TraffiSense AI: A Quantum-Inspired Congestion Prediction and Signal Scheduling Assistant

Traffic congestion remains one of the primary challenges in developing smart cities, primarily caused by rapid urban growth and infrastructure expansion. Existing intelligent traffic management systems often depend on cloud-based processing, which introduces latency, higher operational costs, and data security risks. To overcome these limitations, this paper introduces TrafficSense AI, an offline-first, quantum-inspired traffic optimization framework designed for efficient, real-time signal coordination without relying on cloud infrastructure. The system reformulates the traffic signal scheduling problem as a Quadratic Unconstrained Binary Optimization (QUBO) model and applies Quantum-Inspired Evolutionary Algorithms (QIEA) to identify near-optimal solutions. Additionally, a Reinforcement Learning (RL) controller dynamically adjusts signal timings based on live traffic conditions to improve flow and minimize congestion. Implemented using Python and the SUMO traffic simulator on realistic urban road networks, TrafficSense AI demonstrates a 21% increase in throughput and a 17–32% reduction in vehicle waiting times compared to conventional fixed-time systems, even under CPU-only local execution. With its privacy-preserving, scalable, and cloud-independent architecture, TrafficSense AI offers a promising foundation for future advancements such as multi-intersection coordination, Vehicle-to-Infrastructure (V2I) integration, and autonomous route optimization, paving the way for efficient and sustainable urban mobility.