Edge AI-Powered Smart Traffic Light System

The Edge AI-Powered Smart Traffic Light System operates by combining embedded computer vision with local microcontroller-based hardware control. At the heart of the setup is the Sipeed Maixduino Kit for RISC-V AI + IoT, which integrates a powerful vision processor capable of running machine learning models entirely at the edge without relying on cloud infrastructure. A convolutional neural network, compiled using NNCase and deployed via Python and MaixPy IDE, processes live video feeds to analyze traffic density in real-time. Based on the detected volume of vehicles, the microcontroller dynamically calculates optimal signal timings. To physically represent the intersection signals, the Maixduino outputs digital control signals to a bank of LEDs. Specifically, the system manages two red LEDs and two green LEDs to simulate a directional traffic control layout. Because digital output pins on microcontrollers supply logic-level voltage that can otherwise overcurrent and damage standard semiconductor junctions, four 330-ohm resistors are integrated in series with the LEDs to limit forward current to safe operating levels. As the AI model registers fluctuating vehicle counts, it switches the digital states of the pins, seamlessly transitioning the red and green illumination patterns to demonstrate intelligent traffic flow management.