MUMBAI, India, July 30 -- Intellectual Property India has published a patent application (202641088592 A) filed by H Parthasarathi Patra; Dr. Aravind Papasani; Motamarri Dhatri; Navya Nagireddy; Mamidi Amrutha Varshini; and Madugula Jahanavi on July 21, 2026, for Multi - Agent Reinforcement Learning For Traffic Signal Optimization With Emergency Vehicle Priority In Smart Cities.
Inventors include H Parthasarathi Patra; Dr. Aravind Papasani; Motamarri Dhatri; Navya Nagireddy; Mamidi Amrutha Varshini; and Madugula Jahanavi.
The application for the patent was published on July 24, 2026, under issue no. 30/2026.
Abstract: Urban traffic congestion remains a critical challenge in modern smart cities, where conventional signal control methods such as fixed-time and actuated controllers fail to adapt to dynamic traffic conditions. This paper proposes a Decentralized Multi-Agent Reinforcement Learning (MARL) framework for adaptive traffic signal optimization across a six-intersection urban road network. Each intersection is controlled by an independent Deep Q-Network (DQN) agent that observes local traffic states and selects optimal signal phases to minimize vehicle waiting time and queue length. The system is implemented and evaluated in the SUMO (Simulation of Urban Mobility) microscopic traffic simulator via the TraCI interface. An Emergency Vehicle Priority (EVP) module overrides normal signal control to grant prioritized signal access to emergency vehicles detected within 150 metres of an intersection. Statistical Model Checking (SMC) is applied to probabilistically verify fairness and demand-responsiveness of the learned policies. Experimental results demonstrate that the proposed MARL-DQN framework achieves up to 97.0% reduction in vehicle waiting time and 82.7% reduction in queue length compared to fixed-time control, reduces emergency vehicle delays by up to 100% when EVP is active, and achieves zero fairness-violation probability as verified by SMC
Disclaimer: Curated by HT Syndication.
