Sensor for Classification of Material Type and Its Surface Properties Using Radial Basis Networks

Abstract

An application of the power of neural networks in the implementation of a novel sensor for classification of material type and its surface properties by means of a lightweight plunger probe and optical mouse sensor is presented in this paper. An experimental prototype was developed which involves bouncing or hopping of the plunger-based impact probe freely on the plain surface of an object under test. During the bouncing of the probe, a time-varying signal is generated from optical mouse that is recorded in a data file on PC. Features of the signals are then extracted using signal processing tools to optimize neural network-based classifier used in the existing system. The classifier is developed using radial basis function neural networks (RBF NNs). For this, an optimum RBF NN model is designed to maximize accuracy under the constraints of minimum network dimension. Levenberg-Marquardt learning algorithm, which provides faster rate of convergence, has been found suitable for the training of RBF NN. The optimal parameters of RBF NN model based on classification accuracy on the testing data sets even after attempting different data partitions are determined. The classification accuracy of RBF NN is found consistently reasonable in respect of rigorous testing using different data partitions, and multifold cross-validation.