Convolutional Neural Network Method for Determining the Optimum Number of Transport Data Loggers

Abstract

This article shows a method based on the convolution neural network approach of the embedded dimensions of the acceleration sensor values recorded on the MEMS-based Transport Data Logger (TDL) for tilt detection. The ultimate aim is to define a system defining the optimum number of data loggers covering the nonlinear scaling regions of the original accelerometer signal of the TDL tilting experiment.