A technique to increase performance of a distributive tactile sensing system through an application of the principal component analysis (PCA)

∗ Department of Production Engineering, Faculty of Engineering King Mongkut's Institute of Technology North Bangkok Abstract This paper describes a one-dimensional distributive tactile system that can be used to determine simple contact parameters such as weight, size and position. The system is a simply supported beam on which load position was determined from the induced deflection. This research implements the Principal Component Analysis (PCA) that can effectively increase the accuracy by pre-processing sensory signals before putting through to the interpretation algorithm. The application of the PCA has shown an improved accuracy in determining an applied load position as the determination error reduces to 0.44% compared to 1.77% before applying the PCA. In addition to improved accuracy, the system also benefits from the fundamental principle of the PCA as it has shown that the original data of beam deflections can be compressed into a more meaningful form and a smaller number of transformed data are required to infer the contact parameter.