Adaptive Real-Time Gap Detection System: A Multi-Algorithm Approach Integrating Ultrasonic Sensing and Machine Learning for Robust Structural Analysis

This paper presents an adaptive real-time gap detection system designed to address the challenges of automated low-cost structural inspection in environments with complex geometries. While conventional ultrasonic sensing methods often face limitations, particularly when utilizing lowcost components, our novel framework integrates ultrasonic sensing with integration machine learning to achieve robust detection capabilities. The system incorporates three key components: Kalman filtering for real-time noise reduction, DBSCAN clustering for geometric pattern recognition, and Isolation Forest for anomaly detection. Experiments were conducted using low-cost ultrasonic sensors, demonstrating significant performance improvements with a 95 % success rate across all test conditions. Confidence scores ranged from 0.86 to 0.88 in complex settings, with a processing speed of <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$400-500\ {ms}$</tex>.