Object-Size-Driven Design of Convolutional Neural Networks: Virtual Axle Detection based on Raw Data

As infrastructure ages, the need for efficient monitoring methods becomes increasingly critical. Bridge Weigh-In-Motion (BWIM) systems are crucial for cost-efficient load and thus residual service life determination of road and railway infrastructure. However, conventional BWIM systems require additional sensors for axle detection, which have to be installed in potentially inaccessible locations or in locations that interfere with bridge operation. This study addresses this challenge by replacing dedicated axle detectors with a novel approach to real-time detection of train axles using sensors arbitrarily placed on bridges. The proposed Virtual Axle Detector with Enhanced Receptive Field (VADER) has been validated on a single-track railway bridge, demonstrating that it achieves to detect 99.9% of axles with a spatial error of 3.69cm using only acceleration measurements. Using raw data as input outperforms the state-of-the-art spectrogram-based method in both speed and memory usage by 99%, making real-time application feasible for the first time. Additionally, we introduce the Maximum Receptive Field (MRF) rule, a novel approach to optimise hyperparameters of Convolutional Neural Networks (CNNs) based on the size of objects, which in this case relates to the fundamental frequency of a bridge. The MRF rule effectively narrows the hyperparameter search space, potentially replacing the need for extensive hyperparameter tuning. Since the MRF rule is theoretically applicable to all unstructured data, it could have implications for a wide range of deep learning problems from earthquake prediction to object recognition.