LENS 13 — CONSTRAINT ANALYSIS
"What assumptions must hold — and how fragile are they?"

CONSTRAINT MATRIX SUMMARY

Nine constraints govern Annie's navigation system. They fall into three categories: compounding failures, artificial impositions, and physics limits.

WiFi latency is HIGH fragility — uncontrollable. Household RF is shared infrastructure. A microwave three meters away spikes the channel from 15 milliseconds to 300 milliseconds without any visible indicator. This cannot be debugged or patched. Partial relaxation is available: activating the idle Hailo-8 on the Pi 5 as an L1 safety layer moves obstacle detection off WiFi entirely — YOLOv8n runs at 430 FPS locally with under 10 milliseconds of latency.

Single 120-degree camera is LOW fragility — it's artificial. A 15-dollar rear USB camera and an available Pi USB port exist today.

8 gigabytes of VRAM on Panda is MEDIUM fragility. Gemma 4 E2B consumes 4 gigabytes, leaving 4 gigabytes of headroom. Retiring IndicF5 in session 67 bought 2.8 gigabytes. SigLIP 2 for embeddings needs 800 megabytes. Partial relaxation: if Hailo-8 takes over the L1 safety layer, approximately 800 megabytes of Panda VRAM frees up — exactly the SigLIP Phase 2d budget identified in Lens 03.

llama-server API limits are MEDIUM fragility, patchable. The embeddings blocker has a clean workaround via SigLIP 2 as a separate extractor.

The SLAM prerequisite is MEDIUM fragility. Phase 2-a and 2-b run fine without SLAM. Phase 2-c through 2-e are fully blocked.

No wheel encoders is HIGH fragility — hardware constraint. Dead-reckoning drift of 0.65 meters per room-loop was observed in session 92.

Glass and transparent surfaces is HIGH fragility — fundamental physics. Both sensors fail simultaneously. Lidar light passes through glass; camera sees reflection instead of obstacle. No software fix exists.

Motor overshoot on small turns is HIGH fragility — but artificially sustained. 5 degrees commanded produces 37 degrees of actual rotation at motor speed 30. The fix is a one-session firmware task.

Pico I M U stability is HIGH fragility — crash to R E P L is unpredictable, silent, and leaves the system with no graceful degradation.

NARRATIVE

Three constraints form a compounding failure cluster. WiFi latency, Pico I M U stability, and motor overshoot interact in a way that is worse than their individual impacts. When the Pico drops to R E P L, the nav loop falls back to open-loop motor commands — exactly the regime where momentum overshoot is most dangerous, because no I M U correction is available. If WiFi simultaneously spikes, stale commands arrive to a robot already spinning uncontrolled.

The glass surface problem is the most fundamentally hard constraint — and the one most likely to be ignored until it causes a real incident. Every other constraint has a workaround, software fix, or hardware upgrade path. Glass fails both sensors simultaneously.

Two constraints are genuinely artificial. Motor overshoot has a documented fix. The llama-server embedding blocker has a clean workaround via Sig Lip Two.

Technology will relax the V RAM and model-size constraints first. One-billion parameter V L Ms will match today's 2-billion capability within 3 years. The Hailo Eight on the Pi Five partially relaxes two matrix constraints at once: activating it as an L One safety layer moves YOLOv8n detection off WiFi and off Panda's G P U. The I R O S dual-process paper, arXiv 26-01-21506, measured 66 percent latency reduction and 67-point-5 percent navigation success versus 5-point-8-three percent for V L M-only. The relaxation is not free — it introduces Hailo R T and a distinct model compilation pipeline as a new subsystem to maintain.

The matrix reveals that the constraints most amenable to technology relaxation are the ones least urgently in need of fixing, while the constraints most urgently dangerous — WiFi jitter, Pico crash, glass — are the ones technology either cannot fix or requires hardware changes to address.

THINK BOX

Which single constraint removal would make Annie's navigation system qualitatively more capable — not just quantitatively faster or more accurate?

The SLAM prerequisite. Every other constraint improvement is incremental. But SLAM deployment is a phase transition. With SLAM, V L M labels become spatial memories that persist across sessions. Annie can answer "where is the kitchen?" from accumulated observation rather than real-time inference. Without SLAM, Annie is permanently a reactive navigator with no persistent world model.