DRISHTI-GRID

Turning E-waste into autonomous Edge-AI sensors to detect road hazards, optimize repairs, and mathematically eliminate infrastructure fraud.

Comment

💡 The Real-World Problem

Modern cities face a three-front infrastructure crisis:

  1. Safety: Reactive road maintenance leads to thousands of lethal accidents, delaying emergency ambulances during the critical "Golden Hour."
  2. Systemic Fraud: Municipal governments lose millions to "ghost fixes" and contractor material over-billing due to a lack of transparency.
  3. Sustainability: Millions of functional smartphones are discarded annually, fueling a massive E-waste crisis.

What if we could upcycle this E-waste into a zero-cost, autonomous AI grid that holds infrastructure contractors accountable? That vision became DRISHTI-GRID 2.0.

⚙️ What it does (Innovation & Impact)

DRISHTI-GRID 2.0 is an autonomous GovTech platform that manages the entire lifecycle of road maintenance:

  • Zero-Cost Edge Sensors (Mobile PWA): Discarded smartphones (e.g., Samsung M12) are mounted on municipal garbage trucks. Running our Next.js Edge app, they autonomously scan roads using a Tri-Trigger engine (Heartbeat & HTML5 DeviceMotion jerks).
  • AI Fraud Prevention (AI/ML): Google Gemini 3 Flash performs Monocular Depth Estimation on 2D images to calculate the exact 3D volume of the hazard. This mathematically caps the raw materials a contractor can bill for.
  • Smart Dispatch (Web Platform): The Command Center clusters work orders by contractor jurisdiction. A Live Weather Engine blocks "Hot-Mix" asphalt dispatch during monsoons, forcing temporary gravel fixes to prevent washed-away repairs.
  • Closed-Loop AI Verification: Contractors are only paid when our Edge Node re-drives the exact coordinate and the AI visually verifies the road is smooth, committing the proof to an immutable ledger.

📐 UI/UX & Design Philosophy

We engineered a dual-interface system with strict Role-Based Access Control:

  • The Edge HUD: Designed for fleet drivers, featuring a cyberpunk-inspired, dark-mode laser scanner HUD that requires zero manual input while driving.
  • The PWD Command Center: A military-grade, high-contrast Web UI for city officials, featuring dynamic geospatial plotting (Leaflet Carto-Dark), SLA matrices, and real-time financial tracking.

🛠️ Technical Implementation (How we built it)

  • AI Brain: Google Gemini 3 Flash Preview via REST API for ultra-low latency multimodal processing and structured JSON schema generation.
  • Frontend Web: Next.js 14 (React PWA), TypeScript, Tailwind CSS.
  • Database & Cloud: Google Firebase Firestore for real-time NoSQL spatial Web-Socket syncing.
  • Device APIs: HTML5 Geolocation API and DeviceMotion API.

🚧 Challenges we ran into

API Spam from Moving Vehicles: If a fleet vehicle gets stuck in traffic over a pothole, it would spam the AI API and database. Solution: We implemented an offline edge-cache utilizing the Haversine Formula. If the active GPS lock is $< 1$ meter from a cached scan, the cloud sync is mathematically suppressed, saving battery and compute costs.

🏆 Accomplishments that we're proud of

  • Achieving a $0.00 Hardware Deployment Cost by proving that discarded E-waste can run enterprise-grade IoT edge computing.
  • Successfully engineering a multimodal AI prompt capable of reliable 3D Monocular Depth Estimation for financial bounding.
  • Achieving a zero-error strict TypeScript build deployed seamlessly on Vercel.

🚀 What's next

  1. Web3 Integration: Migrating our AI Verification Ledger to decentralized Smart Contracts to enable automated, tamper-proof payouts to contractors.
  2. OBD2 Telemetry: Integrating real-time vehicle data to dynamically adjust camera capture rates based on truck speed.

Built With

Share this project:

Updates