The Research And Design Of Robot Artificial Skin Based On Low Density Tactile Array

Tactile sensation is not only a complement to vision,but also has a strong sensitivity,which can directly measure the various attributes and characteristics of objects.Artificial skin has become the main structure of tactile perception,and has become a research hotspot in the field of robot tactile.The essence of artificial skin is a kind of flexible tactile sensor array.In this paper,the research and design of artificial skin flexible tactile sensation are studied.Several algorithms for calculating the pressure distribution of haptic perception are proposed.Two kinds of flexible haptic tactile sensor array are designed and analyzed theoretically and simulated by finite element method.The tactile array was made and the experiment was completed.Finally,these tactile sensors can be used for quantitative detection and surface localization of the vector pressure on the external surface of the skin.The main research contents are as follows:1)Read the relevant literature,analyze the current situation of robot tactile sensors and robot artificial skin,and summarize the trends and shortcomings of robot artificial skin development in the future.Combined with a variety of tactile sensor array and tactile perception algorithm,a low pressure array robot skin pressure tactile program is proposed.2)A triangular pressure array is designed and encapsulated into Polydimethylsiloxane(PDMS)to form an artificial skin structure unit.A positive pressure distribution detection algorithm based on the geometric relation of triangle skin is proposed.The mechanical properties of the artificial skin were analyzed by using the theory of elasticity,and the analysis results were verified by ANSYS software.Finally,the feasibility of the artificial skin on the detection of the positive pressure distribution is verified by tests.3)In order to improve the detection accuracy of the above skin pressure,another pressure location algorithm is proposed.The algorithm is based on BP neural network and support vector machine(SVM).Using the above experimental data,1600 groups of training samples are constructed by the normal distribution approximation.MATLAB is used to construct and train BP neural network and support vector machine(SVM),and the feasibility of the algorithm is verified.4)Two algorithms and experimental results were designed.A cross symmetrical pressure array based on finger structure was designed to realize the detection of vector pressure.In the same time,the mechanical analysis of the structure is carried out by using the theory of elastic mechanics,and the analysis result is verified by ANSYS software.An algorithm for detecting the horizontal force component of multidimensional forces is also proposed.Finally,a physical object is fabricated and its performance is verified.Finally,it summarizes the research situation of the whole subject,analyzes the current research results and research deficiencies,and puts forward some improvement measures...