A neural network based diagnostic test system for armored vehicle shock absorbers

摘要

In this paper we describe the development of a neural network based diagnostic system that is being utilized to evaluate the condition of armored vehicle shock absorbers. We begin by providing the motivation behind the development of the Smart Shock Absorber Test Stand (SSATS). We then describe the theory required to evaluate the condition of shock absorbers. This theoretical discussion leads to a description of the type of data that are acquired during a shock absorber test and how it can be analyzed to determine the condition of a shock absorber. The next section describes how the data are transformed and processed in order to develop a neural network classification scheme. The training and testing process of a fully connected feedforward backpropagation neural network is then described. Finally, we explain the integration of the neural network into the SSATS system and the results of utilizing the system to test Bradley Armored Vehicle shock absorbers.