<p>Bengaluru: Close to 1,100 AI based and machine learning prototypes were developed to address the traffic problem in the city during the ‘Gridlock Hackathon 2.0’ organised by Flipkart in collaboration with <a href="https://deccanherald.com/tags/bengaluru">Bengaluru </a>Traffic Police, MapmyIndia, Arcadis, and HackerEarth.</p><p>The participants were encouraged to build solutions using anonymised, real-world datasets through the Bengaluru Traffic Police ASTraM (Actionable Intelligence for Sustainable Traffic Management) platform.</p><p>Among the winning solutions was ‘Gridlock Oracle’ which predicts how a single traffic incident cascades into city-wide congestion using a Hawkes Process model, giving the Traffic Police a live control-room dashboard to deploy resources proactively before gridlock spreads.</p>.'Bengaluru's Traffic, roads, discipline only getting worse': Infosys Kris Gopalakrishnan. <p>The second place was awarded to PRAHAR (Predictive Resource Allocation for High-Impact Area Response)-- a software-only predictive traffic intelligence platform built on real Bengaluru incident data to forecast congestion severity, junction risk, resolution time, and cascade effects. It recommends optimal deployment of officers, barricades, diversion routes, and dispatch stations.</p><p>The third place was given to ParkSight — Parking-Induced Congestion Intelligence which analyses parking violation records to identify priority congestion and enforcement zones, incorporating hotspot detection, forecasting, patrol optimisation, and validation against ASTraM congestion data.</p><p>The initiative received 33,600 registrations from students, developers, and working professionals across the country. </p><p>“Participants developed solutions across themes including congestion prediction, traffic violation detection, movement-pattern analysis, parking intelligence, and smarter mobility decision-making. The hackathon followed a two-stage format, beginning with an online machine learning challenge and progressing to prototype development for shortlisted participants, resulting in 1,100+ prototypes addressing real-world urban mobility challenges,” a statement by <a href="https://deccanherald.com/tags/flipkart">Flipkart</a> said.</p>
<p>Bengaluru: Close to 1,100 AI based and machine learning prototypes were developed to address the traffic problem in the city during the ‘Gridlock Hackathon 2.0’ organised by Flipkart in collaboration with <a href="https://deccanherald.com/tags/bengaluru">Bengaluru </a>Traffic Police, MapmyIndia, Arcadis, and HackerEarth.</p><p>The participants were encouraged to build solutions using anonymised, real-world datasets through the Bengaluru Traffic Police ASTraM (Actionable Intelligence for Sustainable Traffic Management) platform.</p><p>Among the winning solutions was ‘Gridlock Oracle’ which predicts how a single traffic incident cascades into city-wide congestion using a Hawkes Process model, giving the Traffic Police a live control-room dashboard to deploy resources proactively before gridlock spreads.</p>.'Bengaluru's Traffic, roads, discipline only getting worse': Infosys Kris Gopalakrishnan. <p>The second place was awarded to PRAHAR (Predictive Resource Allocation for High-Impact Area Response)-- a software-only predictive traffic intelligence platform built on real Bengaluru incident data to forecast congestion severity, junction risk, resolution time, and cascade effects. It recommends optimal deployment of officers, barricades, diversion routes, and dispatch stations.</p><p>The third place was given to ParkSight — Parking-Induced Congestion Intelligence which analyses parking violation records to identify priority congestion and enforcement zones, incorporating hotspot detection, forecasting, patrol optimisation, and validation against ASTraM congestion data.</p><p>The initiative received 33,600 registrations from students, developers, and working professionals across the country. </p><p>“Participants developed solutions across themes including congestion prediction, traffic violation detection, movement-pattern analysis, parking intelligence, and smarter mobility decision-making. The hackathon followed a two-stage format, beginning with an online machine learning challenge and progressing to prototype development for shortlisted participants, resulting in 1,100+ prototypes addressing real-world urban mobility challenges,” a statement by <a href="https://deccanherald.com/tags/flipkart">Flipkart</a> said.</p>