NETRA

NETRA AI

Autonomous Swarm Intelligence

Perception

YOLOv8 Vision

Real-time threat & incident detection pipeline

Telemetry

Full Situational Awareness

Low-latency data flow.

Navigation

A* Pathfinding

Dynamic global routing & NFZ avoidance

Coordination

Swarm Logic

Autonomous incident dispatch & scaling

Launch Simulation View Architecture

64 incidents resolved·avg response 1.4 min·8 drones active·12.3 km² coverage64 incidents detected·avg response 1.4 min·8 drones active·12.3 km² coverage

Incidents Detected

Drones Deployed

0.0m

Avg Response

0.0 km²

Coverage Area

Autonomous Protocol

PHASE 01

Detection

Computer vision identifies incidents across CCTV feeds in real time, filtering noise with high precision confidence models.

PHASE 02

Dispatch

The nearest available drone is autonomously assigned and routed, dynamically avoiding no-fly zones through A* pathfinding.

PHASE 03

Resolution

Live feed, telemetry, and on-scene response—all securely logged and managed autonomously for real-time situational awareness.

Intelligence at the Edge

NETRA's architecture is built for continuous, low-latency operations. It processes raw CCTV feeds using YOLOv8, maps detections to coordinates, and broadcasts dispatch payloads via Server-Sent Events—all asynchronously.

Next.jsTypeScriptPythonYOLOv8LeafletZustandFastAPISSE

POST /api/ingest

Explainable AI

Transparent Matrix

Real-time decision logs surfacing exact mathematical dispatch weights

Smart Charging

Automated RTB

Self-regulating fleet with autonomous Pod routing logic

Custom Engine

High-Fidelity Physics

Custom TypeScript physics and environment state logic

Built by Team Phoenix · COEP Technological University · CSI Hackathon 2026