A Raspberry Pi 4 edge application that uses PIR, LDR and ultrasonic sensors to throttle YOLOv8 inference resolution and frame rates to reduce mean SoC power consumption and temperature. Serial I/O and TFLite tasks are pinned to isolated CPU cores to maintain stable latency.
pythonraspberry-piiotcomputer-visionfinite-state-machinequantizationultrasonic-sensorldrpir-sensorldr-sensorclaheedge-aiyolov8low-power-inferencesensor-based-inference
-
Updated
Jul 12, 2026 - Python