MuMMER · Project

Autonomous Social Robot That Engages Shoppers in Real Malls

digitalPilotedTRL 6

Imagine walking into a shopping mall and being greeted by a friendly robot that can tell you jokes, play games, give directions, and even collect your feedback about the stores. That's what MuMMER built — a Pepper robot that can see, hear, and navigate around crowds on its own, holding natural conversations with real people. The team spent four years teaching it to read social cues, move safely around shoppers, and behave in ways people actually enjoy rather than find creepy. They tested it in a real Finnish shopping mall over an extended period, not just in a lab.

By the numbers

EUR 4,297,835

EU research investment in social robot technology

consortium partners across research and industry

countries involved (CH, FI, FR, UK)

total deliverables produced

integrated robot system demos (initial + final)

29%

industry partner ratio in consortium

The business problem

What needed solving

Shopping malls and public venues are losing foot traffic to online retail and need new ways to engage visitors, collect real-time feedback, and create memorable experiences. Hiring staff for entertainment and customer interaction is expensive and inconsistent. An autonomous social robot that can entertain, guide, and gather feedback without breaks or bad days addresses both the experience gap and the staffing challenge.

The solution

What was built

The team built a fully integrated autonomous social robot system on the Pepper platform, capable of natural speech interaction, non-verbal communication, social-signal processing, audiovisual scene understanding, and human-aware navigation in crowded spaces. They delivered 31 project outputs including an initial and a final integrated robot system, both demonstrated in a real Finnish shopping mall.

Audience

Who needs this

Shopping mall operators and retail property management companiesHotel chains and airport authorities looking for automated guest engagementRobotics system integrators building commercial service robotsTheme parks and entertainment venues wanting interactive robot guidesLarge retailers exploring in-store customer experience automation

Business applications

Who can put this to work

Retail & Shopping Centres

enterprise

Target: Shopping mall operators and retail property managers

If you are a shopping mall operator dealing with declining foot traffic and bland in-store experiences — this project developed an autonomous humanoid robot system that engages shoppers with entertainment, wayfinding, and feedback collection. Tested over a long-term deployment in a real Finnish mall with 7 partner organizations, the robot combines speech, gesture, and safe navigation in crowded spaces.

Hospitality & Tourism

enterprise

Target: Hotel chains and airport operators

If you are a hospitality company struggling with front-desk staffing and guest engagement — this project built a robot that handles natural conversation, provides guidance, and reads social signals to adjust its behaviour. The system was designed for dynamic public environments, meaning it already handles the unpredictability of real crowds rather than controlled lab settings.

Consumer Robotics & System Integration

any

Target: Robotics integrators and service robot manufacturers

If you are a robotics company looking to add social intelligence to your platforms — this project produced 31 deliverables covering audiovisual scene processing, social-signal processing, human-aware navigation, and high-level action selection. The full stack was integrated on the Pepper platform, giving you a proven reference architecture for socially aware service robots.

Frequently asked

Quick answers

What would it cost to deploy a similar robot in my venue?

The full research programme cost EUR 4,297,835 across 7 partners over 4 years, covering fundamental research and integration. A commercial deployment would be significantly less since the core software is developed, but hardware (Pepper platform), customisation, and maintenance costs would depend on your venue size and use case. Based on available project data, no per-unit pricing was published.

Can this scale beyond a single mall?

The system was designed for dynamic public environments with unpredictable crowds, not a single controlled setting. The architecture separates perception, social reasoning, and navigation into modules, which supports adaptation to different venues. However, the long-term field study was conducted in one Finnish mall, so multi-site deployment would require further validation.

Who owns the IP and can I license it?

The consortium of 7 partners across 4 countries (CH, FI, FR, UK) developed the technology under a Horizon 2020 RIA grant, meaning IP is shared among partners per their consortium agreement. The University of Glasgow coordinated. Commercial licensing would need to be negotiated with the relevant partners — particularly those holding IP on specific modules like navigation or social-signal processing.

How does the robot handle real-world safety and regulations?

The project specifically developed human-aware navigation so the robot moves safely among crowds. It was deployed in a real public shopping mall in Finland, meaning it met local safety requirements for public-space robotics. EU robotics regulations have evolved since the project ended in 2020, so current compliance would need reassessment.

How long did it take from concept to working system?

The project ran from March 2016 to February 2020 — a full 4-year timeline. An initial integrated robot system was delivered first, followed by the final integrated robot system. For a commercial deployment building on this work, the integration timeline would be shorter since the core research is complete.

Does it only work with the Pepper robot?

The project was built on Aldebaran's Pepper platform specifically. The software modules — audiovisual processing, social-signal processing, action selection, and navigation — are in principle adaptable to other humanoid platforms, but based on available project data, no multi-platform testing was reported.

Consortium

Who built it

The MuMMER consortium brings together 7 partners from 4 countries (Switzerland, Finland, France, UK), with a balanced mix of 2 universities, 3 research organisations, and 2 industry partners (29% industry ratio). The University of Glasgow led the coordination. Notably, there are zero SMEs in the consortium, which suggests the technology was developed in a research-heavy environment. The 2 industry partners likely contributed the robotics platform and deployment expertise. For a business considering this technology, the academic-heavy consortium means strong scientific foundations but you would need a systems integrator to bridge the gap to commercial deployment.

UNIVERSITY OF GLASGOWCoordinator · UK
TEKNOLOGIAN TUTKIMUSKESKUS VTT OYparticipant · FI
FONDATION DE L'INSTITUT DE RECHERCHE IDIAPparticipant · CH
HERIOT-WATT UNIVERSITYparticipant · UK
CENTRE NATIONAL DE LA RECHERCHE SCIENTIFIQUE CNRSparticipant · FR
ALDEBARANparticipant · FR

How to reach the team

University of Glasgow, United Kingdom — the coordinating institution for all MuMMER IP and partnership enquiries

Order a report on this consortium See UNIVERSITY OF GLASGOW profile →

Next steps

Talk to the team behind this work.

Want to explore how social robot technology from MuMMER could work in your venue? SciTransfer can connect you with the right research partners and help evaluate fit for your business case.

Order a report on this topic More in Digital & ICT