Skip to content

Autonomous Rescue System

Decouples conversational voice intake from mathematical optimization to slash emergency CAD dispatch latency by over 60% and guarantee zero allocation errors.

Business Domain
Tech Domain
Stall Details
Student Innovation Acharya Bangalore B-School Students

Stall Participants & Presenters

TP-09
Rahul Kumar Khatua

Rahul Kumar Khatua

Student Innovator
Team Member
Manoj Kumar S

Manoj Kumar S

Student Innovator
Team Member

Overview

Over 4.2 million preventable pre-hospital deaths occur worldwide each year due to the “Golden Hour” penalty: patient survival chances plummet by 7% to 10% for every single minute an emergency response is delayed.

Legacy 911 and 112 Computer-Aided Dispatch (CAD) systems suffer from three acute bottlenecks:

  1. Manual Ingestion Delays: 3 to 7 minutes of stressful verbal interrogation while frantic callers struggle to communicate clearly.
  2. Greedy Dispatch Heuristics: Traditional CAD platforms dispatch the single nearest available ambulance unit, leaving downstream high-severity (S5) trauma victims completely unserviced just blocks away.
  3. Operational Fragility: When an emergency vehicle breaks down or road flooding obstructs an artery, manual dispatcher recalculations take 15 to 30 minutes.

The Autonomous Rescue System (Crisis Command AI) replaces this broken workflow by decoupling conversational voice intake from deterministic mathematical optimization. The platform cuts emergency dispatch latency by over 60%, eliminates allocation hallucinations, and dynamically recalculates citywide fleet assignments in under 500 milliseconds, directly advancing SDG 3 (Good Health and Well-Being), SDG 9 (Industry, Innovation, and Infrastructure), and SDG 11 (Sustainable Cities and Communities).

Autonomous Ingestion & Multi-Agent Dispatch Engine

  • Conversational Voice Ingestion: Employs Whisper Voice Activity Detection (VAD) coupled with Google Gemini 1.5 Flash to parse panicked, unstructured emergency calls, transcribing medical symptoms and extracting critical location markers in real time.
  • Deterministic Mathematical Optimization: Decouples natural language processing from resource allocation by feeding structured incident constraints into Google OR-Tools CP-SAT solver, guaranteeing zero allocation hallucinations.
  • Global Fleet Optimization: Avoids greedy nearest-unit pitfalls by optimizing global survival probability across all active incident clusters, preserving advanced life support units for high-severity trauma cases.
  • Dynamic Real-Time Rerouting: Automatically ingests live road closures, vehicle telemetry, and weather alerts to trigger microsecond fleet recalculations in under 500ms.
  • Dispatcher Copilot UI: Delivers clean visual telemetry to dispatch directors with automated routing recommendations, patient triage scores, and ETA projections.

Architecture & Technology Stack

  • Voice Ingestion & Triage: Google Gemini 1.5 Flash with Whisper Voice Activity Detection (VAD) for real-time speech-to-text and unstructured entity extraction.
  • Optimization & Routing Core: Google OR-Tools CP-SAT constraint programming solver combined with geospatial routing engines for fleet assignment.
  • Backend & Data Pipeline: Python / FastAPI event-driven services handling high-concurrency emergency incident streams and telematics queues.
  • Real-Time Telemetry & Tracking: WebSocket and GPS telemetry relays broadcasting vehicle states, transit coordinates, and traffic corridor updates.
  • Frontend Dashboard: Modern dispatcher command portal visualizing citywide emergency incidents, ambulance coordinates, and predictive ETAs.