Ragnarok Solutions / Robotics / Defense Defense and public sector

The robots that carry, clear and resupply need eyes you can trust.

Ragnarok tests and hardens the perception of autonomous systems so they read the world correctly and say when they can't. We do not build weapons or targeting systems.

What we do for defense and public buyers

six lines, all of them protection, upkeep or testing
Sustainment

Sustainment and maintenance

Robots and AI that read gauges, fluids, wear and damage on vehicles, aircraft and equipment, scored by condition before they are trusted in a motor pool or a depot. The misses go into a test set the next version has to pass.

Position

Navigation without GPS

Resupply and reconnaissance robots in jammed or denied areas, underground and in cities. We score whether the robot knows it is lost and how fast it says so, not only how far it drifted.

Protection

Force-protection detection

Sensors that warn people of drones, intrusions, and buried or camouflaged threats, scored for misses and false alarms by condition. The output goes to a person who decides. It does not feed a weapon.

Hardening

Sensor trust

Your sensors against dazzle, spoofing and jamming. We test how a system behaves when one sensor is lied to and the others are not, and we harden the layer that decides which to believe.

Referee

Independent testing

Any system where a misread puts people at risk, including friend-or-foe perception. We test as an independent evaluator with no stake in fielding and no part in the build. You get the misidentification rate by condition, sealed before the run.

The team

Human and machine together

Operators and robots tested as one team in degraded visibility, cold and lost communications, with wrong calls seeded in. What the operator catches, what gets waved through, and how many robots one person can supervise when the link drops.

Where we stop

the policy, in full view

Testing makes a system safer, not more lethal.

  • We do not build weapons, targeting or fire-control systems. This covers sensors, software and AI whose purpose is to select or engage targets.
  • We do not build detection whose output feeds a weapon automatically. We may test its misidentification rate.
  • We do not jam or spoof anyone else's sensors. We defend and test yours.
  • Human judgment stays in any decision to use force. We align with DoD Directive 3000.09.
  • Custom work for an armed platform, or any contract where our output feeds an engagement decision, needs the founder's written sign-off before it starts.

What you get

evidence a program office can file

A scorecard by condition

Smoke, dust, low light, cold, motion, jamming, interference. One number per condition, with the key sealed before the run and the hash in your hands first.

The misses, with the truth beside them

Every reading the system got wrong, every item it should have declined, every false alarm. Ranked by what it would cost in the field.

A test set that outlives the version

The next software drop has to pass the same items before it goes out. The record of what passed and what did not stays with the program.

What a false alarm costs

A sensor that cries wolf every shift gets switched off. The one that misses once gets people hurt.

Both failures are measurable before fielding. Neither is measured by a vendor demo on a clear day. A sealed test is a fixed fee and a few weeks.

Questions we get

Do you build weapons or targeting systems?

No. Ragnarok tests and hardens the perception of autonomous systems so they read the world correctly and say when they can't. We do not build weapons or targeting systems.

Will you test a military system?

Yes, as an independent evaluator with no stake in fielding, where a misread puts people at risk. Testing makes a system safer, not more lethal. Human judgment stays in any decision to use force.

Will you build navigation for an armed platform?

We do not sell a navigation product today. Under our policy a generic product sold to everyone would be fine. Custom work for an armed platform, or any contract where our output would feed an engagement decision, needs the founder's written sign-off before it starts.

Who decides the gray areas?

The founder, in writing, against a published policy. If a request sits on the line, we say so before we quote.

Tell us the system and the conditions it has to work in.

What it reads, where it operates, and what happens when it is wrong. We reply within one working day with how we would test it and what the policy allows.

Photos: bomb-disposal robot, U.S. Army photo by Sandra Wilson, cropped. Legged Squad Support System, DARPA, cropped. Robotic payload carrier, U.S. Army photo by Sgt. Chandler Coats, cropped and edited.