Functional safety · AI safety · Autonomy certification
We bridge IEC 61508, the emerging AI-safety family (ISO/IEC TR 5469 · TS 22440 · TS 25223) and UL's certification standards (UL 4600 · UL 3115) into a single, evidence-based safety case — for AGVs, mobile platforms, robot arms and humanoids alike.
The problem
Functional safety, AI trustworthiness and certifiability are three different failure modes. A platform can clear one and quietly fail the other two — and the gap only shows up once it's deployed.
A motor controller drifts out of calibration, a bus message is corrupted, a watchdog misses its deadline. Classic malfunctioning behaviour of an E/E/PE safety-related system — random hardware faults and systematic faults in hardware and software.
Every actuator is healthy and every joint is within tolerance, and the arm still grips the wrong part or the AMR still misreads the aisle — because the model's confidence was never calibrated to the situation it met. Insufficiency in perception or decision-making, not a fault.
The robot works in the lab and in every demo you've run. A customer, insurer or certification body needs an argument, not a demo reel — evidence of what was tested, what wasn't, and why the residual risk is acceptable.
The seam is the risk. A mobile platform can pass every IEC 61508 safety function test and still misclassify a person as a pallet. It can have a well-behaved perception model and still have no assurance case a customer's insurer will accept. Only an argument that spans all three can close the case.
What we do
One hazard. Three families of cause. One safety case that traces every hazard through every applicable framework to evidence — whatever the robot is built to do.
The safety-related control system is faulted. Answered with SIL determination, safety functions, diagnostic coverage and hardware fault metrics.
The system is fault-free but its learned behaviour can't be trusted. Answered with AI risk analysis, safety requirements for adaptive systems and calibrated uncertainty.
The behaviour may well be safe, but nobody can show it. Answered with a structured safety case and independent certification evidence.
Services
Engage one where you have a specific gap, or all three where you need a complete argument. Either way the work product is designed to slot into the same safety case.
01 / Functional safety
Safety-related control system architecture, SIL determination, hardware fault metrics, systematic capability, and the functional safety case for the E/E/PE parts of a robotic or autonomous system.
02 / Artificial intelligence
AI risk analysis, safety requirements for adaptive and learned behaviour, model evaluation and calibrated uncertainty — from ISO/IEC TR 5469 guidance through to the emerging TS 22440 and TS 25223 specifications.
03 / Autonomy & certification
Safety-case-based evaluation of autonomous products and certification of AI-based products, mapped to the evidence a customer, insurer or certification body will actually accept.
Who we serve
The physical form changes what can go wrong. It doesn't change the questions we ask about it.
Automated guided vehicles and autonomous mobile robots navigating warehouses, factories and hospitals — where ISO 3691-4 and ANSI/RIA R15.08 meet functional safety and AI perception.
Wheeled and tracked platforms carrying payloads or people, where safety-rated stopping, obstacle detection and autonomy assurance have to work together.
Fixed and mobile manipulators performing tasks alongside people or fully autonomously, where safety functions and AI-driven grasping both need to be proven safe.
Bipedal and human-scale robots combining actuation, perception and decision-making in one platform — the hardest single case for an integrated safety argument.
Engagement
Consulting that produces artefacts, not slideware. Every engagement ends with something you can hand to a certification body, an integrator or a customer.
Gap and readiness assessment against IEC 61508, the emerging AI-safety standards and UL's certification requirements at once. A prioritised finding list mapped to clauses, not a generic maturity score.
Safety functions, decomposition and the allocation of requirements across functional safety, AI safety and certification evidence — so the architecture carries the argument rather than fighting it.
A verification and validation strategy that says what each test proves: fault injection, model evaluation, uncertainty calibration and the coverage rationale behind them.
A structured safety case — goals, strategies, context, evidence — with a defensible residual risk statement, plus support through independent assessment and certification.
Regulatory readiness
Standards conformance is necessary but not sufficient. We shape the technical evidence so it maps directly onto the regulatory instrument or certificate you actually need.
The new EU rulebook for machinery, applicable from January 2027, with expanded scope that now explicitly covers autonomous mobile machinery, AI-based safety systems, industrial robots and cobots.
The core Type-C safety standard for driverless industrial trucks and their systems — the reference point for AGV and AMR safety functions, person detection and braking performance.
Industrial mobile robot safety requirements for the US market, covering the robot itself, system integration and — in its forthcoming Part 3 — the user's operating environment.
Where standards conformance turns into a certificate: UL 4600 safety-case evaluation and UL 3115 AI-product certification, mapped against ISO, IEC and NIST frameworks for cross-recognition.
Business-level fluency
About
Build your trustworthy system with me
safeautonomy is an independent consultancy for the safety of robotic and autonomous systems. We work with robot manufacturers, systems integrators and logistics and manufacturing operators at exactly the point where functional safety, AI safety and autonomy certification stop being separate disciplines.
Where we work
Headquartered inside one of Europe's largest concentrations of industrial robots, AGVs and automated logistics — Volkswagen's manufacturing plants and their supplier base — and working with clients well beyond it, on site wherever a programme needs us, not just reviewing documents from a distance.
Headquartered in one of Europe's largest robotics and automation manufacturing hubs.
Remote-first engagements across the EU, North America and Asia.
Get in touch
A 30-minute call is usually enough to tell whether your gap is in functional safety, in the AI evidence, in the certification case — or in the traceability between them.
Phone / WhatsApp
+352 661 553 282Based in
Wolfsburg, Germany — working worldwide