<path_to_map>

or with the launch file:

ros2 launch mr_goto launch.py map:=<name_of_map>

For example:

ros2 launch mr_goto launch.py map:=cave

Note: Per default, we are using the line map.

Launch

You first need to execute

make build-ws02

in the project's root directory.

Launch using

ros2 launch mr_goto launch.py

Launch with map

You can also specify the map you want to use by using an additional map argument.

ros2 launch mr_goto launch.py map:=<map_name>

For example:

ros2 launch mr_goto launch.py map:=cave

Note: Per default, we are using the line map.

Optional ekf or pf

For launching the ekf or pf node, execute

ros2 launch mr_goto launch.py localization:=<node>

For example:

ros2 launch mr_goto launch.py localization:=ekf

Launch with parameter file

Optionally, you can start the launch file with a parameter file by executing:

ros2 launch mr_goto launch.py params_file:=ws02/src/mr_goto/config/params.yaml

The params.yml file is a relative path to your YAML configuration file. A sample params file would look like this:

goto:
  ros__parameters:
    map: cave
    localization: pf

Note: As described in the previous sections, the parameters have their default values set to map:=line and localization:=ekf.

Documentation

Members

• Thomas Khlebovitch (01427030)
• Benjamin Gallauner
• Julius Salamon (51824212)
• Vincent Kenbeek (12229949)

Tasked Completed

• Thomas Khlebovitch

  • Create launch files (80 points)

• Benjamin Gallauner

  • Set target pose (20 points)
  • Drive to target without obstacles (25 points)
  • Drive to target without obstacles and correct the pose at finish (25 points)
  • Make robot avoid small movable obstacle (25 points)
  • Navigate cave obstacle (25 points)

• Julius Salamon

  • published Map(45points)
  • Initialize self-localization and trigger driving using RViz (50 Points)
  • Added parameters to node
  • Adjusted EKF and PF to publish the pose and get the init pose

• Vincent Kenbeek

  • Create a new node (50 points)
  • Plan waypoints to target pose and publish them (50 points)

Create a new node

As can be seen from the image below. We created a new node in ROS. In the beginning this node only had a simple timer callback but was later extended to incorporate the other completed tasks as well.

Self localization using own implementation

We implemented our self localization using our own method. The screenshot below shows our node printing the x coordinate of the robot.

Set target pose using Rviz

The target pose for our robot is set in Rviz as can be seen from the following screenshot.

Driving towards target

After setting the target pose, the robot drives towards the target and stops there as can bee seen from the following sequence of screenshots.

Turning into correct pose at target

Once the robot has reached the target location it will turn into the right pose as can be seen from the following screenshot.

Avoiding a small obstacle

On the way to the target the robot will avoid small obstacles.

Navigating cave obstacle

The following screenshots show that the robot is also capable of navigating cave obstacles.

Plan waypoints and publish them

The node will once a target pose has been set calculate a path to the target goal through breadth-first search. Once the path has been calculated it will publish the waypoints to ROS nav_msgs/Path. The screenshot below shows the waypoints being published and it also gives a visualization of the proposed path.

Publish used map to RViz

Basic launch file

Documented in section Launch.

Initialize self-location and trigger driving using RViz

As shown in the screenshot below, the node listens to /initialpose and /goal so it can be initialized and the driving can be triggered through RViz.

Points distribution

We believe that with the completed tasks we have managed to get 415 points, in addition, we have also used a branching strategy where completed tasks were first push to devel to be tested, this gives us an additional 20 points for a total of 435 points. We believe that the work balance was ditributed equally and so we want these points to be divided between us evenly.

The image below shows that we used a branching strategy for our git