1 of 1

AdaptiSync: AI-based Real-Time Adaptive Traffic Control using Edge-Computing

Amit Agarwal, Karthik Krishnan O, Dhinesh Kumar R, KV Pranav, Dhish Kumar Saxena

Lead Inventor’s E-mail: amitfce@iitr.ac.in

  • The product that can be installed at a new signalized intersection or can be integrated with an existing traffic intersection.
  • It works in real-time and supports multiple vehicle classes using the Indian Traffic Dataset (ITD).
  • It is a wireless and portable product, which makes the installation quick and has little to no interruption to traffic and damage to the pavements.
  • It is based on edge computing, i.e., configured AI-based algorithms for vehicle identification, classification, and counting to support traffic conditions in India.
  • The architecture is designed to support different single-board computers (SBCs) for long-term compatibility.
  • It is integrated with cloud computing for data storage and signal timing computation using Max Pressure Control and Reinforcement Learning (RL) algorithms.
  • The product is economical, scalable, transferable, and suitable under the Smart City Mission by the Govt of India.

INDIAN INSTITUTE OF TECHNOLOGY ROORKEE