HIVEMIND is an advanced visual analytics platform for observing, debugging, and optimizing decentralized autonomous systems. Built on a custom high-performance physics engine, it allows researchers and developers to visualize emergent behaviors in real-time.
HIVEMIND is engineered for high-count agent simulation without sacrificing UI responsiveness.
-
Non-blocking Simulation: A dedicated Web Worker handles the
$O(1)$ spatial hashing and per-agent physics, keeping the UI thread strictly for 60fps rendering. - Spatial Partitioning: Agents are indexed using a high-density spatial hash grid, enabling real-time proximity queries for flocking, food foraging, and collision avoidance.
- Time Mastery: A 300-tick ring buffer enables bidirectional time-traveling. Step forward to predict convergence, or step backward to analyze the exact moment an anomaly occurred.
- Force Vector Overlay: Multi-colored tactical arrows visualize the specific pulls (cohesion, alignment, separation) acting on every agent.
- Voronoi Partitioning: Discrete real-time grid showing "territory" ownership and spatial dominance.
- Heatmap Coverage: Persistent trail analysis to identify gaps in swarm exploration.
- Detailed Telemetry: Click any agent to slide in an inspector panel showing velocity, heading, and algorithm-specific metadata.
- Anomaly Detection: Passive background detection identifies agents that are "stuck" or clustering pathologically, highlighting them in tactical red.
HIVEMIND is designed to model several critical autonomous workflows:
- Search & Rescue (PSO): Visualizing how a swarm of drones can collectively find a signal maximum (e.g., a heat source) in a complex environment.
- Logistics & Foraging (ACO): Optimizing pathing between a base and multiple dynamic resource points using pheromone-based stigmergy.
- Formation Control (Boids): Maintaining rigid or fluid formations through obstacle-heavy channels for coordinated movement.
HIVEMIND is the foundation for a broader suite of autonomous tools:
- AEGIS Safety Auditor: Integrating formal verification to guarantee agents stay within specific geofences.
- Multi-Robot Communication Analysis: Simulating packet loss and latency between agents to test swarm resilience.
- Mission Planning Interop: Direct export of optimized swarm paths to ROS2-compatible mission files.
npm installnpm run dev
npm run buildvercel deploy
Developed with ♥ by ThryLox