Scenario View
Practical Work Flow
- Record telemetry and tag detections.
- Build reference trajectory and tag map offline optimization.
- Relocalize live robot pose against record trajectory.
- Evaluate deviation metrics and visualization outputs.
Results
Organized by real-robot scenarios and simulation validation .
Phase A: Recording
Real-Robot Scenario 1: Recording
Recording / Driving Session
Output 027 - Basic Trajectory
Output 027 - Trajectory with Axes
Additional Scenario Views
Real-Robot Scenario 2: Recording
Recording / Driving Session
Scenario View
Output 020 - Basic Trajectory
Output 020 - Trajectory with Axes
Phase B: Relocalization
Real-Robot Scenario 1: Relocalization
Relocalization Session Video
Additional Scenario Views
Relocalization Point Photos
Relocalization Output Graphs
Real-Robot Scenario 2: Relocalization
Relocalization Session Video
Notes
This scenario includes video evidence and trajectory outputs; no still-point photo set was captured.
Validation in Simulation: Recording + Relocalization
Reference world: limo_ws/src/gz_apriltag_env/worlds/walls_apriltag_limo.sdf
Simulation Environment
Simulation Summary
Offline optimized trajectory, ground-truth comparison, and tag-map consistency were validated on the walls scenario.
Simulation Output Graphs
Simulation Videos
Relocalization Demo 1
Relocalization Demo 2
Key Metrics
| Metric | Value | Comment |
|---|---|---|
| Mean Distance to Path | -- m | Lower is better |
| P95 Distance | -- m | Robustness indicator |
| Mean |Heading Error| | -- deg | Orientation consistency |
Practical Work Example
Important: both real and simulation practical sessions require manual driving via joystick/teleop during the teach run.
Suggested student exercise:
- Record one run in simulation or on robot.
- Generate reference trajectory and map outputs.
- Run online relocalization and capture RViz evidence.
- Report distance and heading metrics with interpretation.
Simulation Annexes
Two annex simulations are provided to support the practical work:
- Annex A - Parking + Control Demo: simulated parking workflow with control integration.
- Annex B - Experiment Validation: same pipeline used to validate recording and relocalization experiments in simulation.
Annex A Commands
cd limo_ws
source /opt/ros/humble/setup.bash
bash profiles/build_sim.sh
bash profiles/launch_sim_parking.shAnnex B Commands
cd limo_ws
source /opt/ros/humble/setup.bash
bash profiles/build_sim.sh
bash profiles/launch_sim_tags_dataset.sh
bash profiles/launch_teleop_joy_sim.shValidation scenario file: limo_ws/src/gz_apriltag_env/worlds/walls_apriltag_limo.sdf
Students who want to study or implement additional control laws can use: PythonRobotics.
These control methods can be connected to the outputs of this pipeline .
Scope Boundaries
In Scope
- Trajectory recording and offline reconstruction
- Online nearest-point relocalization
- Distance and heading deviation metrics
Out of Scope
- Full global navigation in arbitrary worlds
- Dynamic obstacle avoidance stack
- Multi-robot mission planning
Quick Start Commands
cd limo_ws
source /opt/ros/humble/setup.bash
bash profiles/build_real.sh
# or
bash profiles/build_sim.sh