ManaKrishi

Real-time drone telemetry and field insights for agriculture

WebSocketsAWS EC2Node.jsReactMQTTIoTFirebase

The Problem

Drone-based crop spraying gave operators no live picture of what was happening in the field. Telemetry arrived after the fact, so a bad run could not be corrected while it was still running — and the farmer records and IoT streams behind it had no network isolation.

The Solution

A real-time telemetry backend that streams flight data from drone transponders straight to operator dashboards over WebSockets, running on AWS EC2 inside a multi-tier VPC that keeps farmer records and device traffic on isolated subnets.

Key Features

Live drone telemetry streamed to operator dashboards over WebSockets
Multi-tier VPC with isolated subnets and least-privilege IAM for farmer and device data
IoT sensor ingestion for soil moisture and field conditions
Drone survey imagery for crop health monitoring
Market price tracking and weather-aware advisories for farmers
Low-bandwidth UI designed for rural connectivity

Architecture

Node.js WebSocket service on AWS EC2 inside a multi-tier VPC — public subnet for the load balancer, private subnets for application and data tiers, with security groups and least-privilege IAM between them. Drone transponders and IoT sensors publish over MQTT to a lightweight broker; a React frontend renders live telemetry, with Firebase handling auth and Firestore backing the farmer-facing data.

Results

Operators see flight data live instead of reviewing it afterwards, and the platform piloted with a local farming community where farmers checked prices before selling.

Timeline

Ongoing — built with a small team

Lessons Learned

· Streaming telemetry is a network problem before it is a UI problem — subnet layout and security groups shaped the design more than the dashboard did.
· Designing for low bandwidth changes every decision — payload size, offline caching, and font choices all matter.

Future Improvements

· AI crop disease detection from phone photos
· Voice-first interface in regional languages

Next project

This Portfolio