A Layered Spatial Graph Framework for Intelligent Multi-Floor Indoor Navigation

Finding your way inside a massive, multi-story building is a different kind of puzzle than navigating a city street. Outside, the Global Positioning System, or GPS, acts as a reliable guide, bouncing signals off satellites to tell us exactly where we stand. But those same signals struggle to penetrate the thick walls, concrete ceilings, and dense materials of modern architecture. Once inside a large university campus, a hospital, or a shopping mall, the satellite connection fades, leaving people to rely on static signs and their own memory. This is particularly difficult in complex structures with dozens of rooms on multiple levels, where a wrong turn at a staircase can lead to a significant delay before a lecture or meeting. The challenge for scientists and engineers has been to create a digital guide that works without needing expensive sensors or specialized hardware installed throughout the building.

A team of researchers at Vellore Institute of Technology has addressed this problem by creating a web-based system called INDOORNAV. Instead of trying to track a person's location with radio waves or Bluetooth beacons, their approach focuses on the building itself. They treat the floor plans as a digital map that can be read by a computer, turning rooms, corridors, and staircases into a connected network. By using a well-known method for finding the quickest route through a network, the system calculates the shortest path between any two points and displays it clearly on a screen. The result is a tool that anyone can access through a standard web browser, requiring no app download, to guide them from their starting room to their destination, even if they are moving between different floors.

The core of this work lies in how the researchers translated a physical building into a language a computer understands. They began with high-resolution architectural blueprints of a three-story building containing over thirty rooms per floor. Using digital mapping tools, they traced the outlines of every room, corridor, and doorway, converting these shapes into a standard digital format known as GeoJSON. This format acts like a universal translator for geographic data, allowing the computer to see the building not just as an image, but as a collection of shapes and connections. From this digital blueprint, the team constructed a graph, which is essentially a map of connections where every room and intersection is a point, and every walkable path between them is a line.

To find the best route, the system uses a search method called the A* algorithm. Imagine standing at a crossroads with a map and a destination in mind; you naturally look for the path that seems to lead most directly toward your goal. The A* algorithm does something similar but with mathematical precision. It looks at the distance already traveled and estimates the remaining distance to the goal, constantly choosing the path that appears most efficient. This allows the system to ignore dead ends and unnecessary detours, quickly identifying the optimal route. Unlike older methods that might check every single possible path, this approach is faster and uses less computing power, making it ideal for real-time use on standard devices.

The researchers tested their system extensively to ensure it worked reliably. They ran two hundred different navigation scenarios, including trips that stayed on a single floor, journeys that required moving up or down between floors, and emergency routes to the nearest exit. In every single test, the system successfully calculated a valid path. The speed of the response was remarkably fast, with the system finding routes in less than a millisecond on average. For emergency situations, where every second counts, the system was even quicker, finding the path to an exit in roughly one-third of a millisecond. The system also proved to be very light on memory, using a tiny fraction of the storage space available on a typical computer, which means it could run smoothly on almost any device without slowing it down.

One of the most practical features of this system is its ability to handle vertical movement. In many indoor navigation attempts, moving between floors is a complex problem, but this system treats staircases and elevators as simple connection points between the different layers of the building. When a user needs to go from a room on the first floor to a room on the third, the system automatically identifies the nearest staircase or elevator, calculates the path to it, and then continues the calculation on the next floor. This seamless integration means that a user does not need to switch maps or change their approach when changing levels; the system handles the entire journey as one continuous path.

The design of INDOORNAV also prioritizes accessibility and ease of use. Because the system runs entirely in a web browser, users do not need to download or install any software. They simply open a website, type in their starting room and their destination, and the system generates a clear, highlighted map showing the route. This map is generated instantly and can be viewed on a smartphone, tablet, or laptop. The researchers demonstrated that this approach works effectively for a three-story building, but the underlying structure is designed to be scalable. If a new wing is added to a building or a new floor is constructed, the system can be updated by simply adding the new digital floor plan to the network, without needing to rebuild the entire system from scratch.

Beyond everyday navigation, the system includes a specific mode for emergency evacuation. In a crisis, such as a fire, finding the quickest way out is critical. The system allows a user to input "exit" as a destination, and it immediately calculates the shortest path to the nearest emergency door, bypassing any internal obstacles or complex room layouts. This feature is built into the same framework used for daily navigation, ensuring that the safety instructions are consistent and reliable. The researchers found that this emergency routing was the fastest type of calculation the system performed, highlighting its potential value in critical situations where speed is paramount.

The success of this project demonstrates that complex indoor navigation does not require a network of expensive sensors or specialized hardware. By leveraging existing digital floor plans and efficient pathfinding algorithms, it is possible to create a robust guidance system that is both cost-effective and easy to deploy. The system proved that a web-based approach could handle the complexities of multi-floor buildings, offering a solution that is accessible to students, faculty, and visitors alike. While the current version operates on a static map, the researchers have outlined future possibilities, such as integrating voice commands or augmented reality to make the experience even more intuitive. For now, however, the system stands as a functional, reliable tool that turns the confusion of a large, unfamiliar building into a clear, guided path.