Design and Implementation of Intelligent Pet Anti-Barking Collar Control System Based on Multi-Sensor Fusion

Abstract

Traditional single-microphone anti-bark collars suffer high false alarms, poor multi-dog adaptability, missing safety safeguards and failure to pass overseas EMC certifications. This paper develops an embedded control system based on decision-level multi-sensor fusion, using STM32L4 low-power MCU integrated with MEMS audio, capacitive wearing detection and NTC temperature sensors to form a perception-decision-execution closed loop. It supports two spray volumes (0.25 mL / 0.5 mL), ±30% adaptive recognition thresholds and a 12-hour continuous-use safety lock. Verified by third-party FCC Part 15 and CE-EMC labs, the system is mass-produced for TA.TIAOM products sold on Amazon UK. Real household tests deliver an 89.3% effective correction rate, and multi-sensor joint judgment drastically reduces noise-induced misfires, providing a viable engineering solution for pet wearable behavior collars.