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.
python raspberry-pi iot computer-vision finite-state-machine quantization ultrasonic-sensor ldr pir-sensor ldr-sensor clahe edge-ai yolov8 low-power-inference sensor-based-inference
-
Updated
Jul 12, 2026 - Python