Dynamic Hallway Navigator
Advanced AutonomyIntermediate Robotics plus perception pipelines, control loops, coordinate frames, and real-time behavior.
Before you start
- Complete Camera Module and Robot Module
- Understand perception pipelines, coordinate frames, planning, and control loops
- Have an instructor present for every motion run
Build an advanced supervised project that chooses a short path toward a goal while reacting to people and objects in a hallway.
Project Outcome
Navigate between two instructor-approved checkpoints at low speed, stop when the scene is uncertain, and replan when the direct route becomes blocked.
Camera Input
- Object detections and depth estimates
- Hallway boundary or free-space cues
- Robot-relative target and obstacle positions
Build Stages
- Detect free space while the robot remains stationary.
- Convert camera observations into a robot-relative local representation.
- Select a short collision-free target without sending motor commands.
- Test planning and control with recorded data or mocks.
- Add uncertainty, lost-path, and stale-data stop conditions.
- Review limits and cleanup before a supervised low-speed run.
First Prototype
Choose between left, center, and right short-horizon paths and print the selected route when the direct path is blocked.
What To Measure
- Free-space and obstacle-detection accuracy
- Planning latency
- Stop behavior when data is stale or uncertain
- Path smoothness and clearance during supervised runs
Extensions
- Add two sequential checkpoints.
- Compare reactive steering with a small occupancy-grid planner.
- Track moving obstacles across several frames.
- Log selected paths, clearances, and stop reasons.
Safety Boundary
Automated navigation requires instructor supervision, a clear test boundary, a reachable E-stop, conservative speed limits, and fail-closed behavior. Never begin with live motion; validate perception and planning independently first.