Benchmarks

Real measurements where we have them. Schema where we don't. Bad numbers next to good.

§ 01 · Perception

Hardware & software

BoardRaspberry Pi 5 (Model B Rev 1.1), 1 GB
OSDebian Trixie 13.0 · kernel 6.12.75
Python3.13.5
Runtimeonnxruntime 1.26.0 (CPU EP)
ModelYOLOv8n (stock COCO baseline — not a litter model) · ONNX fp32 · 640×640 input
Coolingbare board, fanless, no active cooling
Ambient22 °C
Run date2026-05-13

Burst vs sustained

The two regimes are not commensurate — cold-cache peak FPS is a headroom indicator, not what we will run in the field. Cards side-by-side make that explicit.

BURST · 50 ITER · COLD CACHE 2026-05-13
5.55 FPS peak
p50
180 ms
p95
181 ms
p99
182 ms
regime
50 inferences · cold cache
thermal
run too short to settle

Reference profile only — not what we will run in the field.

ONE-HOUR SOAK 2026-05-13
2.13 FPS sustained
p50
219 ms
p95
226 ms
p99
245 ms
thermal
73.6 °C peak · 0 throttle
memory
140 MiB RSS · flat
drift
p50 minute 1 → 60: +0.05 %
regime
7,649 inferences · no error

Low-end, fanless, no active cooling. No GPU.

Sustained-run notes

  • 1-hour soak, 7,649 inferences, no error, no thrown frames.
  • p50 minute 1 → minute 60: +0.05 % drift. No accumulator pathology observed.
  • Thermal: 73.6 °C peak; firmware throttle threshold is 80 °C. 0 throttle events.
  • Memory: 140 MiB RSS, flat across the run. No leak.
  • Power: USB-C PD 27 W supply (laptop USB-C undervolts and was the bug behind the first run).

These numbers are edge-feasibility evidence: a stock COCO model proving the board can sustain inference. No litter model has been benchmarked on the Pi 5 yet — the litter detector below runs server-side.

Litter detector (server-side)

The detector behind Trash Trail's photo pipeline. v2 (RF-DETR-S, trained 2026-08-02) replaced the v1 YOLO11n baseline after evaluation on a shared internal validation split (826 images / 1,004 boxes):

Metric v1 · YOLO11n v2 · RF-DETR-S
mAP@0.5, internal val (43 classes)0.0740.215 (~3×)
Hard-negative false-positive rate0.0490.024 (halved)
Hazard classes (needles, batteries) recall0.00.0 — unsolved, stated plainly
Training cost8.06 h · one RTX 4070 · 1.613 kWh

Internal validation only — there is no clean external litter benchmark we trust yet, and the 3× bundles architecture and data improvements together. These are relative signals, not headline accuracy claims.

Not-yet-measured

Metric Status Gating
Litter-model FPS / accuracy on Pi 5Edge-sized student model (int8 distillation)
Degraded-view recall (buried / submerged)Field photo collection
End-to-end latency (capture → plan)Chassis assembly + sensor wiring
Thermal envelope at 35 °C ambientSummer enclosure prototype
Power draw at sustained inferenceUSB-C PD meter on bench
§ 02 · Operational safety

The schema is published; first measurements are pending chassis assembly. The schema being visible is the credibility signal even before numbers land.

Human-safety perception

Adult-detection latency
Child-detection latency
Pet-detection latency
Wheelchair-detection latency

Mechanical safety response

E-stop latency
Watchdog cycle
Brake engagement
Thermal cutoff @ 80 °Cimplemented · never fired (soak peaked 73.6 °C)

Hazardous-content handling

Red-list item recall
Operator escalation latency
Hazard-pin coverage on routes