All defense POCs
POC 17BStability guaranteeProvable today on a workstationWave 2VALIDATED
SolvFilterSolvSRK

SolvFilter — Scalable Multi-Target EKF

140-state Riccati EKF prediction at fleet scale — 74× cheaper than UKF with best NEES consistency. The tracking step the counter-UAV kill chain depends on.

140/140
Riccati survival @ dim=140
74×
Cost vs UKF at dim=140
0/280
scipy baseline survival

The scenario

Set the picture

A multi-target tracker must propagate dozens of coupled target states and full covariance matrices every radar update — under noise, at 10 Hz, on embedded hardware. When the prediction integrator fails or blows up NEES consistency, the track picture lies and the kill chain breaks upstream.

UKF is the usual answer at moderate dimension. At dim=140 it becomes the cost and consistency bottleneck.

What it costs today

UKF at dim=140: accurate but expensive — unsustainable at fleet concurrency on embedded processors.

Standard scipy EKF predict: 0% survival on the SF-3 Riccati slate at dim=140 under the validated noise model.

What changes with SolvNum

SolvFilter is a dedicated EKF prediction engine — not a general ODE solver — optimized for multi-target tracking at scale.

Bbounded prediction step

SolvFilter propagates state and covariance with NEES-consistent statistics at dim=140 where UKF cost dominates. Built on the same noise-survival foundation as SolvSRK for the underlying integration when coupled dynamics are required.

Measurable outcome

What we'll claim — and how it survives review

Each line below maps to a captured number in the demo section. Every number is reproducible from the SolvNum validation suite.

  • SF-3 CONFIRMED: 140/140 Riccati cells survived at dim=140 (20 seeds) vs 0/280 for scipy baselines.
  • 74× lower computational cost than UKF at dim=140 on the validated benchmark.
  • Best NEES consistency among compared filters at scale.
  • SF-4 CLOSED: fast-path + dim=140 approved; IEKF stiff variant parked.
  • Operational multi-target field tapes: POC hardened — partner replay (G-036).

The demo

What was tested. How. What the script printed.

SF-3 validation slate: continuous Riccati EKF at dim=140, 20 seeds, scipy UKF and RK45/Radau baselines for comparison. Measured survival, NEES, and cost ratio.

Live simulation

Animated in-browser simulation of what the demo proves. The numbers underneath are the captured demo output.

Kill chain cycle 1

25ms SolvSRK
📡 DETECT
1
Detect25ms

Radar signal ODE integration (FMCW beat-frequency, pulsed Doppler I/Q)

2
Track33ms
3
Identify850ms
4
Prioritize77ms
5
Assign100μs
6
Engage50ms
7
Assess25ms
Total kill chain cycle1.06s

Full counter-UAV kill chain: detect → track → identify → prioritize → assign → engage → assess. SolvSRK powers the ODE integration at every physics-intensive step. Times shown are per-drone wall-clock from the validated SolvSRK defense-autonomy envelope. The entire cycle completes in under 1.1s per threat — fast enough to close the loop before the swarm’s next maneuver window.

Composes with

Where this POC sits in the substrate

Every POC reinforces — and is reinforced by — others. Click through to see how each piece locks into the larger picture.

Evidence pointers

Where the claims live in the repo

These are the files a reviewer should run, read, or grep to re-derive every number on this page.

  • solv/products/solvfilter/overview.md
  • solv/17-defense-validation-gaps.md (G-036)
  • SF-3 triage CONFIRMED 2026-05-30

Want to see this in your environment?

Brief us on a program where this POC matters.

ITAR-aware. Air-gapped delivery available. Every claim above traces back to a script in the public repo.

Brief us