Skip to content

a major European technology museum

Museum Guide Robot Fulfilled 2,400 Tour Requests in Six Days

See how a major European technology museum's autonomous guide fulfilled 2,400 tour requests across six days, with only six left unfulfilled.

2,400
tour requests fulfilled
6
requests unfulfilled
47 hrs
operating time
18.6 km
distance traveled

Based on a documented real-world deployment. Figures are from public reporting; the organization is not named.

Visitors explore hands-on exhibits in a busy technology museum gallery.
Photo: Ray Zhu

A Gallery That Refused to Stand Still

A major European technology museum placed an autonomous tour-guide robot in direct contact with visitors. The machine had to move at roughly walking pace through dense crowds while protecting people, valuable exhibits, and itself.

The galleries were unusually difficult to read. Glass cases, metal bars, and display platforms could escape the robot's sensors, while visitors and frequently moved stools altered usable routes throughout the day.

Interaction posed an equally demanding test. Visitors arrived without technical training, chose tours at the robot, and expected immediate, intelligible responses. Remote users could also observe the museum and submit destination requests through a Web interface.

  • Navigate safely through dense, unpredictable visitor traffic
  • Recognize mapped hazards that onboard sensors might miss
  • Reroute when people or movable furnishings blocked passages
  • Make tour selection clear to visitors without prior instruction

Mapping, Motion Control and Visitor-Led Tours

The documented project centered on a purpose-built field deployment rather than a commercial procurement. An accurate metric map established the operating area, then live occupancy mapping revised that model as crowds and obstructions changed the floor.

Navigation combined probabilistic localization, path planning, live sensor data, and mapped hazard data. The robot could slow near obstacles, calculate detours, resist attempts to draw it outside its approved area, and continue toward an exhibit when conditions allowed.

The visitor interface blended text, graphics, prerecorded speech, sound, and simple physical controls. Visitors selected a tour or requested more information, while execution monitoring supplied a default action when a person made no selection. At the end of a tour, the robot returned to its designated starting position.

The paper does not report a manufacturer selection exercise, formal phased rollout, staff training curriculum, or contracted service plan. It does document advance mapping, live public operation, scheduled battery changes, and corrections to map omissions observed during deployment.

  • Prepare a bounded gallery map with known exhibits and exclusion areas
  • Update routes as crowds and movable objects change the environment
  • Translate visitor choices into navigation, narration, and exhibit-facing actions
  • Monitor execution and keep the interaction moving when input stops
A visitor uses an interactive display while exploring a museum exhibition.
Photo: Vika Glitter

Sustained Service Under Real Visitor Pressure

Across the six-day deployment, the autonomous guide operated for approximately 47 hours without significant downtime. It traveled approximately 18.6 kilometers and fulfilled 2,400 tour requests from physical and virtual visitors.

Only six requests went unfulfilled, mostly because they arrived during scheduled battery changes. The paper reports an overall success rate of 99.75 percent, unusually concrete evidence from a public-facing deployment rather than a controlled demonstration.

The operating record was strong, but not flawless. Six low-speed collisions occurred and caused no damage. Only one was attributed to a software failure; the others arose from hardware failures and omissions in the manually prepared map, which were corrected after discovery.

From Research Evidence to an Accountable Deployment

Museum staff review plans together for operating visitor services across the gallery.
Photo: cottonbro studio

This was not a Service Robot Co. client engagement, and the historical result should not be treated as a promise for another venue. It does show what a serious robot pilot program must examine: crowd behavior, hidden hazards, route recovery, visitor comprehension, battery service, and measurable request fulfillment.

Service Robot Co. brings that discipline to US businesses as a full-service, vendor neutral robot integrator. The company selects across manufacturers, then handles site assessment mapping, robot deployment and integration, financing, staff training, go live support, and ongoing service through a nationwide US engineer network.

A visitor attraction can evaluate commercial robot rental, robot leasing for business, sale, or a robot as a service structure without dividing responsibility among unrelated vendors. Remote triage, on-site dispatch, and commercial robot repair service remain with the same accountable partner across the operating lifecycle.

Frequently asked questions

The documented robot navigated dense visitor traffic by combining probabilistic localization, live obstacle sensing, mapped hazards, and dynamic route planning. Suitability still depends on the venue, so a free site assessment and controlled robot pilot program should precede wider operation.

More case studies

Find the robot that fits your site.

Free site assessment. We tell you what actually works before you spend a dollar.