Stanford researchers build AI teen driver safety system

SafeStudent Driving is a multimodal driver-safety system designed to support teen drivers, who face disproportionately high crash rates due to inexperience and inconsistent attention to basic traffic rules. Built by Max Liu and Yu Sun, the system runs on a Raspberry Pi and a Flutter-based mobile app, using three YOLO-based computer-vision models to detect traffic lights, light-bulb colors, and road signs, plus OCR to read speed-limit values and an audio model with IMU data to infer whether turn signals are used. An analysis layer smooths detections over time and triggers prioritized voice prompts via text-to-speech or pre-recorded audio. The project tackled challenges like model accuracy in varied lighting, fast inference on limited hardware, and designing prompts that inform without distracting the driver. Experiments on sign detection and turn-signal recognition highlighted strengths and failure modes, guiding future improvements—and demonstrating a practical, low-cost way to help novice drivers build safer habits in real traffic.