EU AI Act Article 50 — 20 days to seal | Get passport

DEFONEOS Real-World Scenarios

6 SCENARIOSEND-TO-ENDTIMELINE-VERIFIEDBFT-GOVERNED

Six scenarios. End-to-end timelines. Sensor fusion. BFT-governed responses. SIGIL-audited.

Scenario 1 · Yorkshire Flood Response

MET OFFICE + EA + COASTGUARDRESILIENCE FORUM

Trigger: Met Office forecasts 200mm of rain in 24h across North Yorkshire. Environment Agency river-level sensors show River Ouse rising 0.3m above seasonal normal at Skelton. DEFONEOS must correlate forecast, sensor, demographic, and infrastructure data; classify the threat; and pre-position response assets.

Sensor inputs (fused in real time)

SourceChannelUpdate
Met OfficeRainfall forecast (NetCDF)Hourly
Environment AgencyRiver level (m AOD)15-min
Ordnance SurveyTerrain, flood plain polygonsStatic
Companies HouseCare home & sheltered housing locationsWeekly
OpenStreetMapRoad network + elevationDaily
ONSIMD decile, vulnerable-person densityQuarterly
Sentinel-1SAR backscatter — ground saturation6h
Batear acousticUnusual river noise (debris impact)Real-time

Timeline (T+0 = forecast issue)

T+00:00 · Met Office issues amber rainfall warning. DEFONEOS noise→intuition pipeline detects convergence pattern across 4 data sources. BFT council auto-convened (quorum 23/33 within 90 sec).
T+00:02 · intuition_engine raises recommendation: pre-position 12 high-volume pumps at Skelton, 6 at Linton-on-Ouse, 8 at York city centre. Confidence 0.91.
T+00:03 · BFT vote: 26 for, 1 against, 6 abstain. Action:DispatchPumps SIGIL-emitted to NYLRF (North Yorkshire Local Resilience Forum).
T+02:15 · First EA river-level data confirms forecast. DEFONEOS recommends second-stage pre-position: 4 rescue boats, 2 Coastguard hovercraft, MEDEVAC helicopter on standby at RAF Topcliffe.
T+06:00 · Sentinel-1 pass. SAR shows 14km² already saturated. BFT escalates from pre-position to active response. Action:ActivateFloodPlan emitted.
T+09:30 · First 999 calls begin. DEFONEOS auto-triages: 47 calls in 15 minutes, 12 categorised as P1 (life risk). Maps against care home + vulnerable-person register. 8 P1 calls map to known vulnerable residents; resilience forum dispatched welfare checks.
T+14:00 · DEFONEOS composes public alert via EA Flood Warning Service + DEFRA. 23,000 households notified. SIGILed.
T+24:00 · Rain tapers. River levels peak 1.8m above normal. Zero fatalities in DEFONEOS-supported response area (3 in non-supported neighbouring area).

Outcomes

Forecast-to-preposition lead time3 minutes (vs 90 min industry avg)
P1 calls triaged automatically47 / 47 (100%)
Vulnerable residents identified312 (vs 47 manually known)
Public alerts sent23,000 households
SIGIL entries written14,302 (every decision + sensor reading)
Post-event report auto-generated17 minutes after T+24

Scenario 2 · Dark Vessel Detection — North Sea

AIS + SATELLITE + RFMARITIME OPS CENTRE

Trigger: AISstream shows a 3-hour gap in tracking for a 290m tanker off the coast of Northumberland. DEFONEOS must determine if this is an AIS equipment fault, a sanctioned vessel attempting to spoof, or a "dark" vessel engaged in illicit activity (sanctions evasion, illegal fishing, smuggling).

Sensor inputs

SourceChannelUpdate
AISstreamVessel position, heading, speedReal-time
Sentinel-1 (Copernicus)SAR ship detection6h revisit
Sentinel-2Optical confirmation (clear sky only)5-day revisit
RF geolocation (PX4 radar)Marine radar cross-sectionReal-time
EO/IR gimbalVisual ID at long rangeOn-tasking
UK Sanctions ListIMO + flag matchDaily
OFAC SDNBeneficial ownershipDaily
GDELTNews events in region15-min

Fusion timeline

T+00:00 · AIS gap detected. Track.lastSeen > 3h flagged. DEFONEOS opens incident MOC-2026-07-05-001.
T+00:01 · Predicted position estimated via dead reckoning from last heading + speed + metocean drift. Uncertainty ellipse: 4.2 × 1.8 nm.
T+01:30 · Sentinel-1 pass. SAR detects 1 vessel of compatible size inside the uncertainty ellipse at 55.4N -1.5W. Match probability: 0.74.
T+02:10 · RF geolocation sensor (coastal) detects strong marine radar return at 55.42N -1.52W (range 8 nm). Range ambiguous; cross-bearings needed.
T+02:30 · EO/IR gimbal tasked. Aircraft tasked from RAF Lossiemouth via pre-approved cooperative flight plan. ETA 90 minutes.
T+04:00 · EO/IR imagery returns. Visual classification: oil tanker, dark hull, no name visible. Length 280-300m. Confidence 0.88.
T+04:02 · BFT council session: Promote to "Dark Vessel — High Confidence". 24/33 for, 3 against, 6 abstain. Action:AlertMOC + Action:DispatchNavalTasking.
T+04:15 · Alert reaches UK Maritime Operations Centre + Royal Navy Flag Officer Sea Training. Decision: send surface vessel for positive ID.
T+08:30 · HMS example closes to 1 nm. AIS reactivates under inquiry. Vessel identified as sanctioned shadow fleet tanker. Boarding action initiated.

Outcomes

Detection-to-MOC alert latency4 minutes (vs 4-8 hours industry avg)
Sensor-to-positive-ID8h 30m total
Sanctions match (post-boarding)YES — vessel on OFAC SDN list
BFT quorum reached24/33 (above 23/33 minimum)
SIGIL chain (replayable)142 entries

Scenario 3 · Urban Drone Swarm Defence — Central London

RF + ACOUSTIC + EO/IR + BFT EMERGENCYCOUNTER-SWARM

Trigger: Three unidentified drones detected converging on a Category A government building in central London. DEFONEOS must classify, track, predict trajectory, and recommend response — without escalating to kinetic effectors absent explicit BFT + human authorisation.

Sensor inputs

SourceChannelUpdate
Batear acoustic arrayDrone rotor signatures, multiple classesReal-time
PX4 RF radarDrone control link (2.4/5.8 GHz)Real-time
EO/IR gimbal (multiple)Visual classificationOn-tasking
Air traffic (OpenSky)Manned aircraft exclusionReal-time
GDELTCoordinated event signals15-min
NPCC / MET PoliceC2 coordinationOn-tasking

Fusion timeline

T+00:00 · Batear array detects 3 simultaneous drone rotor signatures within 800m of target building. Distinct frequency signatures (DJI Mavic class 1, DJI Phantom class 1, custom fixed-wing class 3).
T+00:05 · PX4 RF radar confirms: 3 control links in 2.4 GHz, one also on 5.8 GHz. Bearing correlates with acoustic.
T+00:08 · EO/IR slewed. Visual classification: 2 quadcopters (commercial, possibly DJI), 1 fixed-wing. Confidence 0.92, 0.89, 0.78.
T+00:12 · Trajectory prediction: all 3 converging on building rooftop. Estimated time-to-arrival 4 min 30s.
T+00:15 · EMERGENCY BFT SESSION auto-convened. Reduced quorum (21/33) due to urgency. Recommend: (a) RF jamming authorisation request, (b) alert NPCC + MET, (c) prepare counter-drone swarm.
T+00:18 · BFT vote: 22 for, 2 against, 9 abstain. Action:RequestRFJamAuthority + Action:AlertNPCC.
T+00:30 · Senior Civil Servant (Gold command) approves RF jamming. SIGILed with SCA Ed25519 signature.
T+00:32 · RF jam activated on 2.4 + 5.8 GHz (JSP 440 compliant — non-emergency services frequencies only). All 3 drones enter controlled descent.
T+02:00 · Drones recovered in soft-landing zones. Counter-swarm never deployed (jam sufficient). Post-event forensics: 2 commercial drones modified for longer range; 1 custom built from open-source plans. No explosives. Investigation ongoing.

Defensive design choices

Scenario 4 · CBRN Detection — Port of Felixstowe

CHEMICAL + RADIOLOGICAL + BFT EMERGENCYPHE + DSTL

Trigger: Customs container scanner flags anomalous gamma reading on a 40ft container declared as "machine parts" from a high-risk origin. DEFONEOS must triage, escalate, and coordinate with Public Health England + DSTL.

Sensor inputs

SourceChannelUpdate
Port scannerGamma spectroscopy, neutronPer-container
Met OfficeWind direction/speed (plume model)Hourly
ONS demographicsPopulation density within plume radiusStatic
OS Open DataCritical infrastructure locationsQuarterly
GDACS / ReliefWebInternational event signalsReal-time
PHE knowledge baseSubstance ID + treatment protocolsCurated

Fusion timeline

T+00:00 · Gamma reading 0.8 µSv/h above background, consistent with industrial source or low-grade radiological material. DEFONEOS opens CBRN incident.
T+00:30 · Isolation zone auto-defined: 50m radius. Container moved to dedicated isolation pad by remote-controlled straddle carrier.
T+02:00 · PHE specialist team dispatched. DSTL on standby. DEFONEOS plume model: worst-case wind (NNE 15 km/h), 2 km exclusion, 20 km shelter zone.
T+04:00 · Spectroscopy confirms Cs-137 (industrial, sealed source). Confidence 0.96. NOT a dispersal device — likely orphaned medical/industrial source in cargo.
T+04:30 · BFT vote: 27/33 for safe containerisation + return-to-sender per IAEA guidance. Action:ContainmentPlan + Action:AlertIAEA.
T+24:00 · Source containerised in IAEA-compliant Type B(U) package. Origin country notified via HM Customs.

Defensive design choices

Scenario 5 · National Grid Power Outage — Multi-County

SCADA + NETWORK + ECONOMIC IMPACTDSO + NG ESO

Trigger: National Grid ESO loses visibility on 3 grid supply points (GSPs) simultaneously. 240,000 customers across 3 counties lose power. DEFONEOS must triage cause, predict restoration, coordinate with DSO and emergency services.

Sensor inputs

SourceChannelUpdate
NG ESO SCADAGSP status, frequency, demandReal-time
DSO telemetryFeeder status, transformer loadingReal-time
Telco (BT + Vodafone)Cell tower power status (battery backup)5-min
Water utilityPump station power5-min
Traffic (National Highways)Signal blackouts, congestionReal-time
OpenAQBackup generator emissions (CO, NOx)Real-time
RTSP cameras (CCTV)Street lighting + signal visibilityReal-time

Fusion timeline

T+00:00 · 3 GSPs drop. DEFONEOS opens incident GRID-2026-XX-001. Correlates with NG ESO telemetry.
T+00:02 · Root cause hypothesis: transmission fault on supergrid. Backup routes triggered automatically.
T+00:15 · 40% of customers restored via auto-reclose. 60% remain down (240k → 144k).
T+00:30 · DEFONEOS impact map: 23 care homes without power (auto-paged to DSO priority list), 4 water pump stations on battery, 7 cell towers on backup.
T+01:00 · BFT session: prioritise restoration sequence. 28/33 for proposed order. SIGILed.
T+02:30 · All 23 care homes restored (DSO crews + portable generators). Water utility alerted re: pump station batteries.
T+06:00 · Full restoration. Post-event: cause identified as third-party cable damage during construction. National Grid instructed to inspect similar routes.

Scenario 6 · Mass Casualty Incident — Motorway Pile-Up

AMBULANCE + POLICE + FIRE + AIR AMBULANCEMETHANE + JESIP

Trigger: M6 collision in Cumbria. 17 vehicles involved. Low visibility + ice. Initial 999 calls indicate 30+ casualties. DEFONEOS must implement METHANE (Major Incident Medical Hazard Assessment) + JESIP (Joint Emergency Services Interoperability) protocols automatically.

Sensor inputs

SourceChannelUpdate
999 CADCall text + location + triage categoryReal-time
National Highways CCTVVisual confirmation of sceneReal-time
Met OfficeWeather (visibility, ice, temp)15-min
Air Ambulance (NWAA + GNAA)Aircraft availability + ETAReal-time
NHS trusts (UHMB + NCUH + RLI)ED capacity + trauma team availabilityReal-time
Blood stock (NHSBT)O-negative, plateletsReal-time
ONS demographicsVulnerable persons known to be in transitStatic

Fusion timeline (METHANE)

M — Major Incident declared · 17+ vehicles, 30+ casualties confirmed by 999 calls + CCTV. T+00:01.
E — Exact location · M6 J36-J37 northbound. Lat/lon auto-confirmed via ANPR + 999 cell triangulation.
T — Type of incident · RTC, multi-vehicle, fog + ice. Predicted severity: high (visibility near-zero, pile-up likely expanded).
H — Hazards · Fire risk (fuel), ice, low visibility, traffic build-up approaching. Auto-paged: Fire + Air Ambulance.
A — Access · M6 NB closed; emergency corridor via J36 entry. HETS (Highways England Traffic Officers) dispatched for contraflow.
N — Number of casualties · 47 confirmed by 999 + CCTV (initial 30 was under-estimate). Triage: 8 P1, 22 P2, 17 P3.
E — Emergency services · 12 ambulances, 3 air ambulances, 5 fire engines, 2 HART teams, 1 police helicopter auto-dispatched.

DEFONEOS-specific value-add

Outcomes (illustrative baseline vs DEFONEOS)

Time to all services on scene22 min (vs 35 min baseline)
Casualties to definitive care8 P1 in 47 min (vs 75 min baseline)
Preventable deaths (modelled)0 (vs 2-3 baseline estimate)