Spring Validation Demonstration (SVD)
Logistics:
- Location: NSH B506
- Equipment: Master workstation, Client workstation, Platform representing Zamboni (if Zamboni is not available), Intel RealSense D435i, Printed ArUco marker Board
Objective:
- Demonstrate the functionality of the autonomous convoy software stack in simulation,
- Demonstrate the detection and pose estimation algorithm on the physical leader vehicle platform and physical sensors
Demonstration:
- Autonomous Leader Follower Convoy System in Simulation
# | Procedure | Success Criteria |
1 | In the simulation, place the leader Zamboni, attached with a board of ArUco markers, 5 meters in front of the follower Zamboni with all sensors (IMU, encoders, camera, lidar) equipped | – Follower will estimate the relative position of the leader within 0.1 m accuracy |
2 | Launch the leader Zamboni in Gazebo and teleoperate it along a predefined resurfacing path with a maximum velocity of 2.5 m/s | – Follower will localize itself relative to leader with accuracy in x direction of 0.1m, accuracy in y direction of 0.1m and estimate the yaw angle with 1 deg accuracy.- Follower will estimate the velocity of the leader to anaccuracy of 0.2 m/s |
3 | Command the follower Zamboni to start following the leader autonomously and stop the leader Zamboni using keyboard control after moving 2 laps in the ice rink | – Follower will follow the leader and maintain a constant distance of 2m- The system will maintain a lateral offset with the leader of 1.98 m to an accuracy of ±0.15m |
- Leader Follower Autonomy Subsystem Validation on Hardware Platform
# | Procedure | Success Criteria |
1 | – Place a manual-driven vehicle, e.g. a mobility scooter or remotely controlled RC Car, as a leader with an ArUco marker board attached at the rear- Place a Zamboni backup platform (e.g. F1/10 Platform), equipped with the same set of sensors we will use on the follower Zamboni, on the back of the leader with a longitudinal distance of 5 meters and a lateral offset of one body width | – Follower will estimate the relative position of the leader within 0.1 m accuracy |
2 | Start driving the leader with a maximum velocity of 2.5 m/s | – Follower will localize itself relative to leader with accuracy in x direction of 0.1m, accuracy in y direction of 0.1m and estimate the yaw angle with 1 deg accuracy.- Follower will estimate the velocity of the leader to anaccuracy of 0.2 m/s |
3 | The RC is commanded to autonomously follow a leader based on the ArUco marker board | – Follower will follow the leader and maintain a constant distance of 2m |
Fall Validation Demonstration (FVD)
Logistics:
- Location: The FVD will be demonstrated at the Hunt Armory Ice Rink in Shadyside, Pittsburgh or, if the ice rink is unavailable, at a back-up ice rink (Baeirl Ice Complex, UPMC Lemieux Sports Complex), or in a spacious parking lot. To demonstrate some of the functionalities, we will have to drive a Zamboni, or a vehicle of a similar size if the Zamboni is unavailable, in the location specified above. Tests involving only software will be conducted in the MRSD Lab in NSH B506.
- Personnel: We will need one driver to drive the vehicle and others to set up the required system to demonstrate the validation plan. The team members in Team I are sufficient to conduct all the tests. One team member will document the test results, while another operates the leading vehicle. A third team member will remain seated on the follower as a safety driver. In the event that members of the team are not allowed to drive the vehicle, an operator provided by the ice rink will be required for tests involving the Zamboni.
- Equipment: For all tests involving hardware, the vehicle (Zamboni ice resurfacer or ATV) and the operator computer will be used. The start-up procedure involves powering on the units separately, connecting each unit to the dedicated router, and launching processes on the follower platform/robot via the operator computer. Once the start-up processes are launched on the robot/follower zamboni, the visualizer is launched on the operator computer.
Objective:
- Demonstrate drive-by-wire, autonomous leader-following capabilities, and obstacle detection on a Zamboni ice resurfacer.
Demonstration:
# | Procedure | Success Criteria |
---|---|---|
1 | – Place a human-driven Zamboni machine as a leader fitted with ArUco markers; – Place the follower Zamboni equipped with drive-by-wire and follower autonomy capability; – Maintain head-to-tail longitudinal offset of 6.00 m and center-to-center lateral offset of 0.98 m ±0.50 m between leader and follower. | – The follower platform should detect the leader in ≥95% of the frames |
2 | – Launch the perception sub-system, follower autonomy sub-system and the drive-by-wire subsystem on the follower Zamboni machine – Start driving the leader and maintain a velocity of 2.5 m/s – Complete 1 full loop on the ice rink using autonomous Zamboni convoy | – The follower will estimate the relative pose of the leader within 0.1 m accuracy – The follower will estimate leader’s velocity to an accuracy of 0.1 m/s – The follower will localize itself within 0.1 m accuracy – The follower will follow the leader with a head-to-tail longitudinal distance of 6.00 m ±1.00 m – The follower will follow the leader with a center-to-center lateral offset of 0.98 m ±0.50 m |
3 | Repeat steps 1-2 but place an obstacle in the path of the follower after it completes half loop on the ice rink | – The follower will detect the obstacle kept in its path and stop before it hits the obstacle |
Note: In case we cannot get the follower zamboni within the project timeline from the sponsor, we will demonstrate the same autonomous leader following convoy validation on a surrogate platform (Yamaha ATV) on Ice Rink or at Gascola facility depending upon where it is feasible and allowed. We will use a manual driven mid size SUV as a leader if we don’t get a Zamboni machine and a driver on the FVD from the sponsors.