SIET CampusNav

AI-powered smart campus navigation platform that helps students find buildings, rooms, facilities, and routes across SIET.

Comment

Inspiration

Finding a specific classroom, laboratory, department, office, or facility on a large college campus can be frustrating, especially for new students, visitors, and students attending events in unfamiliar areas.

At Sri Shakthi Institute of Engineering and Technology (SIET), we wanted to build something more useful than a traditional campus information page. Our idea was to create a smart campus navigation platform that combines interactive maps, search, route guidance, and AI-powered campus assistance in one place.

The inspiration behind SIET CampusNav came from a simple question:

“What if finding anything on campus was as easy as finding a place on Google Maps?”

This led us to design CampusNav as a map-first platform where students can search for a destination, identify their starting point, view a route, and receive step-by-step navigation instructions.

What it does

SIET CampusNav is a smart campus navigation and information platform designed for students, faculty, staff, and visitors.

The platform is designed to help users:

Search for buildings, classrooms, laboratories, departments, offices, and facilities. Find nearby campus facilities. Select a starting location and destination. View routes on an interactive campus map. Receive step-by-step navigation instructions. Support indoor navigation through building, floor, and room information where verified data is available. Use current device location for outdoor positioning where supported. Manually select a starting point when GPS is unavailable. Find accessible routes when accessibility data is available. Save frequently visited destinations. View relevant campus information and announcements. Ask an AI Campus Assistant questions using verified campus information. Access important campus information through an offline-first experience.

The goal is not to replace a normal map application, but to provide college-specific navigation and information that general-purpose maps usually do not understand.

How we built it

We designed SIET CampusNav as a modular full-stack application.

Frontend

We use:

React TypeScript Vite Tailwind CSS React Router Lucide Icons Responsive design

The interface follows a map-first design philosophy rather than presenting the user with a traditional dashboard.

The UI was designed to be:

Modern Professional Student-friendly Mobile responsive Accessible Simple to navigate Fast and visually clear Backend

The backend is designed using:

Python FastAPI Pydantic SQLAlchemy Alembic

The backend handles authentication, campus information, search, navigation requests, user preferences, AI requests, and other application services.

Database

We use:

PostgreSQL PostGIS

The database structure separates campus information into entities such as:

Campuses Buildings Floors Rooms Departments Facilities Points of Interest Entrances Navigation Nodes Navigation Edges Events Announcements Favorites Navigation History

This structure allows the system to represent the campus as a searchable and routable environment rather than simply displaying a static map.

Navigation Engine

For navigation, we model the campus as a graph consisting of nodes and connections.

The system can use:

A* pathfinding Dijkstra-based routing where appropriate Route distance calculation Estimated travel time Turn-by-turn instructions Accessibility-aware routing

The routing engine is designed around verified campus geometry rather than invented routes or coordinates.

AI Campus Assistant

The AI component is designed as an assistant that works with campus-specific information.

Instead of allowing the AI to freely generate campus facts, the system uses a grounded architecture:

User Question ↓ Intent Detection ↓ Campus Database / Navigation Engine / RAG ↓ Relevant Information ↓ AI Response ↓ Map or Navigation Action

For example, a user could ask:

“Where is the library?”

The assistant can provide the relevant verified information and, where supported, connect the answer directly to the navigation experience.

RAG Knowledge Base

We also designed a Retrieval-Augmented Generation (RAG) layer for approved campus documents such as:

Campus guides Official information Facility information Student resources Campus policies Other verified documents

This helps reduce hallucinations and keeps AI responses grounded in available campus information.

Offline-first support

The platform is designed with an offline-first approach.

Important campus information, map data, and basic navigation data can be cached locally using technologies such as:

Service Workers IndexedDB Cache API PWA architecture

Basic cached navigation can continue without connectivity when the required campus data is already available.

AI services that require cloud-based models still require an internet connection.

Challenges we ran into

One of our biggest challenges was campus data accuracy.

A navigation application cannot simply invent building coordinates, room numbers, corridors, entrances, or walking distances. Even a small incorrect route can make the application misleading.

We therefore designed a data-verification approach where campus information can be classified as:

Verified Requires verification Demo/sample data

Another challenge was combining outdoor and indoor navigation.

Normal GPS works well outdoors but cannot reliably determine a user's exact indoor position inside every building. To handle this honestly, the system supports manual starting locations and indoor map data where available, while keeping advanced indoor positioning as a future enhancement.

We also had to balance AI functionality with reliability. We did not want CampusNav to become just another chatbot that generates plausible but incorrect answers. The AI therefore acts as a supporting layer connected to the campus database, navigation system, and verified knowledge base.

Finally, designing a professional interface that works equally well on mobile and desktop required careful attention to navigation panels, map controls, search, accessibility, and responsive layouts.

Accomplishments that we're proud of

We are proud that SIET CampusNav is designed as a complete campus navigation ecosystem rather than just a map or chatbot.

Our key accomplishments include:

Designed a map-first campus navigation experience. Created a structured campus database model. Designed a graph-based navigation architecture. Planned A* pathfinding for intelligent route calculation. Designed indoor and outdoor navigation support. Added accessibility-aware routing concepts. Designed an AI Campus Assistant connected to navigation. Designed a RAG-based campus knowledge system. Included location privacy considerations. Designed an offline-first PWA architecture. Created separate student and administration experiences. Designed an administrative map-management system. Built the architecture to support future expansion of campus data.

Most importantly, we focused on transparency and accuracy instead of presenting fake routes, fake GPS accuracy, or unsupported campus information.

What we learned

Through this project, we learned that building a navigation application is much more than putting a map on a webpage.

We learned about:

Graph-based pathfinding. A* and shortest-path algorithms. Geospatial data. PostgreSQL and PostGIS. Indoor navigation concepts. Responsive UI/UX design. Progressive Web Apps. Offline-first architecture. RAG and AI grounding. API architecture. Data validation and verification. Location privacy. Designing AI systems that know when they do not have enough information.

One of the biggest lessons was that AI should not replace reliable application logic.

For CampusNav, navigation should be handled by the navigation engine, campus facts should come from verified data, and AI should help users interact with those systems naturally.

What's next for SIET CampusNav

Our next goal is to move from the core platform architecture toward a more complete real-world campus deployment.

Future improvements include:

More accurate campus mapping

We plan to build a verified digital representation of SIET's:

Buildings Floors Rooms Entrances Corridors Walkways Facilities Accessibility paths Advanced indoor positioning

Future versions could explore technologies such as:

Bluetooth Beacons Wi-Fi positioning QR-based indoor positioning Device sensors Other indoor positioning technologies Smarter AI

The AI Campus Assistant could become more capable of:

Understanding natural-language navigation requests. Recommending nearby facilities. Answering campus-related questions. Connecting answers directly to map actions. Providing contextual assistance based on the user's destination. Live campus information

We plan to integrate:

Temporary route closures Events Room changes Announcements Facility availability Campus alerts Accessibility

We want to improve accessibility routing by providing better support for:

Wheelchair users Elevator routes Ramps Accessible entrances Reduced-stair routes

Built With

Share this project:

Updates

Submission history