The doctrinal task is not to invent new hardware. It is to compose existing capability layers into a stack that survives a given mission.
The DA Survivability Engine turns counter-UAS doctrine into a quantified, mission-optimized deliverable. It accepts a mission profile, threat envelope, infrastructure, and constraint set; it returns the stack composition that maximizes projected survivability under those constraints, with confidence-bounded probability terms continuously refined from operator-validated combat data nobody else holds structured access to.
The Math §5.1
P_survive(M) = ∏i∈T [ 1 − ∏j∈S (1 − p_detectij · p_classifyij · p_engageij · p_killij) ]N_i
Every variable named, every value source identified, objective and constraints separated. Mission-specific optimization is a procedure, not a judgment call.
The Moat §7
The framework can be replicated. The data backbone cannot. RAND, IDA, MITRE publish on multi-year cycles; the Ukrainian operator iteration cycle is measured in weeks. Each customer engagement refines friendly-side terms; each operator engagement refines threat-side terms. The flywheel is unique to entities holding both ends.
The Wedge §8
Three engagement pathways already monetized: worked-example mission advisory for U.S. defense customers under existing DA contract authorities (NSWC CSO, SOFWERX, OT); operator-validated capability-feedback subscription for product companies and primes; hardware channel via NVECTECH thermal distribution. Cash-engine funds the platform layer to maturity.
Public Institutional Admission §1.1
| 28 Aug 2025 | JCO disestablished; replaced by JIATF 401 reporting to Deputy SecDef ($50M/effort approval). |
| 27 Sep 2024 | Replicator 2 directed: SecDef memo, second line of effort, counter-small-UAS focus. |
| 31 Mar 2025 | CRS R48477: FY2025 NDAA §1090 directs SecDef formal C-UAS strategy. |
| Sep 2025 | CNAS Countering the Swarm: U.S. lacks sufficient purpose-built counter-drone systems and SHORAD. |
| Apr 2026 | USSOCOM Capabilities of Interest V26-3 IO 5.0 / Mob 12.0–12.1: counter-UAS named requirements. |
| FY26/27 | NSWC CSO H92240-26-S-C001 §1.8.3.5 names "fiber optic"; §1.8.3.12 names "swarm + diversion decoys". |
What This Prototype Demonstrates
Five tabs, each anchored to a section of the DA Counter-UAS Survivability Proposal v1.0:
- Mission Composer — input panel: threat envelope T, infrastructure I, constraints (B_acq, T_max, W_max, S_eligible).
- Survivability Engine — live computation of P_survive(M) with the full p_detect · p_classify · p_engage · p_kill matrix exposed.
- Recommended Stack — argmax_S P_survive(M) under constraints; residual gap list; constraint-binding analysis.
- Tactical Demonstration — animated worked example: SOCOM small team facing the JCO Yuma 50+ saturation profile.
- Data Backbone — anonymized source catalog, threat catalog, equipment performance, TTPs — every value in §5 traces back here.
Press ▶ Run Guided Walkthrough in the nav for a paced tour, or jump in via Mission Composer.
Mission Composer
Proposal §6.1 – §6.3 · Survivability Function Inputs (M, T, E, I, R, constraints) Specify the mission profile, threat envelope, environment, infrastructure, and binding constraints. The engine returns the doctrinally-correct stack composition.Mission Profile (M) M
SOCOM small team conducting fixed-site overwatch of partner-nation FOB. Threat assessed to include JCO-Yuma-class saturation engagement (50+ simultaneous, mixed Group 1–2 + decoys + fiber-optic).
Threat Envelope (T) T_i, N_i
Per Intelligence Stack §3 · expected count N_i per engagement window
Constraint Set §5.2
$7.5M
3
25 kg
Honest framing: PRC-origin (e.g. TATUSKY TTSKL03) excluded by NDAA/TAA toggle. Non-productized (e.g. UK channel sonar v1) excluded unless explicitly allowed.
Inputs valid · 4 threats in envelope · ready
Survivability Engine
Proposal §5.1 — Mission Survivability Function Live computation of P_survive(M) over the threat envelope and capability stack. Probability terms initialized from §4 equipment intelligence and §5 TTPs; refined from operator-engagement records via Bayesian update.
Psurvive(M) = ∏i ∈ T [ 1 − ∏j ∈ S ( 1 − p_detectij · p_classifyij · p_engageij · p_killij ) ]N_i
i threat classes in envelope T
j defensive layers in stack S
N_i expected count of threat-i per engagement window
p_* conditional probabilities · §4 capability data
Active Stack S 0 layers
Probability Matrix · pij = p_detect · p_classify · p_engage · p_kill layer-vs-threat
p ≥ 0.50 (high)
0.20 ≤ p < 0.50 (med)
0 < p < 0.20 (low)
p = 0 (modality blind)
in recommended stack S*
P_survive (Baseline · Existing I only)
—
existing infrastructure I unsupplemented
P_survive (Recommended S*)
—
argmax over feasible S subject to constraint set
Δ Survivability
—
absolute improvement attributable to S*
Per-Threat Survivability Breakdown
◐ Operator Data Flywheel · Live Bayesian Refinement
Each probability term is initialized from manufacturer-claimed performance and refined per documented engagement. The system's recommendation quality compounds with use.
Initialized from OEM-disclosed (handheld RF vs T-014 fiber-optic FPV):
p_detect[D-RF6, T-014] = 0.05 → after OP-A network corroboration (n=18 engagements, May 2026):
p_detect[D-RF6, T-014] = 0.00 (modality-blind confirmed)
Initialized from manufacturer claim (mesh-net pistol vs T-014b RPG-armed FPV):
p_kill[F-NET, T-014b] = 0.45 → after OP-M direct combat observation (warhead fires forward on intercept):
p_kill[F-NET, T-014b] = 0.05 (operator-rejected; bunker doctrine substituted)
p_detect[D-RF6, T-014] = 0.05 → after OP-A network corroboration (n=18 engagements, May 2026):
p_detect[D-RF6, T-014] = 0.00 (modality-blind confirmed)
Initialized from manufacturer claim (mesh-net pistol vs T-014b RPG-armed FPV):
p_kill[F-NET, T-014b] = 0.45 → after OP-M direct combat observation (warhead fires forward on intercept):
p_kill[F-NET, T-014b] = 0.05 (operator-rejected; bunker doctrine substituted)
Recommended Stack S*
Proposal §5.4 · §6.4 — argmax_S P_survive(M) subject to constraints Output of the optimization. Each layer addition is justified by its marginal contribution to P_survive against the threat envelope, traced to source intelligence in the Data Backbone.Δ Survivability vs Baseline
Layer Composition
Residual Gap List P_defeat < threshold
Constraint-Binding Analysis §5.2
Binding constraints tell the customer what additional resource availability would change the recommendation. Slack constraints indicate where additional spend is wasted at this composition tier.
What the System Returns to the Customer §5.4
1. Recommended stack S* as a list of capability-class additions to existing infrastructure I (above).
2. Projected survivability number P_survive(M | S*) with confidence bounds — —.
3. Enumerated residual gap list — threat classes for which P_defeat is below threshold despite additions (right column).
4. Delta against baseline — survivability with vs. without recommended additions: —.
5. Constraint-binding analysis — which constraints are binding so the customer can see what additional resource availability would change the recommendation (right column).
6. Source-traced citations — every probability used in the computation traces to entries in the Data Backbone tab (S-001, OP-*, OEM-Partner, OPEN).
Tactical Demonstration · Worked Example
Proposal §6 — SOCOM Small Team · JCO Yuma 50+ Threat Profile Saturation engagement: 30× Group 1–3 baseline, 15× FPV swarm with decoys, 5× fiber-optic FPV, <1× satellite-bridged FPV (contingent). Stack composition from §6.4 active. Engagement chain visualized in real time.◉ Engagement Live
Threats inbound0
Detected0
Classified0
Engaged0
Defeated0
Leaked to asset0
Realized P_survive
—
Engagement Log · Live
Threat-Class Composition N = 50
Engagement-Chain Logic
Cueing-to-engagement window for FPV-class engagement is seconds to low-tens-of-seconds at engagement geometry. Operator decision authority must be pre-delegated (§5 TTP "Cueing-to-engagement window discipline"). Kill chain: detect → classify → engage → kill, layered economy-of-defeat (cheapest applicable defeat first).
Residual Risk Notes
T-019 (sat-bridged FPV) has no Ku-band detection in the recommended stack — procedural mitigation only. Decoy-survivors in saturation: HPM is sparing of interceptors but does not defeat 100%. Operator-decision-authority dependency assumed delegated to small-team level — must be operationally validated in pre-deployment training.
Data Backbone
Intelligence Stack v1.0 · §2 sources · §3 threats · §4 equipment · §5 TTPs The defensible asset. Every probability term in §5 traces back to entries here. Operator identities, units, and locations excluded per partner-confidentiality and personal-security protocols.Source Catalog (Anonymized) §2
| ID | Class | Background |
|---|---|---|
| S-001 | DA Principal · CEO · Author of Doctrine | U.S. Army Infantry 11B (08–11) · KRG Peshmerga Advisory (16) · UAF combat-support (22–23) · UA MoD Bn Revanche (24) · HUR/GUR SO + assault regiment service |
| S-002 | DA Program Manager | U.S. citizen · French Foreign Legion 13e DBLE · operational role w/ Ukrainian intel service · identity withheld for personal security |
| OP-A — OP-L | Ukrainian FPV operator network | 12+ vetted operators · 700+ recent combat hours combined · 1,800+ total flight hours combined · UA MoD FPV creds · Victory Drones / Dignitas Foundation creds |
| OP-M | Allied counter-FPV operator | Direct combat experience: kinetic counter-FPV · mesh-net, shotgun, bunker tactics · Western/PRC/UA equipment-channel awareness |
| OEM-Partner | UA capability-provider engineering teams | Engineering and product personnel from UA OEMs in active partnership conversation · NVECTECH (executed 13 Mar 2026), Specter Aero NDA (24 Apr 2026), Unwave (engineer recruitment) |
| OPEN | Public authoritative literature | DoD doctrine · federally-funded research · CSIS, CNAS, RAND · UN Security Council reports · manufacturer pages |
Confidence Label Definitions §2
DCODirect combat observation — S-001 or specific OP-* with personal in-theater observation
ONCOperator-network corroborated — ≥2 OP-* sources reporting independently
SOCSingle-operator current — one OP-* source within recent 90-day window
OEMOEM-disclosed — OEM-Partner reporting product capability
OSCOpen-source confirmed — multiple authoritative public sources independently reporting
ASSAdjacent-source supported — federally-funded analytical body or major think tank reporting; operator corroboration incomplete
Standing rules. Capability-provider rule: "currently testing in live combat" claims require source-company confirmation before external citation. Operator intelligence rule: external sharing requires explicit principal-level approval and source-class verification. No "combat-proven" claims (preferred: "operationally refined" / "combat-informed design").
Threat Catalog §3 · 7 entries shown · refresh: weekly review · publish: quarterly
| ID | Threat Class | Key Operating Characteristic | Confidence | Source |
|---|---|---|---|---|
| T-019 | Satellite-bridged FPV | Ku-band uplink (~14 GHz) / downlink (~12 GHz); range >100 km; operator hundreds of km away; no nearby ground emission. Outside conventional Western RF coverage. | SOC | OP-* (90d) |
| T-014 | Fiber-optic FPV | Wired control via paying-out fiber spool; multi-km engagement; no RF emission; airframe detectable by acoustic / visual / thermal / radar; trailing fiber vulnerable to mechanical defeat. | ONC OEM | OP-* + OEM |
| T-012 | Coordinated FPV swarms with decoys | Mixed-platform engagements with decoy patterns; 50+ simultaneous threats per JCO Yuma demonstration June 2024. Saturation-defeat is a fusion-and-prioritization problem before a kinetic problem. | ONC OSC | OP-* + OPEN |
| T-008 | Loitering munitions (Shahed/Lancet) | Multi-hundred-km range for Shahed-class; Group 2-3 spectrum; legacy Western stack reasonably effective in fixed-site IADS; cost-per-defeat economics constrain interceptor employment at scale. | ONC OSC | OP-* + OPEN |
| T-001 | Commercial COTS in non-permissive contexts | DJI / Autel / Parrot in NSAG employment; documented JNIM bottle-IED FPV use in Burkina Faso (Feb 2025); Iranian Mohajer/Shahed transit through Port Sudan since late 2023. | OSC | OPEN |
| T-022 | Adversary EW evolution against friendly UAS | Continuous Russian GNSS denial, control-link jamming, video downlink spoofing across UA theater; drives friendly migration to fiber-optic and pre-programmed autonomous platforms. | ONC OSC | OP-* + CSIS |
| T-014b | RPG-armed FPV (10-inch class) | 10-inch FPV airframe carrying RPG warhead. Defeats most kinetic point engagement at standoff: warhead fires forward on net interception, putting the shooter at risk even on a successful net hit. Operator judgment: open-air engagement is doctrinally inadvisable. | DCO | OP-M |
Equipment Performance Catalog §4 · selected · refresh: continuous · publish: quarterly
| Capability Class | Observation Against Modern Threat Surface | Confidence | Source |
|---|---|---|---|
| Conventional handheld RF (sub-6-GHz) | Effective vs RF-controlled FPV. Blind to fiber-optic FPV (no emission). Blind to satellite-bridged FPV (Ku-band outside coverage). | ONC | OP-* + S-001 |
| Active radar (electronically scanned, sub-15 GHz) | Effective at airframe-level detection of small UAS at multi-km ranges. Most robust answer to fiber-optic FPV in operator-network observation. Active emission creates counter-detection vulnerability. | DCO ASS | S-001 + OPEN |
| Counter-FPV SIGINT (video downlink intercept) | Provides intent-classification data unavailable from other passive sensors. Effective vs analog/lightly-encrypted; degraded vs strong-encrypted or fiber. | ONC OEM | OP-* + OEM |
| NVECTECH PATRIOT 2 thermal handhelds | Operationally refined since 2022 in UA theater. Distribution agreement to U.S. market executed 13 Mar 2026. Specific spec performance: manufacturer-claimed; DA verification in progress. | OEM | OEM-Partner |
| Allied homebuilt sonar detection station (UK channel) | ~$1k–$2k; ~500 m position accuracy; active acoustic w/ reflected return. Detects fiber-optic FPV and conventional RF FPV — closes the fiber-optic detection gap. Provenance: single-developer homebuilt prototype (UK channel); v2 in development pending builder funding. | DCO | OP-M (single builder) |
| Mesh-net handheld pistol (kinetic interceptor) | ~50 m effective range, accurate-shooter dependent. Operator-rejected against RPG-armed FPV: warhead fires forward on net interception, shooter exposed in the open. OP-M field history: had two units, "tossed them out week after getting them." | DCO | OP-M |
| TATUSKY TTSKL03 handheld decoder (PRC origin) | Marketed 3 km range, waterproof. CRPC 2.0 data-chain decoding (graphic/video, SN, geo-loc, trajectory, altitude, speed, RC position). Operator caveat: combat-validation pending. PRC-origin compliance posture excludes blanket NDAA / TAA representation. | OSC SOC | OPEN + OP-* |
Critical operator-validated equipment-performance gaps no current Western system addresses adequately:
(1) Ku-band coverage in handheld and vehicle-mounted RF detection · (2) wide-band acoustic detection meshing · (3) protocol-aware reactive jamming with operator-validated waveform library · (4) open-architecture fusion accepting third-party sensor feeds with operator-validated classifier · (5) operator-validated training data for Western fusion classifiers.
Operator-Validated TTPs §5 · selected · refresh: continuous · review: monthly
| TTP | Description | Source / Confidence |
|---|---|---|
| Multi-modal redundancy | Always pair RF detection with at least one signature-independent modality (acoustic or thermal) when fiber-optic threat is in envelope. | OP-* + S-001 · ONC |
| Passive-first posture in low-vis SOF | Passive sensors only when friendly signature exposure is mission-critical; active radar / jamming reserved for moments when ROE permits emission. Aligns with USSOCOM SOF SUD signature-management framing. | OP-* + S-001 + OPEN · ONC |
| Cueing-to-engagement window discipline | FPV-class engagement window is short (seconds to low-tens-of-seconds). Operator decision authority must be pre-delegated. | OP-* + S-001 · ONC |
| Layered economy-of-defeat | Cheapest applicable defeat mechanism engaged first; expensive interceptors reserved for residual threats. Underlies ATP 3-01.81 layered-defense principle. | OP-* + S-001 + OPEN · ONC |
| Engagement reveals position | Any engagement against hostile UAS reveals friendly position to follow-on hostile employment. Mission planning premise: avoid engagement where mission profile permits. | OP-* + S-001 + OPEN · ONC |
| Bunker + early notice doctrine vs RPG-armed FPV | Against 10-inch RPG-armed FPV: "perfect bunker and notice beats standing in the open firing at a ten inch and rpg." Mesh-net rejected; shotgun secondary; bunker primary. | OP-M · DCO |
| Detect-before-counter premise | "You need to see the threat before trying to counter it." Detection capability is the binding constraint on counter-FPV mission outcomes; investing in a defeat layer without proportionate detection capacity is doctrinally inadvisable. | OP-M + S-001 · ONC |
| Sonar-dominant detection composition for sub-$5k fixed-site | For sub-$5k fixed-site composition with fiber-optic FPV in envelope: "sonar only — RF won't pick up fibre." Excludes RF-decode-class detection at this composition tier; concentrates budget on active sonar. | OP-M · DCO |
CEO Five Combat-Observation Judgments §5.1 · S-001
- Pace of iteration. Ukrainian-side iteration speed against Russian counter-drone employment exceeds anything observed in the prior decade of U.S. counter-VEO operations.
- Cost asymmetry as doctrine-shaping force. Western counter-UAS doctrine that does not internalize cost-per-defeat economics will be incorrect against the cost asymmetry the modern threat surface imposes.
- Operator decision authority is the rate limiter. No fusion stack survives if the operator cannot decide and engage within the engagement window.
- Trust is the moat. The structural advantage of the firm's operator network is trust developed through service — not buyable by capital alone.
- U.S. doctrinal language must evolve. Published language ("counter-small UAS," "Group 1-3") reads correctly for the threat surface of 2018; it does not adequately partition the modern threat surface (satellite-bridged, fiber-optic, autonomy-stack-attacked).