Back to 5.2.4 Padding

5.2.4 Padding — Testing & Evaluation

V-Model Traceability: This page validates RM-6 (Absorb repeated strikes without damage; provide impact detection) through acceptance criteria PAD-AC-1 and PAD-AC-2, plus additional subsystem-level tests designed to ensure reliability under real sparring conditions.

Performance Criteria

The padding subsystem is validated against the following acceptance criteria, cascading from system-level requirement RM-6:

ID Criterion Target Test Method Status
PAD-AC-1 IMU strike detection rate ≥ 95% true positive 60-punch controlled test across three body zones Pending
PAD-AC-2 Impact force differentiation Monotonic: light < medium < heavy IMU L2 vector norm comparison (3 users) Pending
Cross-reference: PAD-AC-1 and PAD-AC-2 also appear on the Arm Actuation Testing & Evaluation page, since they validate the integration between the padding IMU pipeline and the arm's FSM response.

Test Procedures

PAD-AC-1: Strike Detection Rate Test

Equipment: BoxBunny robot (arm stationary at nominal height), padding assembly with 4× MPU6050 IMUs active, GUI V3 with IMU Diagnostics Tab, video camera.
Setup: IMU threshold set to 2.0 g. Robot arm stationary. All four I²C channels verified responsive.
Procedure: Deliver 20 controlled punches per zone (Head, Celiac Plexus, Liver) = 60 total. Wait 2 seconds between each punch. Record each punch event manually (tally) and compare against the /strike_events topic log.
Acceptance: ≥ 57 of 60 punches detected (TP rate ≥ 95%).
Evidence: GUI screenshot showing detection log, confusion matrix (detected vs actual per zone), video recording of test session.

PAD-AC-2: Force Differentiation Test

Equipment: Three users of different skill levels, GUI V3 with L2 norm logging.
Setup: Logging mode enabled in IMU Diagnostics Tab. Centre Body zone selected for consistency.
Procedure: Each user delivers 10 light, 10 medium, 10 heavy punches. Peak L2 norms recorded per strike via the scan-window peak detection algorithm.
Acceptance: Mean L2 norms satisfy mean(Light) < mean(Medium) < mean(Heavy) for all users. Statistical significance confirmed via one-way ANOVA (p < 0.05).
Evidence: Box plot of L2 norms by force category, ANOVA summary table.

Subsystem-Level Robustness Test — Mechanical Endurance

Supplementary test; not a formal acceptance criterion. Validates long-term physical durability of the padding assembly.

Criterion: Padding survives repeated high-velocity strikes without delamination or permanent deformation.
Procedure: 200-punch endurance session, visual inspection after completion.
Acceptance: No visible delamination, foam returns to original shape within 5 seconds.

Subsystem-Level Robustness Test — I²C Communication Reliability

Supplementary test; not a formal acceptance criterion. Validates bus stability under the Hard-STOP recovery protocol during active sparring.

Criterion: IMU communication stable during active sparring with Hard STOP protocol.
Procedure: 5-minute sparring session, monitor for I²C bus hang events via Teensy debug log.
Acceptance: Zero bus hang events.

Results

Pending: Test results will be added after the controlled testing sessions are completed using the Test Validation GUI. This section will include:
  • Measured values and pass/fail verdicts for each criterion
  • GUI screenshots from the IMU Diagnostics Tab
  • L2 norm distribution plots and confusion matrices
  • Video evidence of sparring test sessions