HPS Field Test Checklist — Free Humanoid Robot Audit Tool

HPS Field Audit Checklist

Humanoid Performance Score · Six Axes · NIST / ISO / ASTM

Rev 2.0 · Expert Review Edition · 2026

Checklist progress: 0 / 18 items verified

Score: 0 / 100 pts

Mobility — Gait Symmetry Index

GSI · Patterson et al. (2010) · NIST TN 2160 agility framework · Target < 2.5% · 20 pts

20 pts

Stride period simulator — adjust sliders to explore GSI

Left T_L 0.28 s

0.18 s0.40 s

Right T_R 0.29 s

0.18 s0.40 s

GSI result (bar saturates at 5% · exact value shown right)—

Field checklist

Heel-strike events captured at ≥ 120 fps; ≥ 20 valid strides per condition; GRF threshold 20 NGait Posture 31(2)

Zero-fall recovery on 10° incline at nominal gait speedField test

Obstacle clearance > 5 cm threshold confirmed on instrumented walkwayField test

Dexterity — Force Overshoot

FO · ISO/TS 15066:2016 §A.2 · Compliant Cylinder Object · 20 pts

20 pts

Force overshoot — three grasp geometries

F_applied 1.0 N

0.1 N3.0 N

F_min 0.8 N

0.1 N2.0 N

FO ratio (pass < 0.2 · lower = better)—

Field checklist

Standard compliant cylinder — zero permanent deformation after 5 trialsYCB/HPS-prop.

All three grasp geometries completed; median FO reported per geometryField test

Tactile sensor feedback latency < 30 ms (ATI Nano17 6-axis F/T @ ≥ 1 kHz)< 30 ms

Intelligence — Perception Reaction Latency

PRL · ISO/TR 23482-1:2019 §5.4 · Target < 250 ms · 20 pts

20 pts

Stimulus–response test — simulated latency distribution

— ms

PRL latency (target < 250 ms · bar = PRL / 400 ms)

Field checklist

Dynamic obstacle avoidance — path re-plan PRL < 250 ms (median, 20 trials)HPS / ISO ctx

Zero teleoperation confirmed via hardware signal audit (no remote IP traffic)Signal audit

NL command latency: VAD trailing edge → first motor command, median reportedField test

Endurance & Energy — Cost of Transport

COT · Tucker (1975) normalisation · Target < 1.0 on standard surface · 15 pts

15 pts

COT calculator — adjust parameters

Power P 200 W

50 W600 W

Mass m 60 kg

20 kg120 kg

Speed v 1.4 m/s

0.2 m/s3.0 m/s

COT (target < 1.0 · bar: 50% = at target)—

COT vs speed curve — blue line · red dot = operating point

Field checklist

DC-bus power logged at ≥ 1 kHz over ≥ 60 s steady-state window; COT < 1.0COT < 1.0

Thermal stability — all actuator temps < 55°C after 20 min nominal walkThermal audit

Standby power consumption metered and recorded (W)Meter reading

Safety & Compliance — Contact Force & Kinetic Energy

ISO 13482:2014 Annex B · Force, pressure & Ek limits · 15 pts

15 pts

Kinetic energy indicator (secondary · non-normative)

Mass m 60 kg

10 kg120 kg

Velocity v 1.0 m/s

0.1 m/s4.0 m/s

Kinetic energy (Ek) — non-normative indicator only—

Field checklist

Physical E-Stop response time < 100 ms; confirmed by hardware oscilloscope traceISO 13482

Contact force audit (ISO 13482:2014 Annex B): quasi-static head/neck ≤ 25 N, chest ≤ 105 N; transient skull/neck ≤ 130 N, chest ≤ 140 N; ISO 13482 Annex A compliant manikinISO 13482 AnnB

Software geofencing boundaries verified; robot halts at boundary with ≤ 5 cm overshootSignal audit

Adaptability — Versatility Index

Cross-domain success rate · variance-weighted · 10 pts

10 pts

Versatility index calculator

Attempts 10

120

Successes 8

010

Variance σ² 0.15

0.001.00

Versatility index V—

Field checklist

Domain switch test: industrial → domestic; autonomous reconfiguration < 60 s with zero manual interventionField test

Unstructured environment: success rate ≥ 0.7 on defined task battery (10 tasks)Field test

New task acquisition: few-shot (≤ 5 demos) and zero-shot; both reportedDefined battery

Humanoid Press Footer