From what to build to how to build it — our platform helps you understand, plan, and solve SIH 2026 problem statements with complete insights into requirements, solution approaches, technology stacks, architecture, AI/ML, and implementation roadmaps.
Search for your problem statement, and start building your solution.
| Description: | Student Innovation-There is a need to design drones and robots that can solve some of the pressing challenges of India such as handling medical emergencies, search and rescue operations, etc. |
| Organization: | AICTE |
| Department: | AICTE, MIC-Student Innovation |
| Category: | Software |
| Theme: | Robotics and Dronestd> |
| Youtube Link: | Not provided |
| Dataset Link: | Not provided |
| Contact info: | Not provided |
This research and architectural dossier establishes an end-to-end blueprint for "GarudaResQ"—a Software-First Autonomous Drone & Robot Rescue Orchestration Platform [INFERENCE] designed to address India's urgent medical emergency logistics and disaster Search & Rescue (SAR) challenges [PS]. While the problem statement highlights hardware deployment in complex Indian topographies [PS], the competition entry is officially categorized under Software [PS]. GarudaResQ bridges this gap by delivering a unified, edge-resilient Mission Control & Swarm Dispatch System [INFERENCE] capable of interfacing directly with open-source autopilots (PX4/ArduPilot via MAVLink and ROS2) [KNOWN] while executing full Software-In-The-Loop (SITL) simulations [KNOWN] for live judging. The platform integrates autonomous lawnmower/spiral search path planning [KNOWN], edge-based AI survivor detection using thermal/optical feeds [RESEARCHED], payload-drop stability monitoring for emergency medical delivery [RESEARCHED], and an offline-first mesh synchronization engine for zero-connectivity disaster zones [INFERENCE]. Operating strictly within the Indian regulatory framework mandated by the DGCA Drone Rules 2021 and DigitalSky airspace maps [RESEARCHED], this document provides a complete technical design, data schema, risk assessment, and hackathon roadmap targeted for completion before 30 September 2026 [PS].
| Metadata Field | Details |
|---|---|
| Problem Statement ID | SIH26201 [PS] |
| Title | Student Innovation - Drones and Robots for Medical Emergencies and Search & Rescue [PS] |
| Organization | All India Council for Technical Education (AICTE) [PS] |
| Department | AICTE, MIC-Student Innovation [PS] |
| Category | Software [PS] |
| Theme | Robotics and Drones [PS] |
| Submission Deadline | 30 September 2026 [PS] |
| Ideas Counter | 0 / 500 [PS] |
| Fundamental Objective | Build a tactical, offline-resilient GCS software platform that orchestrates autonomous aerial/ground robotic swarms for rapid medical payload delivery and AI-driven thermal survivor detection under DGCA DigitalSky compliance [INFERENCE]. |
The explicit problem statement provided in the source JSON reads: "Student Innovation-There is a need to design drones and robots that can solve some of the pressing challenges of India such as handling medical emergencies, search and rescue operations, etc." [PS]
To make this actionable for a software-track student hackathon team, the requirement is deconstructed into atomic functional units, explicitly distinguishing between stated requirements and derived operational implications:
"category": "Software" [PS]| Stakeholder Group | Primary Goal | Current Pain Point | Core System Need |
|---|---|---|---|
| NDRF / SDRF Disaster Response Commanders [KNOWN] | Rapidly survey collapse/flood zones and locate survivors within the critical "golden hour" [KNOWN]. | Ground teams are delayed by blocked roads, debris, flooded terrain, and lack of real-time situational maps [RESEARCHED]. | Unified GIS Dashboard displaying real-time SAR heatmaps, survivor GPS coordinates, and autonomous multi-drone coverage progress [INFERENCE]. |
| Emergency Medical Response Officers (108 ERSO / PHC Doctors) [RESEARCHED] | Transport urgent medical supplies (anti-venom, blood, organs, vaccines) to cut-off rural or disaster-hit primary health centres [RESEARCHED]. | Ambulances locked in traffic or delayed by damaged roads; sample/blood degradation in transit [RESEARCHED]. | One-click medical dispatch workflow with live thermal/payload temperature telemetry and automated safe-drop zone selection [INFERENCE]. |
| Field Drone Pilots & Mobile Operations Operators [KNOWN] | Safely fly search-and-rescue sorties under stress without risking vehicle collisions or airspace violations [KNOWN]. | Manual flight control of multiple drones is prone to pilot fatigue, loss of signal, and airspace violations [INFERENCE]. | Autonomous waypoint generation, one-key auto-takeoff/return-to-launch (RTL), and dynamic obstacle/no-fly zone avoidance [INFERENCE]. |
| Ground Search Units (First Responders on Foot) [KNOWN] | Navigate directly to trapped victims with necessary medical aid [KNOWN]. | Inefficient manual grid searches in dangerous environments without precise coordinates [INFERENCE]. | Mobile companion app with offline tactical map caching victim location pins and optimal ground walk paths [INFERENCE]. |
| District Disaster Management Authorities (DDMA) [RESEARCHED] | Ensure all disaster operations strictly comply with national civil aviation laws and airspace safety [RESEARCHED]. | Risk of unauthorized flights near military zones, airports, or hazardous power lines during relief operations [RESEARCHED]. | Automated DGCA DigitalSky Airspace map integration (Green/Yellow/Red zone verification) and flight log auditing [INFERENCE]. |
| Feature / Module | Addressed Requirement | Technical Justification |
|---|---|---|
| Autonomous Lawnmower & Spiral Path Planner | REQ-02 (SAR Operations) [PS] | Guarantees 100% spatial coverage over search polygons without manual pilot intervention [KNOWN]. |
| Edge YOLOv8/v10 Thermal & RGB Detection Engine | REQ-02 (SAR) & REQ-05 (Indian Realities) [PS] | Processes video streams at 30+ FPS directly on companion computers (Jetson/Pi 5) to identify human thermal signatures through canopy or smoke [RESEARCHED]. |
| Medical Payload Drop & Thermal Monitoring Module | REQ-01 (Medical Emergencies) [PS] | Tracks real-time payload container temperature and actuates Servo release mechanisms upon reaching verified drop coordinates [RESEARCHED]. |
| Heterogeneous Multi-Agent ROS2 Bridge (PX4/ArduPilot) | REQ-03 (Drones & Robots) [PS] | Enables simultaneous telemetry ingestion, state estimation, and command translation for both aerial quadcopters and ground UGVs [KNOWN]. |
| PX4 / Gazebo SITL Multi-Vehicle Simulator | REQ-04 (Software Track Execution) [PS] | Allows full end-to-end evaluation and live demonstration of swarm SAR and medical dispatch without needing live physical airspaces during judging [INFERENCE]. |
| Off-Grid LoRa / Wi-Fi Mesh Peer-to-Peer Sync | REQ-05 (Indian Realities) [PS] | Ensures tactical telemetry and victim location data are synced between field workers and command tents when 4G/5G towers are down [INFERENCE]. |
| DigitalSky Airspace Geo-Fence Checker | REQ-05 (Indian Realities & DGCA Rules) [PS] | Parses official Indian airspace XML/JSON geo-json data to restrict flight paths within permitted Green zones or warn on Yellow/Red zone intersections [RESEARCHED]. |
Developing drone and robotic rescue software for Indian field conditions introduces distinct operational realities that dictate system architecture:
+-----------------------------------------------------------------------------------+
| INDIAN FIELD CONSTRAINTS |
+-----------------------------------------------------------------------------------+
| Connectivity : 80% cellular outage in disaster zones (floods, landslides, quakes)|
| Geography : High-altitude low air density (Himalayas) vs. Dense jungle canopy |
| Battery Life : Micro-drones restricted to 20-30 min flight times |
| Regulations : DGCA Drone Rules 2021, DigitalSky Green/Yellow/Red airspace limits |
| User Base : Non-technical field rescue workers requiring simple touch UI |
+-----------------------------------------------------------------------------------+
To ensure technical viability, the student team must investigate the following core research questions:
| Data Type | Operational Purpose | Ideal Real-World Source | Hackathon Substitute | Access Method | Availability | Privacy Sensitivity | Confidence |
|---|---|---|---|---|---|---|---|
| Airspace Boundaries & Geo-Fences | Prevent illegal entry into airport/military zones. | DGCA DigitalSky Portal [RESEARCHED] | Synthetic GeoJSON matching Indian Green/Yellow/Red zone schemas. | REST API / Static File Ingestion | Static files available publicly; API restricted [RESEARCHED] | Public domain regulatory data [KNOWN] | [RESEARCHED] |
| High-Resolution Elevation Models (DEM) | Terrain-following flight path generation in valleys/mountains. | ISRO Bhuvan / USGS SRTM 30m DEM [RESEARCHED] | NASA SRTM 30m open raster tile datasets. | HTTP Download / GeoTIFF parsing | Fully open access [KNOWN] | Non-sensitive geospatial data [KNOWN] | [KNOWN] |
| Thermal & RGB Human Detection Imagery | Training/testing AI models for survivor detection. | Live FLIR Lepton / Boson sensor streams [KNOWN] | OpenDatasets (FLIR Thermal Dataset, SARTAG SAR UAV Dataset) [RESEARCHED] | Kaggle / Roboflow Direct Download | Public academic datasets available [RESEARCHED] | Anonymized thermal profiles; low privacy risk [INFERENCE] | [RESEARCHED] |
| Vehicle Flight Telemetry | Real-time position, attitude, battery, and GPS health. | Live Pixhawk 4 / PX4 MAVLink output [KNOWN] | PX4 SITL MAVLink UDP stream over local socket. | MAVSDK-Python / MAVROS node | Freely generated in simulation [KNOWN] | Operational security only [INFERENCE] | [KNOWN] |
| Medical Payload Health Data | Ensure temperature-sensitive medicine (blood/vaccines) remains viable. | External IoT sensor payload bay (DS18B20 temp sensor) [KNOWN] | Simulated MQTT JSON payload stream (temp_c: 4.2, lock_status: SECURE). |
Serial / MQTT topic | Synthetically scriptable [INFERENCE] | Non-personal IoT data [INFERENCE] | [ASSUMPTION] |
To justify GarudaResQ, existing commercial, open-source, and government solutions were evaluated:
+-------------------------------------------------------------------------------------+
| EXISTING SOLUTIONS & GAPS |
+-------------------------------------------------------------------------------------+
| QGroundControl / Mission Planner : Great single-vehicle fly tool; NO edge AI or SAR |
| ICMR i-DRONE Platform : Great point-to-point medical flight; NO SAR/Swarm |
| Zipline / Commercial Logistics : Closed proprietary ecosystem; fixed hub dependency|
| ROS2 Academic Swarm Frameworks : Highly complex; NO user-friendly tactical UI |
+-------------------------------------------------------------------------------------+
[KNOWN][RESEARCHED][RESEARCHED]When designing a software system to satisfy SIH26201, three candidate technical architectures were evaluated:
| Evaluation Criteria | Approach 1: Pure Cloud Web GCS (SaaS) | Approach 2: Monolithic Desktop ROS2 GCS | Approach 3: Hybrid Edge-Swarm Node + Tactical Mesh GCS + Cloud Command (RECOMMENDED) |
|---|---|---|---|
| System Mechanism | Web browser connects via 4G/5G to cloud server; cloud relays MAVLink commands and runs cloud AI vision. | Single heavy C++ Qt/ROS2 desktop app running directly on a field laptop. | Local edge compute on drone companion computer (AI detection); local P2P Web/Electron GCS on field laptop; async cloud sync when online. |
| Problem Fit | Low [INFERENCE]. Fails in off-grid disaster zones due to connectivity reliance [INFERENCE]. | Medium [INFERENCE]. Strong offline capability, but rigid UI and poor multi-device field sharing [INFERENCE]. | High [INFERENCE]. Perfectly matches Indian off-grid field conditions while maintaining modern UI and edge AI performance [INFERENCE]. |
| Accuracy / Latency | High compute potential, but unviable due to 200ms-2s cloud round-trip video latency [INFERENCE]. | Zero network latency, but bound to laptop GPU capabilities [INFERENCE]. | Sub-30ms local edge video inference; real-time low-bandwidth MAVLink local mesh transmission [INFERENCE]. |
| Data Requirements | Continuous high-speed broadband connection required [INFERENCE]. | Local map tiles and offline spatial files [INFERENCE]. | Local offline vector map tiles, lightweight MAVLink packets over LoRa/Wi-Fi Direct [INFERENCE]. |
| Complexity | Low system integration, high network reliability burden [INFERENCE]. | High C++/ROS2 native compilation complexity [INFERENCE]. | Moderate: Modular split between Web UI (React/TypeScript), Node/Python backend gateway, and C++/Python ROS2 edge nodes [INFERENCE]. |
| Cost | High recurring cloud infrastructure fees [INFERENCE]. | Zero cloud fees; requires high-end field laptop [INFERENCE]. | Low cost; standard laptop + low-cost edge companion hardware (Raspberry Pi / Jetson) [INFERENCE]. |
| Hackathon Feasibility | High [INFERENCE]. | Low [INFERENCE]. Difficult to build polish and present in short window. | Very High [INFERENCE]. Clear separation allows team members to build UI, SITL backend, and AI pipeline in parallel [INFERENCE]. |
| Production Feasibility | Unviable for rural disasters [INFERENCE]. | Feasible, but poor fleet scale [INFERENCE]. | Highly Feasible; aligns with modern military/disaster tactical software architecture [INFERENCE]. |
The team selects Approach 3: The Hybrid Distributed Rescue Orchestration Architecture [INFERENCE].
+-----------------------------------------------------------------------------------+
| RECOMMENDED SYSTEM HYBRID |
+-----------------------------------------------------------------------------------+
| [ EDGE DRONE COMPUTER ] ---> TensorRT YOLOv8 Vision + MAVLink Controller |
| | (LoRa / Local Wi-Fi Mesh) |
| [ TACTICAL FIELD GCS ] ---> Electron/React Web GIS + PX4 SITL + Mesh Gateway |
| | (Cellular / Satcom when available) |
| [ CLOUD COMMAND DASH ] ---> Strategic Overview + DigitalSky Sync + Audit Logs |
+-----------------------------------------------------------------------------------+
===================================================================================
1. USER INTERFACE LAYER (Tactical Field GCS & Commander Dashboard)
- Web GIS Dashboard (React + Mapbox GL JS / Leaflet)
- Mission Planner (Lawnmower SAR Generator, Pin-Point Medical Drop Target)
- Real-Time Telemetry HUD (Altitude, Speed, Battery, Thermal Cam Feed)
- Mobile Field Companion App (PWA / React Native - Offline Victim Vector Maps)
===================================================================================
|| WebSocket / WebRTC Protocols
===================================================================================
2. GCS LOCAL BACKEND GATEWAY LAYER (Node.js Express + Python FastAPI)
- MAVLink Telemetry Parser (MAVSDK-Python / PyMAVLink)
- DigitalSky Airspace Boundary Geo-Fence Evaluator (R-Tree Indexing)
- Mesh Gateway Manager (UDP / Serial Bridge to LoRa radio)
- Offline GIS Map Tile Server (MBTiles / Vector Tile Server)
===================================================================================
|| Local RF / UDP MAVLink Communications
===================================================================================
3. EDGE COMPUTE & AUTONOMY LAYER (Drone / UGV Companion Hardware)
- ROS2 Humble / Micro XRCE-DDS Middleware
- AI Perception Pipeline: OpenCV Video Capture -> ONNX/TensorRT YOLOv8 Thermal Engine
- Waypoint Dispatcher & Emergency Fail-Safe Handler (Auto-RTL on Low Power/Geofence)
- Medical Payload Bay Actuator Service (GPIO / Serial Servo Release + Temp Monitoring)
===================================================================================
|| Serial MAVLink (UART)
===================================================================================
4. HARDWARE AUTOPILOT / SIMULATION LAYER
- PX4 Autopilot / ArduPilot Flight Controller (Physical Pixhawk 4 OR SITL Gazebo)
- Sensors: GPS/GNSS, IMU, Barometer, Optical Flow, Thermal Camera (FLIR)
- Actuators: ESCs, Brushless Motors, Payload Release Servos
===================================================================================
MAV_CMD_NAV_WAYPOINT) and pushed over UDP/Serial to the PX4 flight controller [KNOWN].| Architecture Layer | Selected Technology | Technical Justification | Evaluated Alternative | Reason Alternative Rejected |
|---|---|---|---|---|
| Frontend Framework | React 18 + TypeScript + Mapbox GL JS | Fast vector map rendering, highly responsive components, rich open-source GIS library support [KNOWN]. | Angular / Native C++ Qt | Qt has steep learning curve and slow web integration; Angular is overly boilerplate-heavy for hackathons [INFERENCE]. |
| GCS Local Backend | Node.js (Express) + Python 3.12 (FastAPI) | Dual stack: Python handles heavy MAVLink/GIS math; Node.js handles ultra-fast WebSocket UI streaming [INFERENCE]. | Pure Java / Spring Boot | High memory footprint; slow startup time on field laptops [INFERENCE]. |
| Drone Edge Autonomy | ROS2 Humble Hawksbill + MAVSDK-Python | Native microsecond inter-process communication, pub-sub architecture, industry-standard robotics framework [KNOWN]. | ROS1 Noetic | ROS1 is end-of-life and lacks native real-time QoS (Quality of Service) policies [KNOWN]. |
| AI Perception Framework | PyTorch -> ONNX -> TensorRT (YOLOv8-Nano) | Quantized 8-bit model delivers 45+ FPS on Jetson Orin Nano with minimal battery drain [RESEARCHED]. | Full TensorFlow / OpenCV Haar Cascades | TensorFlow is heavier with higher latency; Haar cascades fail under varying thermal conditions [RESEARCHED]. |
| Autopilot Middleware | PX4 Autopilot Ecosystem | Modern, highly modular, native Micro XRCE-DDS integration with ROS2, robust SITL simulation support [KNOWN]. | ArduPilot | ArduPilot is excellent but ROS2 integration requires more complex MAVROS bridges compared to PX4 [INFERENCE]. |
| Database & Caching | SQLite (GCS Local) + IndexedDB (Browser) | Embedded, zero-configuration relational database ideal for off-grid desktop/PWA environments [KNOWN]. | PostgreSQL / MongoDB | Requires running active database daemons, unnecessary overhead for field GCS [INFERENCE]. |
| Off-Grid Connectivity | MAVLink over 865-867 MHz LoRa / Wi-Fi Mesh | Operates within legal Indian unlicensed ISM bands; penetrates light foliage [RESEARCHED]. | 4G/5G Dongles | Unreliable during actual Indian disaster events due to tower failures [RESEARCHED]. |
| Simulation Engine | Gazebo Harmonic + PX4 SITL | Accurate multi-vehicle physics, sensor simulation (thermal camera, LiDAR), native ROS2 bridge [KNOWN]. | AirSim | Microsoft AirSim is archived/deprecated and extremely resource-heavy for student laptops [KNOWN]. |
AI should only be applied where traditional deterministic code fails. Below is the strict justification breakdown:
+------------------------------------------------------------------------------------+
| AI vs. RULE-BASED LOGIC |
+------------------------------------------------------------------------------------+
| USE AI : Thermal & Optical Survivor Detection (Unstructured video frames)|
| DO NOT USE AI : Flight Stability & Motor Control (Deterministic PID Loops) |
| DO NOT USE AI : DigitalSky Airspace Geofencing (Exact 2D/3D Polygon Geometry) |
| DO NOT USE AI : Fail-Safe Return-to-Launch (Hardcoded Altitude & Battery Limits)|
+------------------------------------------------------------------------------------+
[x_min, y_min, x_max, y_max], class label (human_survivor), and confidence score [0.0 - 1.0] [KNOWN].-- Missions Table
CREATE TABLE missions (
mission_id TEXT PRIMARY KEY,
name TEXT NOT NULL,
type TEXT CHECK(type IN ('SEARCH_RESCUE', 'MEDICAL_DELIVERY', 'RECON')),
status TEXT CHECK(status IN ('PLANNED', 'ACTIVE', 'COMPLETED', 'ABORTED')),
created_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP,
boundary_geojson TEXT NOT NULL
);
-- Vehicles Table
CREATE TABLE vehicles (
vehicle_id TEXT PRIMARY KEY,
type TEXT CHECK(type IN ('QUADCOPTER', 'HEAVY_LIFT_HEXA', 'UGV_GROUND_ROBOT')),
firmware_version TEXT,
battery_level INTEGER,
last_known_lat REAL,
last_known_lon REAL,
status TEXT CHECK(status IN ('DISCONNECTED', 'ARMED', 'IN_FLIGHT', 'RTL'))
);
-- Survivor Detections Table
CREATE TABLE detections (
detection_id TEXT PRIMARY KEY,
mission_id TEXT FOREIGN KEY REFERENCES missions(mission_id),
vehicle_id TEXT FOREIGN KEY REFERENCES vehicles(vehicle_id),
latitude REAL NOT NULL,
longitude REAL NOT NULL,
altitude_m REAL,
confidence REAL NOT NULL,
thermal_snapshot_blob BLOB,
status TEXT CHECK(status IN ('UNCONFIRMED', 'VERIFIED_VICTIM', 'FALSE_POSITIVE', 'RESCUED'))
);
-- DigitalSky Airspace Cache
CREATE TABLE airspace_zones (
zone_id TEXT PRIMARY KEY,
zone_type TEXT CHECK(zone_type IN ('GREEN', 'YELLOW', 'RED')),
min_altitude_ft REAL,
max_altitude_ft REAL,
geometry_geojson TEXT NOT NULL
);
+-----------------------------------------------------------------------------------+
| OFFLINE MESH SYNC ARCHITECTURE |
+-----------------------------------------------------------------------------------+
| [ Field Operator A (Tablet) ] [ Field Operator B (Laptop) ] |
| \ / |
| \---- LoRa / Wi-Fi Mesh ----/ |
| | |
| Vector Clock Conflict Resolution |
| "Master Commander Override Wins" |
+-----------------------------------------------------------------------------------+
lat, lon, alt, battery, state) are broadcast every 1000ms [INFERENCE].To guarantee deliverability before the 30 September 2026 deadline [PS], the project scope is divided into clear functional tiers:
+-----------------------------------------------------------------------------------+
| HACKATHON MVP SCOPE |
+-----------------------------------------------------------------------------------+
| MUST BUILD : 2-Drone SITL Simulation, React GIS HUD, Lawnmower SAR Generator, |
| YOLOv8 Edge Thermal AI, MAVSDK Telemetry Stream, DigitalSky Check |
| SHOULD BUILD : UGV Ground Robot Telemetry Bridge, Medical Drop Lock Trigger UI |
| NICE TO HAVE : LoRa Radio Hardware Loop Test, Mobile PWA Offline Map Viewer |
| OUT OF SCOPE : Custom Drone Hardware Assembly, Live Physical Airspace BVLOS |
+-----------------------------------------------------------------------------------+
MAV_CMD_DO_SET_SERVO) to simulate releasing a medical box over target coordinates [INFERENCE].Because live drone flights inside judging venues are unsafe, illegal, or physically infeasible, the team must execute a high-fidelity Software-In-The-Loop (SITL) Live Simulation [INFERENCE].
+-----------------------------------------------------------------------------------+
| LIVE DEMO SIMULATION PIPELINE |
+-----------------------------------------------------------------------------------+
| [ PX4 SITL + Gazebo ] --(MAVLink UDP)--> [ GarudaResQ Local Node Backend ] |
| | | |
| (Simulated FLIR Video Feed) (WebSocket Real-Time Feed) |
| v v |
| [ YOLOv8 Thermal Detection ] ------------> [ React Web GIS Map Dashboard ] |
+-----------------------------------------------------------------------------------+
px4_sitlgazebo-classic via Docker on the demonstration laptop. Spawns two virtual vehicles at coordinates matching a real Indian disaster site (e.g., Wayanad, Kerala or Kedarnath, Uttarakhand) [INFERENCE].SUCCESS [INFERENCE].The roadmap is structured backward from the deadline of 30 September 2026 [PS]:
+-----------------------------------------------------------------------------------+
| TIMELINE ROADMAP |
+-----------------------------------------------------------------------------------+
| Phase 1: Arch & SITL Setup (May 2026 - Jun 2026) |
| Phase 2: Core GCS & Path Generator (Jun 2026 - Jul 2026) |
| Phase 3: AI Vision & Telemetry (Jul 2026 - Aug 2026) |
| Phase 4: Integration & Edge Mesh (Aug 2026 - Sep 15, 2026) |
| Phase 5: Demo Lock & Freeze (Sep 16, 2026 - Sep 30, 2026) |
+-----------------------------------------------------------------------------------+
| Test Category | Target Metric / Scenario | Validation Methodology | Success Threshold |
|---|---|---|---|
| Functional Flight Test | Lawnmower waypoint upload and auto-mission execution in SITL. | Automated MAVSDK test script verifying mission completion. | 100% mission item execution without flight command timeouts [INFERENCE]. |
| AI Perception Performance | Precision and Recall on thermal human detection under partial occlusion. | Test inference on 500 ground-truth thermal test images. | [email protected] >= 0.85; Inference speed >= 30 FPS on TensorRT [INFERENCE]. |
| Airspace Safety Validation | Detect intentional flight path overlap into simulated Indian Red Zones. | Unit tests passing 50 illegal GeoJSON polygons through fencer. | 100% illegal paths hard-blocked with zero false negatives [INFERENCE]. |
| Offline Resilience | Disconnect physical WAN/Internet cable during multi-vehicle mission execution. | Kill active internet gateway during live flight in SITL. | Local GCS continues receiving MAVLink telemetry and displaying map tiles without crash [INFERENCE]. |
| Telemetry Latency | End-to-end packet delay from PX4 SITL to React GIS UI meters. | Timestamp log comparison between MAVLink emit and WebSocket UI render. | Latency < 100 ms over local network loopback [INFERENCE]. |
| UI Usability Test | Time taken by non-technical user to dispatch a rescue drone upon alert. | User study measuring task completion time for 5 test subjects. | Complete dispatch workflow in < 3 button clicks / under 15 seconds [INFERENCE]. |
| Risk Event | Likelihood | Impact | Severity | Technical Mitigation Strategy |
|---|---|---|---|---|
| High SITL Laptop Resource Consumption (Gazebo + Docker + Web UI crashes student laptop during live presentation) | High [INFERENCE] | High [INFERENCE] | CRITICAL | Pre-configure lightweight headless Gazebo options (gz sim -s); split frontend and SITL onto two laptops via local Ethernet cable if needed [INFERENCE]. |
| AI Thermal False Positives (Hot rocks/animals flagged as disaster victims) | Medium [INFERENCE] | Medium [INFERENCE] | MODERATE | Require multi-frame verification (target must persist across 5 consecutive video frames) and enforce human commander visual confirmation [INFERENCE]. |
| MAVLink Signal Loss in Off-Grid Mesh (Radio packet drop over long ranges) | High [INFERENCE] | High [INFERENCE] | HIGH | Implement automatic drone fail-safe: if radio link is lost for >30 seconds, vehicle completes search pattern autonomously or executes auto-RTL [KNOWN]. |
| Incomplete DigitalSky Airspace Data (Official DGCA API unavailable or restricted during build) | High [RESEARCHED] | Medium [INFERENCE] | MODERATE | Build a robust offline parser that ingests standard GeoJSON files matching official DigitalSky zone schemas [INFERENCE]. |
| Payload Balance Instability (Medical box drop causes drone aerodynamic flip) | Low [INFERENCE] | High [INFERENCE] | HIGH | Simulate payload release only when drone is in stabilized low-altitude hover (<0.2 m/s velocity) [INFERENCE]. |
To stand out from generic SIH submissions, GarudaResQ incorporates distinct operational capabilities directly linked to SIH26201 requirements:
+------------------------------------------------------------------------------------+
| KEY DIFFERENTIATORS |
+------------------------------------------------------------------------------------+
| 1. Dual-Domain Heterogeneous Coordination (Scout UAV + Cargo UAV + Ground UGV) |
| 2. DGCA DigitalSky Offline Parser (Built specifically for Indian Airspace Rules) |
| 3. Bandwidth-Adaptive Mesh Protocol (Translates video to 3KB thermal thumbnails) |
| 4. 100% Off-Grid Operability (Runs complete GCS stack on local loopback) |
+------------------------------------------------------------------------------------+
Post-hackathon development can extend the system into a commercially viable product:
[ASSUMPTION][ASSUMPTION] Student team will have access to at least one mid-spec Linux/Windows laptop (Intel i7/Ryzen 7, 16GB RAM) capable of running Gazebo multi-vehicle SITL simulation.[ASSUMPTION] IoT medical payload enclosure uses a simple UART/GPIO relay lock actuated via PX4 standard MAVLink servo commands.[ASSUMPTION] Synthetic GeoJSON zone boundaries accurately reflect official DGCA DigitalSky spatial polygon formats.[NEEDS VALIDATION][NEEDS VALIDATION] Exact pricing and procurement lead-time for IP65 rugged tactical laptops in India.[NEEDS VALIDATION] Physical radio range limit of 865-867 MHz LoRa modules through dense Himalayan forest canopy under non-line-of-sight (NLOS) conditions.[NEEDS VALIDATION] Direct REST API access endpoints for live DigitalSky airspace zone queries for non-governmental software developers.Build GarudaResQ as an offline-first, ROS2 and PX4-powered Tactical Ground Control Station with edge-based YOLOv8 thermal vision and automated DigitalSky airspace safety enforcement [INFERENCE]. Focus the hackathon build on delivering a flawless dual-vehicle SITL simulation that proves end-to-end autonomous search, human victim discovery, and precision medical payload dispatch [INFERENCE].
PX4 SITL + Gazebo multi-vehicle simulation stability on the demonstration laptop [INFERENCE].
Laptop hardware lag or crashes during live evaluation while rendering 3D Gazebo physics alongside React Mapbox GIS maps [INFERENCE].
Achieving low-latency, real-time synchronization between edge AI object detection bounding boxes, MAVLink drone telemetry coordinates, and map overlay renders [INFERENCE].
88 / 100 [INFERENCE] (High feasibility: The software stack relies entirely on mature, open-source robotics and web technologies (PX4, ROS2, React, YOLOv8) with zero requirement for physical hardware manufacturing during the hackathon phase).
HIGH [INFERENCE]
(The solution perfectly aligns with the explicit Software category of problem statement SIH26201, addresses real-world Indian operational constraints, and provides a clear path to live demonstration).
What we think about this platform, and how it can help people solve SIH problems.
" 30M+ Google Gemini tokens went into researching, analyzing, and understanding all 233 SIH problem statements to make them easier to explore and solve. "
" The platform makes SIH problem statements much easier to understand. Having everything from the problem details to the procedure in one place. "
" Oh, you got ads? Well, we need to make money too! But, you can use our platform to help you win SIH and take home the prize money. "
Get in touch if you need help bringing your ideas to life, or simply have an interesting idea you'd like to talk through. I'm always open to a conversation and exploring new possibilities together.