Robust real time material classification algorithm using soft three axis tactile sensor: Evaluation of the algorithm

Materials and textures identification is a desired ability for robots. Developing such systems require tactile sensors that have enough sensitivity and spatial resolution, and the computational intelligence to meaningfully interpret sensor data. This paper introduces a texture classification algorithm utilizing support vector machine (SVM) classifier. Data taken from a novel three axis tactile sensor that utilize magnetic flux measurements for transduction was used to obtain the three dimensional tactile data. Frobenius norm calculated from the covariance matrix of the above data and the mean values of the three dimensional sensor data were used as features. Palpation velocity and small vertical load variances had minimum influence on the proposed algorithm. We have compared this algorithm with two other classification methods. They are: classify using the feature spatial period that is calculated from principal frequencies of the textures/material, and classify using neural network classifier with special properties of each material's tactile signals as features. For eight classes of material, the proposed algorithm performed faster and more accurately than the comparators when the scanning velocity and the vertical load varied.

I. Introduction

Biological tactile system is complex. It has hundreds of sensors under a square centimeter of the skin that detect different stimuli. These detect pressure, vibrations, pain, and temperature modalities [1]. Even though tactile sensors are distributed all over the body, humans mostly use their fingers to explore the environment and acquire tactile information. Data from these sensors taken at different internal impedence levels [2] fused with vision and audition information are used to comprehend the surrounding environment. Two main activities benefited from tactile sensing are: humans interpreting tactile signals and using prior experience to classify materials and textures only by touching the surfaces, and grasp objects by applying just enough grip force to hold objects without slipping. Both tasks are important for a robot if it were to operate autonomously in unstructured environments. In attempt to develop artificial systems with the same capabilities as the human hands and fingers, researchers have mimicked human anatomy.