Research On Tactile Perception Based On Bioinspired Microstructures

Touch plays an irreplaceable role in our daily life,so as to robotics.From the perception of the conduct principles,the tactile sensors could be classified as resistive,capacitive,photoelectric and so on.By utilizing these different tactile sensing princi-ples,these tactile sensors possess different characteristics and advantages in different scenarios.In this paper,we make our main focus on the core issue of robot tactile sens-ing.In our paper,inspired by the bionic micro-structure in nature,a flexible tactile sensor which is based on this structure is designed,and then simulations and theoretical studies are launched in detail for in-depth study.Results show that the tactile sensor studied in this topic has the characteristics of low hysteresis,high repeatability,no drift and high signal-to-noise ratio.The main contents of this paper are as follows:Design and fabrication of a flexible tactile sensor for robotics based on bionic surface microstructures.Starting from the widely existing micro-structures in nature,this paper studies the formation mechanism of biological surface structural color,ex-plores the process of fabricating micro-structures on the surface of elastic silica gel polydimethylsiloxane(PDMS)and forming visible structural color,and analyses the change of surface structural color of elastic materials under external contact force from the perspective of theoretical model and simulation respectively.The structure and color change characteristics of the surface are discussed as tactile sensors.Information acquisition and processing method of bionic flexible tactile sensor.The signal characteristics of flexible tactile sensor based on bionic micro-structure are analyzed in detail;the contact position of a single point is detected based on image pro-cessing method,and the continuous position is tracked;the training samples of machine learning are made by collecting and marking a large number of tactile image signals;and the sensor image information is obtained based on ResNet depth residual network learning framework.No.1 is used to learn,and the learning model is used to perceive and predict the contact position of the test samples.Error analysis of bionic flexible tactile sensor.The calibration method of bionic flexible tactile sensor is studied,and a calibration platform is designed according to the structural characteristics of the tactile sensor and the principle of signal generation.After discrete calibration,the interpolation method for continuous arbitrary position is discussed,and the error of the interpolation method and its causes are discussed.This paper studies two perception methods of tactile sensor for the shape of small simple structure,complex structure and surface structure,and analyses the application scenar-ios and errors of these methods.Integrated design and test of bionic flexible tactile sensor.Based on the princi-pie prototype,the research scheme of miniaturization and integration of tactile sensor is studied;aiming at the miniaturization structure scheme,the casting die and manufactur-ing process of elastic film with surface micro-structure are designed,and the universal calibration method is studied;the precision of miniaturization tactile sensor is measured based on deep learning model;and the miniaturization tactile sensor is designed.Error analysis.The practical application of flexible tactile sensor for robots based on bionic mi-crostructures.The feasibility and stability of the tactile sensor integrated into the end of the dual-arm robot are studied.Based on Euler video amplification technology,the sensing method under the condition of micro-contact force is studied,which extracts and senses the tiny signals which can not be extracted under normal conditions from the contact image signals of the sensor.Finally,on the basis of the research in this paper,the future research direction of robot tactile sensor based on surface bionic micro-structure is further prospected,and the sensing ability of this type of robot tactile sensor is further excavated and studied.