Mathematical transforms and correlation techniques for object recognition using tactile data

The authors discuss an object recognition system which uses tactile array data and mathematical transform techniques. The sensor consists of a grid of sensitive sites arrayed in a 10×16 orthogonal array. An object is placed onto the sensor and an instantaneous reading of the tactile image is transferred to an IBM-XT computer where the data are processed using two-dimensional mathematical transforms and correlation algorithms. Several test results are presented to illustrate the performance of the system