Page 173 - Toucpad robotics C11

3. DETECT logic:
In sim: check tile color/object API.
°
With OpenCV: color threshold or QR decode; if found mark GPS (grid coords) and alert.
°
D. Step-by-Step Build (Simulator first)
1. Place obstacles/targets in sim map.
2. Write TAKEOFF routine (altitude set or default).
3. Code SEARCH loop:
Move 1 cell (or fixed distance), sense tile/object; log coordinates.
°
Turn at row ends; continue until full coverage or target found.
°
4. Obstacle handling:
When obstacle flag = true skip to next navigable cell; or implement “detour” subroutine (right-hand wall
°
follow for 3 steps resume sweep).
5. DETECTION:
If target found hover, log, beep/print alert; optionally “drop aid” (sim API).
°
6. RETURN_HOME:
Navigate back to start via shortest path (simple reverse trace or Manhattan path).
°
7. LAND and export mission log (time, visited cells, detections).
E. Optional Python Vision Add-On (PC or Raspberry Pi)
1. Print/display markers (QR or colored cards) on the floor/table.
2. Capture frames (webcam/PiCam) while moving a toy drone/robot or by simulating frames.
3. OpenCV flow:
For color marker: convert to HSV threshold contour compute centroid.
°
For QR: pyzbar.decode(frame) get data log.
°
4. Decision: If marker detected emulate “hover/aid drop” and log coordinates/time.
F. Testing Protocol
1. Dry run on a 5×5 grid; verify full coverage (no cells skipped).
2. Inject obstacles in 3 random positions; confirm detour routine works.
3. Hide targets (1–3 positions); measure time-to-detection; verify alert/log.
4. Edge cases: target at border; obstacle near start; no target (mission completes cleanly).
G. Documentation & Deliverables
1. Algorithm & flowchart (state machine + search + detection).
2. Block code/Python script screenshots.
3. Mission log (CSV/table: step, cell, detection flag, note).
4. 2�3 minute video screen-capture of mission run.
5. Reflection sheet: What improved detection? What failed? Next iteration?
H. Extension (for achievers)
Multi-target search with priority zones.
l
Add battery constraint (simulate: must return when steps > threshold).
l
Introduce no-fly zones and altitude changes.
l
Fuse two sensors (e.g., simulated Lidar + camera) and compare accuracy.
l
I. Assessment Checklist (teacher quick rubric)
Robust coverage path ( ) | Correct obstacle behaviour ( ) | Reliable detection & alert () | Clean landing &
l
log ( ) | Documentation clarity ( )

171
Assignments