University of Leeds Students Achieve Indoor AR Positioning & Navigation Breakthrough with MultiSet’s Visual Positioning System (VPS)

A team of final-year Computer Science students at the University of Leeds has successfully demonstrated a pioneering indoor augmented reality (AR) navigation solution using MultiSet’s Visual Positioning System (VPS). In their capstone project, titled “Unigator: A Smart Indoor Navigation Solution,” the students integrated MultiSet’s scan-agnostic, self-hosted VPS into their prototype and achieved the first end-to-end AR path-planning demo inside the university’s Sir William Henry Bragg Building. This milestone showcases how MultiSet’s visual positioning system can power indoor AR positioning and navigation with high precision while aligning with open-source innovation.

Student Innovation Powered by MultiSet VPS

The Unigator project set out to tackle the perennial problem of wayfinding in large campus buildings by blending LiDAR-based visual localization, Wi-Fi positioning, and an intuitive AR mobile interface. After initially experimenting with a proprietary solution (Vuforia) in an earlier prototype, the team pivoted within days – Vuforia’s high costs and vendor lock-in quickly proved incompatible with the students’ open-source, budget-conscious goals. Switching to MultiSet VPS in their Version 4 prototype delivered immediate benefits: MultiSet’s mapping pipeline yielded sub-20 cm localization accuracy without any physical markers, using only a standard iPhone’s depth sensor. This allowed the students to achieve a live AR path-planning demo throughout the Bragg Building, something never accomplished before in their department.

Critically, MultiSet’s technology fit seamlessly into the team’s existing development stack. By leveraging MultiSet’s Unity AR SDK within the Unity 2022 LTS engine, the students unified their mapping, localization, and routing logic in one cross-platform codebase. The device-agnostic design of MultiSet’s VPS meant the solution worked on mainstream Android phones and iOS devices alike, with AR Foundation ensuring a consistent experience across platforms. All components were packaged into a single mobile app capable of running offline without needing external infrastructure like beacons or QR codes. This alignment with the project’s open, infrastructure-free approach was a key factor in the successful deployment.

Faster Navigation and Fewer Errors in Live Trials

The impact of the MultiSet-powered AR positioning solution was evident in user testing on campus. In trials with over 50 participants, first-time visitors guided by the Unigator AR app reached their destinations 27.7% faster and made 81% fewer navigation errors compared to those relying on conventional static signage. These quantifiable results came from rigorous A/B evaluations, where the AR-guided group consistently outperformed the control group navigating with printed maps and wall signs. The AR interface, powered by precise visual localization, reduced user confusion and wrong turns, validating the hypothesis that live AR guidance can significantly improve indoor navigation efficiency. Participants also reported higher confidence and satisfaction, noting that the AR overlays made wayfinding clearer and more intuitive than reading floor plans or directory boards.

Notably, the visual positioning accuracy enabled by MultiSet VPS (often sub-meter in the complex environment) was a game-changer. Previous indoor navigation attempts using only Wi-Fi or smartphone GPS yielded position errors of several meters, but the MultiSet VPS provided pinpoint localization that kept virtual arrows and directions perfectly aligned to real-world locations. This precision contributed directly to the dramatic 81% drop in navigation errors observed during the trials. In practical terms, students and visitors using the AR app could find labs, lecture halls, and offices with unprecedented ease, without wandering the corridors or consulting multiple signs.

Scan-Agnostic, Self-Hosted VPS – A Key Differentiator

The success of Unigator underscores the advantages of MultiSet’s approach to indoor mapping and localization. MultiSet’s Visual Positioning System stands out as a scan-agnostic, self-hostable engine – it can ingest any 3D scan data (from smartphone LiDAR captures to point clouds) and convert it into an AR-ready map without tying users to a specific platform or hardware ecosystem. This meant the Leeds students could use their own collected scans and data, maintaining control over their mapping assets. Unlike traditional VPS offerings that often require cloud processing or proprietary formats, MultiSet’s solution can be deployed on-premises, ensuring sensitive scan data remains within the university’s infrastructure.

By eliminating cloud lock-in and recurring license fees, MultiSet provided an open, cost-effective VPS option that aligned perfectly with the project’s ethos and the broader needs of enterprise adopters seeking privacy and flexibility.

Equally important for the Leeds team was MultiSet’s robust cross-platform support. The MultiSet VPS SDK includes integrations for Unity, iOS, Android, and even WebXR, allowing developers to build AR navigation apps on the platform of their choice. The students capitalized on the Unity integration in particular, which expedited their development cycle by enabling indoor AR positioning features through familiar Unity APIs (via AR Foundation).

The pivot from a fragmented React Native/Vuforia setup to MultiSet’s Unity-based workflow was completed in one sprint, demonstrating the ease of integration and well-documented developer support. MultiSet AI even provided technical assistance to the team, including quick bug fixes and additional mapping credits for the project’s needs. This collaboration highlights MultiSet’s commitment to supporting innovators and ensuring smooth deployment in real-world scenarios.

Bridging Academia and Industry in AR Navigation With Visual Positioning System

By adopting MultiSet’s visual positioning system technology, the University of Leeds students were able to turn cutting-edge research into a working prototype that solves a tangible problem on campus. The Unigator project exemplifies how academic innovation and industry-grade tools can come together to push the envelope of indoor positioning technologies.

The outcome – a functional AR wayfinding app with measurable improvements in user performance – offers a glimpse into the future of smart campuses and workplaces. Navigating complex buildings could soon be as effortless as following a digital arrow on your phone, thanks to advancements in self-hosted VPS and AR interfaces.

MultiSet’s leadership in this domain is reinforced by the project’s results. A scan-agnostic mapping approach, sub-10 cm localization accuracy, and seamless Unity integration are not just theoretical selling points – they have now been proven in a demanding, real-world environment. For enterprises and developers, the message is clear: a self-hosted VPS like MultiSet’s can deliver reliable, high-precision indoor positioning without the downsides of proprietary platforms. Whether for university campuses, hospitals, airports, or industrial facilities, MultiSet’s technology empowers creators to build their own indoor AR navigation solutions that are vendor-neutral and scalable.

MultiSet AI is proud to have supported the University of Leeds student team in this successful demonstration of AR navigation. For more information about MultiSet’s Visual Positioning System and its applications in indoor AR navigation, visit the MultiSet website. To learn more about the University of Leeds and its Computer Science programs, visit the University of Leeds School of Computer Science page.