01LEADS
AUTHORIZED PROGRAM
Chornobyl NPP and Ukrainian authorities lead the site response; EBRD coordinates donor support, IAEA monitors and advises, and qualified contractors perform assigned work. Swovee would have no independent repair authority.
CHORNOBYL / SENTINEL N01REV A · 15 AUG 2026
NO HEROICS.
JUST EVIDENCE.
Swovee should not arrive at Chornobyl promising to print a new sarcophagus. It should arrive—if invited, qualified and useful—as a radiation-aware inspection and logistics system inside the existing Ukrainian-led restoration program.
CONTROLLED SITE14·04·26
15 M² BREACH / REPORTED
N01REMOTE / HOLD
01 / SURVEY02 / REGISTER03 / ASSIST
00 / CURRENT RECORDA DAMAGED BARRIER.
A DAMAGED BARRIER.
A CONTROLLED SITE.
Two facts can be true at once: the NSC lost important confinement functions, and official monitoring did not show radiological releases above limits as of the plant's 14 April 2026 update. The simulator must preserve that distinction.
REPORTED 14 APR 2026CONTROLLED + STABLE
Chornobyl NPP reported on 14 April 2026 that no releases above applicable limits had been recorded and industrial-site gamma dose rates remained within established control levels.
Official source ↗CONFINEMENTPARTIAL FUNCTIONThe New Safe Confinement no longer fully maintains internal leak-tightness after the February 2025 strike and fire, although its protective envelope continues to provide partial ecological isolation.
Official source ↗STRUCTURENO LOAD-PATH DAMAGE IDENTIFIEDThe plant reported no identified damage to the load-bearing structures of the NSC or the Shelter Object from the strike or firefighting. That does not mean the cladding and systems are undamaged.
Official source ↗RESTORATION BASISEST. ≥ €500M / FUNCTION BY 2030EBRD states that restoring full NSC functionality by 2030 could cost at least €500 million. Preparatory assessment, design, structural monitoring and emergency-tightness work are underway; that is not the same as completed full restoration.
Official source ↗CLAIM DISCIPLINE“Damaged” does not mean “uncontrolled release.” “Stable” does not mean “fully restored.”
SCHEDULE / OFFICIAL PLANNING CONSTRAINTThe Shelter Object localizing structure's documented service life runs through 31 October 2029, subject to a justified extension; EBRD says full NSC functionality is needed by 2030. Neither date is an automatic failure event. ↗01 / HONEST COMPARISONIS SWOVEE
IS SWOVEE
BETTER?
Not as a replacement for the official repair program. The useful question is narrower: can a qualified robot reduce human exposure or improve repeatability for a specific task without adding more risk?
02LEAD
SPECIALIST ACCESS SYSTEMS
Qualified cranes, lifts, survey tools, temporary works and contractor methods are the baseline. A new rover does not outrank proven access equipment merely because it is autonomous.
03MAY COMPLEMENT
SWOVEE SENTINEL N01
A future platform could add repeatable multisensor registration, exposure-aware routing, tool logistics and a traceable digital record—only after qualification and integration into the official work package.
GO / NO-GOBuild a clean-site demonstrator only if an operator identifies a real task that current tools do not already solve safely.
02 / NEW MACHINE FAMILYMEET N01.
MEET N01.
THE SENTINEL.
N01 is not the land Rovalizer with shielding added. It is a proposed system-of-systems for measurement, light tool presentation and controlled logistics. Its first material is trustworthy data.
MISSION DESIGN CENTERReduce dose-hours.
Reduce dose-hours.
Increase evidence.
The concept separates mobility, metrology, radiological safety, access and recovery so that each can be qualified independently. A hardened black box records raw sensor timing, operator commands, AI recommendations, overrides and aborts.
- CONTROL
- Tethered or supervised remote
- AUTONOMY
- Navigation assistance + stop enforcement
- CONTACT
- None by default; force-limited if released
- MATERIAL
- Imported, certified tools and repair media only
N01-GNEW CONCEPT
Ground carrier
A low-profile electric base for clean-to-controlled-zone logistics, perimeter survey and instrument carriage. Smooth, sealed surfaces support inspection and decontamination.
N01-VQUALIFICATION REQUIRED
Vision mast
LiDAR, calibrated imagery, thermal sensing and surveyed fiducials build a repeatable defect map. AI may flag change; a named engineer accepts classification.
N01-RINDEPENDENT SAFETY CHANNEL
Radiation pod
Calibrated dose-rate sensing and accumulated component-dose accounting remain separate from perception software. Saturation, drift or disagreement forces a hold.
N01-AINTERFACE TBD
Access head
A tethered inspection head could use a certified boom, rail, wheel or other site-approved carrier. No magnetic, suction or climbing interface is presumed safe without substrate and load testing.
N01-HLIGHT INTERVENTION ONLY
Hold + present tool
Force-limited tooling could present a camera, gauge, marker, sealant cartridge or qualified temporary patch while authorized crews retain command and verification.
N01-DMANDATORY
Recovery dock
A shielded service bay closes power, communications, contamination survey, tooling, packaging and disabled-vehicle recovery. The rover is not a system without it.
03 / OPERATING MODELSEVEN STEPS.
SEVEN STEPS.
ONE CONTROLLED TOUCH.
The system advances by evidence gates, not confidence theatre. “Return with better data” is a valid mission outcome.
- 00
Authorize the question
The site operator and regulator define the exact inspection question, work zone, stop limits, data custody and human decision owner.
- 01
Rehearse off site
Run the same route and tool sequence on a clean physical mock-up and synthetic digital model. A passing simulation is evidence for a test—not permission for site work.
- 02
Establish the reference
Register official drawings, surveyed control points, weather, dose background and no-entry geometry. Unknowns stay visibly unknown.
- 03
Survey without contact
Collect synchronized geometry, imagery, thermal and dose data from the lowest-risk route. Sensor disagreement or localization loss triggers retreat.
- 04
Compare change
Engineers review deviations against prior surveys and inspect the evidence trail. AI prioritizes review; it does not declare structural safety.
- 05
Assist one approved task
If released, the system stages or holds one qualified item under force, wind, dose and communications limits. Structural repair remains the official contractor's work.
- 06
Recover + account
Return the machine, tools and consumables through survey and decontamination. Publish an as-left map, dose ledger, exceptions and unresolved questions.
04 / FUTURE TRAINING MISSIONCHORNOBYL:
CHORNOBYL:
HOLD THE LINE.
This page specifies a future synthetic training scenario; it does not launch one. Geometry, radiation fields and work procedures must remain fictionalized unless the site authority provides releasable data.
WIND 7.2 M/SWINDOW 10:42
N01
DOSE BUDGET82%
LINK A / BLOCK / LOCK
EVIDENCE3 / 7
01SIMULATED
Build the truth model
Register four survey anchors and validate every sensor before entering the controlled corridor.
02SIMULATED
Reach the observation lane
Plan the lowest-dose route while maintaining redundant communications and a recoverable return path.
03SIMULATED
Resolve the anomaly
Distinguish cladding damage, water ingress, sensor artifact and geometric occlusion without inventing certainty.
04SIMULATED
Protect the work window
Stage a qualified temporary cover and hold it within force limits for an authorized crew or remote tool.
DO NOT GAMIFYRadiation injury, conflict attribution, contamination spread, unauthorized entry or the 1986 loss of life.
The challenge comes from engineering tradeoffs and imperfect information—not spectacle.
05 / MISSION CONTRACTWHAT GOES IN.
WHAT GOES IN.
WHAT MUST COME OUT.
Every recommendation must resolve back to an authoritative input, named owner and reviewable record. AI-generated architecture is not a source of nuclear truth.
CONTROLLED INPUTS
No mission without a complete passport.
- Regulator-approved work package and mission polygon
- Authoritative NSC drawings, surveys and load restrictions
- Current radiological, contamination and weather limits
- Calibrated detectors with response and saturation records
- Qualified access interface, tool and temporary repair media
- Redundant power, communications and time synchronization
- Decontamination, packaging and disabled-machine recovery plan
- Named operator, health physicist and independent stop authority
REQUIRED OUTPUTS
The product is an auditable decision record.
- Versioned, coordinate-registered evidence model
- Raw and processed sensor record with uncertainty
- Vehicle, tool and component accumulated-dose ledger
- Human-reviewed defect and anomaly register
- Route, dwell, wind, contact-force and communications log
- As-placed or as-inspected geometry for each approved task
- Contamination survey and consumables disposition record
- Exceptions, aborts and open questions—never silently omitted
06 / READINESSN01 EARNS
N01 EARNS
EVERY METRE.
A polished demo, successful game mission or accurate scan is not a nuclear qualification. Each gate requires an owner, acceptance method, witnessed test and configuration-controlled evidence.
AUTHORITY
Written mission owner, regulator interface, site access and stop authority.
SURVIVE
Radiation, weather, contamination, power and communications qualification at stated margins.
MEASURE
Calibrated instruments, traceable timing, localization uncertainty and repeat-pass accuracy.
RECOVER
Demonstrated retrieval after lost power, lost link, tool jam or mobility failure.
TOUCH
Approved interface, force limits and representative mock-up testing for one bounded task.
INTEGRATE
Acceptance into the official repair work package, emergency plan and contamination controls.
07 / PRIMARY RECORDSOURCE THE SITE.
SOURCE THE SITE.
LABEL THE CONCEPT.
Current-status statements above are tied to Chornobyl NPP, EBRD and IAEA records. Everything labeled N01, Sentinel or simulation is a Swovee proposal. The sources document the problem and authorized response; they do not endorse Swovee.
BRING A TASK.
BRING THE
EVIDENCE.
Swovee should approach nuclear decommissioning experts with a question, not a claim: which narrow remote task remains dangerous, repetitive and underserved—and what would a credible qualification program require?