Autonomous Disaster-Response Drone
A twin-drone system built to fly a flood-response mission end-to-end without a hand on the stick: scan the disaster area, geotag survivors, and drop survival kits — all from a single ground control station. Built and flown with the UAV Club for NIDAR 2025, India's largest drone innovation challenge.
Overview
NIDAR's Disaster Management mission simulates a flooded coastal town where residents are stranded on rooftops with no food, water, or medicine, and rescue teams need eyes on the area before they can move in. The brief: build two drones that together scan 30 hectares for survivors, geotag every person found, and deliver 5×10×20cm, 200g survival kits — flying autonomously off one command and control station, with manual intervention limited to reloading and resuming the delivery drone.
That constraint — one C2 station, two drones, zero manual flying — was the actual engineering problem. It meant the scout drone's detections had to flow into a shared picture of the mission that both the operators and the delivery drone could act on, and the whole thing had to survive a live 30-minute mission window in front of a judging panel.
What I built
Custom ground control station
Built a ground control station on top of the ArduPilot stack using pymavlink, giving a single command-and-control setup visibility and control over both drones at once — the scout and the delivery drone reporting status, telemetry, and mission progress to one screen, which NIDAR's rules require for a fully autonomous run.
Autonomous flight & mission logic
Worked on the autopilot-side mission logic: uploading the competition's KML search boundary, triggering autonomous area scans, and handling the reload/resume flow between the scan phase and the survival-kit delivery phase without manual stick input.
Computer vision & geotagging
Worked on the detection pipeline that scans the onboard camera feed for survivors during the scout drone's pass and geotags them against GPS telemetry, turning raw flight footage into coordinates the delivery drone (and rescue teams) can act on.
Simulation before hardware
Ran autonomy logic in Gazebo before it ever touched a real airframe — cheaper, faster iteration on mission scripts and failure cases than testing on hardware first.
The team
This was fielded through the college's UAV Club, which builds and flies drones outside of coursework — NIDAR was one of a few competitions the club took on. Follow their build logs and flight tests on Instagram @uav_dr.ait.
About NIDAR
The National Innovation Challenge for Drone Application and Research (NIDAR) is India's largest drone innovation competition — run by the Drone Federation India (DFI), an industry body representing 550+ drone companies and 5,500+ pilots, under SwaYaan, MeitY's initiative to grow India's drone ecosystem toward a 2030 global-hub goal. Multidisciplinary student teams go through shortlisting and progress reviews, then a final on-campus round covering a technical design review, a startup-style business pitch, and a pre-flight inspection — before anyone's allowed to attempt the live autonomous mission.
Participation
Registered, built, and flown end-to-end for the NIDAR 2025 Disaster Management mission.
NIDAR 2025 · Drone Federation India — click to enlarge