Urban expansion has significantly increased vehicular traffic, creating challenges in road monitoring and incident response. We present a cyber-physical system (CPS) integrating IoT sensors, edge computing, and machine learning for real-time traffic management. Our architecture employs probe vehicles with GPS/ultrasonic sensors and fixed infrastructure using RFID cameras, achieving 92% incident detection accuracy. The system reduces average congestion by 30% through adaptive signal control and demonstrates 2.1-minute emergency response times. Experimental results show 40% improvement in traffic flow efficiency compared to traditional systems.