FleetMind

Operating robot fleets traditionally requires complex dashboards, rigid scripts, and specialized ROS knowledge. FleetMind democratizes robotics by allowing operators to control multiple agents using simple natural language commands like "Patrol the north sector" or "Inspect the warehouse center." Our core innovation is a Hybrid AI Architecture that solves the gap between high-level reasoning and low-level control: The Brain (Gemini 2.0 Flash): Acts as the strategic commander. It parses vague human intents, resolves ambiguity (e.g., "Go back" implies memory), and outputs precise JSON mission parameters for multiple robots simultaneously. The Reflex (Reinforcement Learning): Local Q-Learning agents run on each robot to handle real-time survival. Robots autonomously decide to abort missions and seek charging stations when battery is critical—demonstrating self-optimization without hard-coded rules. Digital Twin: A high-fidelity 3D simulation (built with Three.js and React Three Fiber) visualizes the fleet in real-time, featuring simulated physics for collision detection and battery drain. FleetMind also integrates Voice Control via the WebSpeech API and simulated Telemetry streams, providing a complete "Mission Control" experience accessible from any web browser.