Mobile inspection robotics is worthwhile when an inspection round is repeated frequently, is demanding or dangerous for people, and the inspection points are stable enough to be taught in once. The robot itself is rarely the largest cost item – and the most frequent project delays are caused not by technology, but by evidence of conformity, works council co-determination and structural prerequisites at the site.
This guide takes you through the decision: from the question of whether the deployment pays off at all, to a realistic project schedule.
What an inspection robot does
An inspection robot follows a defined route, stops at defined points, aligns its sensors, records measured values and images and makes them available in structured form. It does this with high repeatability, documented and independently of time of day and weather.
The real benefit lies in repeatability. A person measures at slightly different points, from slightly different angles, with varying care. A robot produces series of measurements that are comparable over months. Only this makes gradual changes visible.
When is the application worthwhile?
Three conditions must come together.
Frequency of repetition. The inspection round takes place at least weekly, preferably daily. With quarterly inspections, the set-up effort does not pay back.
Demands on staff or hazard. The inspection round is physically strenuous, takes place under unfavourable conditions, requires safety measures or ties up two people. The more effort manual execution involves, the better the calculation.
Stable environment. Routes and inspection points stay the same for months.
Short check
If you answer these questions mostly with yes, a closer look is worthwhile:
- Is the same inspection round carried out at least weekly?
- Does it take longer than 30 minutes?
- Does it require safety measures or two people?
- Is the site rarely staffed, so that every inspection means a journey there?
- Would more frequent inspections make sense technically, but fail for lack of staff?
- Do the inspection points stay unchanged for months?
- Does the area remain structurally largely the same?
The calculation is particularly favourable at unmanned sites. There, not only the inspection round is eliminated, but also the journey.
Which design fits?
Legged systems are not always the right choice. An honest classification:
| Design | Strength | Limit |
|---|---|---|
| Legged system, four-legged | stairs, gratings, uneven ground, several levels | higher costs, limited payload, Ex approval only in exceptional cases |
| Wheel-leg hybrid | fast on level stretches, yet able to climb steps | larger turning circle, less mature products |
| Tracked vehicle | robust, high payload, Ex approvals available | stairs only to a limited extent, level terrain needed |
| Driverless vehicle | established, inexpensive, high payload | only level, continuous surfaces |
| Drone | height, façades, high-bay racks, large areas | flight time, indoor navigation, aviation law requirements |
| Stationary sensors | continuous recording of individual points, inexpensive | no area coverage, cabling |
Table scrolls sideways.
Rule of thumb: If your plant is level throughout and without steps, look at driverless vehicles first. If the inspection round covers only a few fixed measuring points, look at stationary sensors. Legged systems play to their strengths as soon as stairs, gratings or several levels are in the way.
Levels of autonomy
The most important technical difference between systems is not the payload, but the degree of autonomy.
Remotely operated. A person operates the device throughout. Staff effort hardly falls – the benefit lies solely in nobody having to enter the hazardous area.
Assisted. The system keeps its distance and avoids obstacles on its own; a person still specifies the route.
Automated inspection walk on a defined route. The system follows a previously taught route on its own and returns to the charging station. This is the level from which a real economic benefit arises.
With event response. The system responds to anomalies and adapts its sequence. Technically more demanding and involves considerable regulatory effort, because its behaviour can be predicted less completely.
For procurement, this means: do not check whether a system is “autonomous”, but what exactly happens on its own – and what in practice still requires an operator.
Sensors by inspection task
The sensors are often a larger cost item than the platform. They should be derived from today's inspection scope, not from the product catalogue.
| Inspection task | Sensors |
|---|---|
| Meter readings, pressure gauges, labels | optical camera with optical zoom |
| Overheated terminal points, bearing damage, temperature distribution | thermal imaging camera, radiometric |
| Leaks, oxygen deficiency, hydrogen sulphide | multi-gas sensors |
| Incipient bearing damage, partial discharge, compressed-air leaks | acoustic methods |
| Structural condition, target/actual comparison | LiDAR scanner |
| Documentation of the inspection round | camera with position data |
Table scrolls sideways.
Check the interface: if it is openly documented, sensors can be added or replaced later. If it is closed, you are tied to the manufacturer's accessories programme.
Bear in mind: more extensive retrofits can legally count as a modification of the machine – with consequences for responsibility. More in CE conformity in inspection robotics – a guide for operators and importers.
What belongs to the system
The robot is the visible part. The complete system also includes:
- sensor payload including mounting and cabling
- charging station and its structural preparation
- software for set-up, analysis and multi-device management
- network connection or deliberately chosen isolated operation
- spare batteries and wear parts
- training for operation and set-up
- service contract with a defined response time
- operating concept: who sets up, who analyses, who decides
The last point decides success. A system without a named responsible person produces data that nobody looks at.
Prerequisites at the site
Charging point. Level ground, power connection, clearance all round, often anchoring in the floor. A roof is frequently needed, because charging stations are less well protected than the robot. With systems that dock using reflectors, existing safety markings can interfere.
Routes. Passage widths, doors and gates, stair geometry, ramp transitions, floor condition. Open grating stairs are more demanding than closed ones.
Environment. Temperature range, dust, moisture, corrosion. Check the upper limit critically – manufacturer specifications refer to the ambient air; radiant heat acts in addition.
Network. Radio coverage along the route or isolated operation. In tunnels and basements, both need to be clarified.
Ex zones. Does the planned route lie wholly or partly in a zone-classified area? What counts is the explosion protection document of the plant.
Four hurdles before commissioning
These four points delay projects more often than any technical question.
Conformity. In the EU, an inspection robot may only be placed on the market with CE marking and an EU declaration of conformity. Ask for the declaration before concluding the contract and check which legal acts it names. → CE conformity in robotics
Works council co-determination and data protection. In companies with a works council, the introduction of a camera-equipped system is subject to enforceable works council co-determination. Without a works agreement, the device does not go into operation. → Data protection and works council co-determination
Explosion protection. The IP rating is not an Ex approval. Only a few appropriately certified systems are permitted for zone-classified areas. Retrofitting is ruled out. → Explosion protection and mobile robotics
Dangerous goods. The batteries are Class 9 dangerous goods. This does not affect operation, but it does affect delivery, shipping of spare parts, storage and return shipment in the event of repair.
Cost-effectiveness at a glance
The calculation rarely works out on staff savings alone. More robust are:
- avoided exposure to hazards
- more frequent inspection intervals with the same staff deployment
- comparable, documented series of measurements
- anomalies detected earlier and unplanned downtime avoided
- journeys no longer needed at unmanned sites
In detail in Cost-effectiveness of inspection robots – purchase, leasing or Robot-as-a-Service.
Realistic project schedule
| Phase | Duration |
|---|---|
| Needs assessment, analysis of inspection rounds | 2–4 weeks |
| Supplier selection, demonstration, trial operation | 4–8 weeks |
| Data protection and works agreement (in parallel) | 8–12 weeks |
| Procurement and delivery time | 8–16 weeks |
| Structural preparation of the charging point (in parallel) | 2–6 weeks |
| Installation, teaching in the routes | 2–4 weeks |
| Trial operation to regular operation | 4–8 weeks |
Table scrolls sideways.
From decision to regular operation: six to twelve months. With direct import, rather at the upper end, because delivery times and clarification of conformity are added.
The phases running in parallel are the lever. Anyone who starts on data protection and the works agreement only after delivery extends the project by months.
Common mistakes
The device first, the use case afterwards. A demonstration impresses, but it does not answer which inspection round is to be transferred. Start with the analysis.
Involving the works council late. The most expensive mistake in the entire project – and the most avoidable.
Overestimating autonomy. Ask what actually happened on its own in the demonstration and what was remotely operated.
Forgetting the charging point. Structural preparation needs lead time and sometimes a permit.
Calculating only the purchase price. The follow-up costs over five years regularly exceed the price of the device.
Not checking Ex zones. The question will come from the occupational safety specialist at the latest – better beforehand.
Trialling without a scaling perspective. A single device that covers only one inspection round rarely pays off. Clarify early whether and how the deployment is to be expanded.
Checking site prerequisites
- Charging point: level ground, power connection, clearance, anchoring possible
- Roof available or can be provided
- No interfering reflectors in the docking area
- Passage widths, doors, stair geometry recorded
- Floor condition along the planned route assessed
- Temperature range and ambient conditions checked
- Radio coverage along the route or isolated operation clarified
- Zone classification checked against the explosion protection document
- Responsible person for set-up and analysis named
This checklist will be offered as a download once the template exists.
Sources
- Regulation (EU) 2023/1230 on machinery European Union accessed 15 August 2026
- Regulation (EU) 2016/679 (General Data Protection Regulation) European Union accessed 15 August 2026
- Betriebsverfassungsgesetz Bundesministerium der Justiz accessed 15 August 2026
Change log (1)
- 5 October 2026 Initial publication First publication of the English version.