Maze Safety Classification & Path Planning (AI-based Pathfinding using Perceptron & A* Algorithm)

AI-based web application designed for safe path planning in maze environments.

The system uses a Perceptron model to classify safe and unsafe zones and an A* algorithm to find the optimal and safest path from start to goal.

The project includes:

Interactive maze generation with adjustable size and terrain distribution (Grass, Water, Obstacles).

Real-time pathfinding visualization (start, end, frontier, visited, and final path).

Custom dataset evaluation for model performance (safe vs unsafe samples).

Editable environment and reset functionality for testing multiple scenarios.

Built with HTML, CSS, JavaScript, and deployed on Netlify.