Rover Quest Book - Winter 2025 (W25)
The Great Objective: Autonomous Navigation for the UWRT Rover
The WATO Rover Team is helping build the autonomous navigation system for the UWRT's URC 2025 (opens in a new tab) rover.
Term Objectives Summary
The objectives for Winter 2025 focus on integrating UWRT's existing modules with the WATO Monorepo/watod infrastructure, as well as building out the autonomy software using nav2.
- Point Cloud Generator:
- Create a point cloud using the IntelRealSense D435 cameras and the RealSense SDK, and publish the point cloud message to Nav2 for Costmap Generation, and localization.
- Object Detection:
- Using camera feeds to generate field relative poses for mission-critical objects.
- ArUco Marker Detection:
- Using camera feeds to generate field relative poses for ArUco markers.
- Behaviour Tree:
- Using the Behaviour Tree to plan the navigation of the rover.
- Nav2 Integration:
- Use sensor data from rover to plan and control the rover.
- Use the rover's odometry to plan and control the rover (costmap, localize, plan, and control).
- Simulation:
- Generate rover URDF and TF tree.
- Gazebo simulation for both the rover and desert environment.
Term Objectives and Scoring
- Point Cloud Generator
Score | Criteria |
---|---|
10/10 | Rover is able to reliably generate and publish a point cloud to Nav2. |
5/10 | Rover is able to inreliably generate and publish a point cloud to Nav2. |
0/10 | We were not able to meet the above goals |
Minimum Requirements: Rover is able to inreliably generate and publish a point cloud to Nav2 for a score above 5/10.
- Object Detection
Score | Criteria |
---|---|
10/10 | Rover is able to accurately detect and publish goal pose for Object. |
5/10 | Rover is able to inaccurately detect and publish goal pose for Object. |
0/10 | No progress. |
Minimum Requirements: Rover is able to inaccurately detect and publish goal pose for Object for a score above 5/10.
- ArUco Marker Detection
Score | Criteria |
---|---|
10/10 | Rover is able to accurately detect and publish goal pose for ArUco marker. |
5/10 | Rover is able to inaccurately detect and publish goal pose for ArUco marker. |
0/10 | No progress. |
Minimum Requirements: Rover is able to inaccurately/unreliably detect ArUco marker but not publish goal pose for a score above 5/10.
- Behaviour Tree
Score | Criteria |
---|---|
10/10 | Rover is successfully able to navigate to any goal autonomously. (Object, ArUco, or GNSS Waypoint). |
7/10 | Rover is able to navigate to one of the goals, but is not able to switch between them. |
4/10 | Rover has a basic behaviour tree set up but is not able to navigate to any goal. |
0/10 | No progress. |
Minimum Requirements:: Rover has a basic behaviour tree set up but is not able to navigate to any goal for a score above 4/10.
- Nav2 Integration
Score | Criteria |
---|---|
10/10 | Rover is achieve all mission goals. Costmap, localization, planning, and control all work together and fully implemented. |
7/10 | Rover is able to navigate to all mission goals, but slowly/slightly illegally. Costmap, localization, planning, and control work together but not fully implemented. |
4/10 | Rover is unable to navigate to all mission goals. One more more aspects of Nav2 not implemented. |
0/10 | No progress. |
Minimum Requirements: Rover is unable to navigate to all mission goals. One more more aspects of Nav2 not implemented for a score above 4/10.
- Simulation
Score | Criteria |
---|---|
10/10 | Accurate Gazebo simulation for both the rover and a desert-like environment. |
7/10 | Accurate simulation for the rover but not the environment. |
4/10 | Basic simulation for an example rover with our sensors, but not accurate to our actual rover. |
0/10 | No progress. |
Minimum Requirements: An accurate simulation of the rover for a score above 7/10.
Scoring Template
Quest Name | Description | Due Date | Score |
---|---|---|---|
Point Cloud Generator | Create a point cloud using the IntelRealSense D435 cameras, and publish that to Nav2 for Costmap Generation, and localization. | 2025-03-31 | |
Object Detection | Using camera feeds to generate field relative poses for mission-critical objects. | 2025-03-31 | |
ArUco Marker Detection | Using camera feeds to generate field relative poses for ArUco markers. | 2025-03-31 | |
Behaviour Tree | Using the Behaviour Tree to plan the navigation of the rover. | 2025-04-25 | |
Nav2 Integration | Use sensor data from rover to plan and control the rover. | 2025-04-25 | |
Simulation | Create URDF and TF tree for the rover, get Gazebo simulation running. | 2025-03-16 |