Inspiration
Search and Rescue was inspired by the challenge of emergency response in complex buildings where every second matters. We wanted to create a system that could do more than just visualize a floor plan. Our goal was to build a platform that helps users think through how first responders move, how hazards spread, and how rescue outcomes change based on layout and decision-making. As engineering students, we were especially motivated by the idea of combining spatial design, simulation, and optimization into a tool that could have real-world safety applications.
What it does
Search and Rescue is an AI-powered emergency response simulation platform for planning and analysis. Users can design multi-floor buildings in 2D and 3D, define rescue scenarios, place responders and victims, and run simulations that model how emergency teams navigate a building. The system uses optimization and pathfinding techniques to assign responders and generate efficient routes while accounting for dynamic conditions such as fire or hostile threats. It is designed to help users evaluate layouts, improve preparedness, and better understand emergency response behavior before a real disaster happens.
How we built it
We built the platform as a full-stack system with a Python FastAPI backend, a React frontend, and a Three.js based 3D renderer. The building editor supports floor plan creation with tools for walls, rooms, materials, and object placement, while the scenario builder allows users to define emergency conditions and simulation inputs. On the backend, we implemented routing and assignment logic using A* pathfinding and Hungarian algorithm based matching to connect responders with tasks efficiently. We also developed a JSON-based workflow for importing, exporting, and passing building and scenario data between the editor, simulation engine, and replay system.
Challenges we ran into
One of the biggest challenges was integrating several very different systems into one seamless workflow. The building editor, scenario builder, simulation engine, and 3D visualization all had to share consistent geometry and state data. Another challenge was balancing realism with performance, especially when handling multi-floor environments and dynamic hazards. We also had to think carefully about how to represent emergency logic in a way that was both computationally efficient and visually understandable to the user. On the frontend side, keeping the interface intuitive while supporting many tools and simulation features was an ongoing design challenge.
Accomplishments that we're proud of
We are proud that we built a platform that brings together building design, scenario creation, simulation, and replay in one unified system. Instead of treating emergency planning as a static map problem, Search and Rescue allows users to actively test and visualize response strategies. We are also proud of the technical integration between the 2D editor, 3D viewer, and backend simulation logic, as well as the use of optimization algorithms to make routing decisions more intelligent. Most importantly, we created a project that feels meaningful and applicable to real emergency preparedness problems.
What we learned
This project taught us a lot about full-stack system design, simulation architecture, and spatial data modeling. We learned how difficult it is to keep geometry, user inputs, and simulation state synchronized across multiple application layers. We also gained experience applying algorithms like A* and the Hungarian method in a practical context rather than only in theory. Beyond the technical side, we learned how important usability is in simulation software. Even strong algorithms are only valuable if users can understand, trust, and interact with the results effectively.
What's next for Search and Rescue
Next, we want to make Search and Rescue more realistic, scalable, and collaborative. That includes improving hazard propagation models, adding more responder behaviors, and supporting richer building semantics such as stairwells, blocked exits, and varying mobility constraints. We also want to expand analytics so users can compare strategies across multiple runs and better quantify performance. In the long term, we would like to explore real-world applications in safety training, building design review, and emergency preparedness planning, where simulation can help teams make faster and smarter decisions before lives are at risk.
Log in or sign up for Devpost to join the conversation.