What robotic handling in a radiological environment means
Robotic handling of radioactive materials uses remotely controlled equipment or defined automated functions to perform tasks within an authorised facility. It can physically separate an operator from a hazardous area, but separation alone does not establish that the entire activity is safe. The complete system includes machinery, tools, observation, communications, personnel and procedures.
This entry concerns preventive management of deployment, rather than instructions for intervening on sources or waste. An inspection robot that gathers information should be distinguished from a manipulator acting on materials. Human-controlled teleoperation also differs from a bounded autonomous function. The word robot does not explain which decisions equipment can make or the conditions under which it has been authorised. Those distinctions should remain explicit throughout a project.
Existing evidence and work still at trial stage
On 30 June 2026, NDA and NRS reported Oldbury trials involving teleoperated arms and a separate autonomous sorting programme, Auto-SAS. Active commissioning of the latter was expected around mid-2027. This is a documented development example, rather than evidence of universal availability or fully established autonomy.
When assessing a proposal, distinguish a demonstration, a trial with representative material, acceptance within a facility and routine operation. A news report about a milestone does not establish all these states. We propose requesting evidence matching the intended task, configuration and environment. If a supplier changes a tool or control mode, earlier results should be reviewed before they are used as a safety argument. The procurement record should preserve these distinctions instead of treating technological maturity as a single marketing label.
Sources: [3]
Radiological protection: responsibility and scope
Spain’s CSN links worker protection to prior assessment, optimisation, classification of areas and workers, monitoring and training. The facility operator must establish applicable protective measures. Introducing a robot does not remove these responsibilities or the need to control exposures that may remain during other work stages.
As a planning criterion, the dossier should identify the exposure targeted for reduction and where exposure may persist: preparation, installation, maintenance or removal. Comparison needs to cover the whole task. It is insufficient to state that the operator remains farther away during one movement if an unplanned additional intervention is needed afterwards. Radiological assessment belongs to competent personnel and the facility’s authorised framework, rather than being inferred from the robot’s location alone.
Sources: [2]
Selecting technology for the task and its limits
IAEA guide SSG-47 considers remote handling to avoid high exposures and includes maturity, reliability, maintenance, secondary waste and training in technology selection. Non-radiological hazards also require attention. The guide does not approve any industrial robot for a nuclear environment.
An original selection matrix can compare functional scope, visibility, required precision, environmental compatibility and available evidence without disclosing sensitive site information. Every capability should be linked to an accepted test and an explicit limitation. Include tasks the robot will not perform. This boundary helps prevent equipment approved for observation from later being used for handling, or an accessory change from being treated as trivial. Acceptance should concern the actual system that will be used, rather than an unspecified product family.
Sources: [1]
Considering the complete work cycle
We propose dividing preventive review into preparation, testing, deployment, operation, maintenance and end of use. For each stage, identify who works, which authorisation is needed, which interfaces are used and what evidence confirms that expected conditions remain valid. This structure is an editorial organisational proposal, not a radiological intervention procedure.
Its purpose is to avoid transferring a problem to another person or another stage. Maintenance that is straightforward in an ordinary workshop may require a different assessment after equipment has been used in a controlled area. Similarly, moving, storing or removing components must fit approved procedures. Financial planning should account for these needs rather than covering only the robot’s purchase price and advertised operating hours. Otherwise, essential support tasks can disappear from the comparison.
Human factors and the teleoperation workstation
Distance changes the operator’s relationship with the task. A design review should examine required information, how the interface displays uncertainty and communication with the team responsible for the facility. A camera image should not be assumed to provide all the perception available to someone physically present.
We propose observing visual workload, posture, control layout and understanding of operating modes in safe trials. Control status, information limitations and responsibility for authorising changes should be clear. Assessment should include handovers and infrequent tasks, not only a demonstration performed by the specialist who designed the system. Competence means understanding limitations as well as using the controls. A useful evaluation also records questions raised by operators, since unclear assumptions can otherwise remain hidden behind a successful demonstration.
Training and trials before authorised use
UKAEA describes digital mock-ups and haptic simulators within LongOps for planning, training and testing remote operations in lower-risk environments. These tools support preparation of skills; a simulation alone does not establish the safety of a real intervention.
For a training programme, we propose observable objectives: recognising the active mode, interpreting incomplete information, communicating a discrepancy and implementing a suspension decision under the approved procedure. Record the configuration used and aspects left untested. A simulator representing normal operation alone leaves unanswered questions about coordination and change management. Final acceptance belongs to competent responsible personnel, with documentation and authorisations appropriate to the case. Training evidence should describe demonstrated competence rather than relying exclusively on attendance or time spent using a simulator.
Sources: [4]
Failure, change and continuity under control
Before operation, a project should resolve responsibility for information loss, unexpected behaviour or equipment unavailability. The editorial question is who decides and under which approved procedure, rather than providing robot recovery manoeuvres or instructions for handling radioactive material here. Pressure to finish should not become implicit permission to improvise.
Change control should identify modifications to software, tools, cameras and environmental conditions. An update may alter what the operator sees or how an instruction is executed. We propose requiring an impact review and proportionate evidence before accepting a new configuration. Preserve versions and acceptance decisions so a later investigation can reconstruct the system actually in service. This also helps distinguish a design limitation from a change that was never assessed.
Sources: [1]
Hypothetical example of a cautious assessment
A facility considers adding a teleoperated arm to an authorised task that currently requires personnel to work nearby. In this hypothetical example, the team first tests a mock-up without radioactive material. It discovers that the proposed camera obscures a relevant part of the scene and that shift handover loses information about task status.
The project adjusts observation arrangements and the handover record before requesting the relevant acceptance. It does not extrapolate one successful demonstration to all conditions or announce zero exposure. The useful trial result is identifying limitations and responsibilities early. Any subsequent step involving actual material remains subject to the facility’s safety assessment and specific procedures. The example illustrates a decision process and does not describe a real site or an operating instruction.
Safety and learning indicators
Reasonable monitoring can combine completed authorised tasks, deviations identified before use, availability of necessary functions and closed corrective actions. Radiological data should come from competent systems and professionals, with the relevant interpretations and restrictions. A commercial estimate of dose saved without a verifiable method should not replace those data.
As a management proposal, distinguish observed exposure, expected exposure and benefits that remain hypothetical. Unplanned maintenance, outstanding training and changes requiring renewed validation can also be analysed. An initial increase in reported problems may indicate better detection. Assessment should not reward concealment or compare equipment solely by speed: preparation quality and respect for authorised limits matter too. Indicators need enough context to distinguish a genuine improvement from a change in task mix.
Relationship to OSH and management documentation
This technology connects radiological protection, machinery safety, ergonomics, contractor coordination and waste management. Preventive improvement should be demonstrated across the complete work activity while preserving decision traceability. Robotisation does not turn a specialised activity into an ordinary task that anyone can perform.
A management system can organise evidence of training, risk reviews, responsibilities, versions and actions according to functions actually available. Document management should not replace safety systems or radiological monitoring. Nor does this entry attribute undocumented robotic or nuclear control functions to Sabentis. Its purpose is to support an informed discussion between prevention staff, engineering and radiological protection specialists within their respective competence. The final record should make both the intended benefit and the remaining limitations understandable to those responsible for the work.
Frequently asked questions
Does a robot eliminate radiological risk?
No. Exposure and risks across the entire task, including preparation, maintenance and removal, require assessment.
Does teleoperated mean autonomous?
No. Teleoperation retains human control; autonomy needs to be described through specific functions and boundaries.
Does a trial establish suitability for every facility?
No. Evidence must match the intended task, configuration and environment and the relevant authorised acceptance.
Does this entry authorise an intervention?
No. It explains management criteria. Interventions require competent personnel and specific approved procedures.
