exprob_as2

Assignment 2 for the Experimental Robotics Laboratory course.

Description

4 Aruco markers are placed in a Gazebo world. Under the assumption that around each waypoint there is a marker, the robot is tasked to:

1. Visit a waypoint
2. Find the aruco marker in that waypoint and save its ID
3. After all waypoints have been visited and the 4 markers IDs have been found, go to the waypoint with the marker of lowest ID

Requirements

The following package runs with ROS in Ubuntu 20.04 It requires previous installation of ROSplan (included in the rosplan package), OpenCV libraries, aruco marker models, aruco files (included in the aruco_ros package), and navigation algorithm of move base (included in the planning package).

How to run

Clone this repo, which contains 4 ROS packages and a video folder, in the /src folder of a ROS workspace. Then build your workspace.

In 3 terminals run the following launch files:

1. Gazebo, RViz, Gmapping and Move Base

roslaunch exp_rob_assignment2 test.launch

2. ROSplan nodes, Action interfacees, Aruco detection node and Finder of lowest marker-waypoint node

roslaunch exp_rob_assignment2 test2.launch

3. Call ROSplan services in this order

rosservice call /rosplan_problem_interface/problem_generation_server

rosservice call /rosplan_planner_interface/planning_server

rosservice call /rosplan_parsing_interface/parse_plan

rosservice call /rosplan_plan_dispatcher/dispatch_plan

Note: the decision of splitting the launch files in 2 is because the Move Base launch file throws repeatedly warnings, which makes impossible to visualize other logs in the terminal regarding other nodes.

Solution Approach

First, a .xacro and .gazebo files where created to incorporate to a mobile robot a camera and a laser scanner.

For solving the tasks mentioned above, a domain and problem file were creating with PDDL. 3 durative actions where created:

• goto_waypoint for task 1.
• rotate_and_detect for task 2.
• goto_lowest_marker_wp for task 3.

In addition, for each of the three actions, a .h and .cpp file were created for the action interface of the given action. Aruco marker detection .py script was done implementing OpenCV and Aruco libraries.

Demo: Watch/Download the video

Visualization

The execution of each action can be followed by the terminal, Gazebo and RViz:

1. Terminal: It tells which new marker has been recognized (topic /detected_marker_id ), the marker ID and its associated waypoint (topic /aruco/marker_waypoint_pairs), the dispatched action (topic /rosplan_plan_dispatcher/action_dispatch), and finally the activation of marker detection, goal position, predicate updates in the same terminal where test.launch is launched.
2. RViz: Shows the robot, the camera feed, the highlighted recognized Aruco markers and the robot in the map (add Map from topic) of obstacles, free and unknown space.
3. Gazebo: Shows the Gazebo world with the obstacles, the robot and the aruco markers in position.