Study On Slip Detection Modeling And Surface Recognition Of Flexible Tactile Sensor Array

With the rapid development of the intelligent robots in unstructured environments,their demand for tactile sensing technology is increasing.Mounted on the robotic hands,the tactile sensor array can be used for tactile sensing such as slip detection and texture recognition,which improves the intelligent level of the robots.Supported by National Nature Science Foundation of China(Grant No.51575485)and the Natural Science Foundation of Zhejiang Province for Distinguished Young Scientists(Grant No.R19E050011),the mechanical model of flexible tactile sensor array was carried out,based on which the slip detection method and surface recognition technology were studied.Firstly,the mechanical model of the flexible sensor unit was developed based on Beam-Bundle theory.The Localized Displacement Phenomenon on the contact surface was then studied.Secondly,the surface morphology of different objects was reconstructed based on statistical self-similar fractal theory.Combining with the established mechanical model,the influences of surface texture on the force sustained by the sensing unit during the gross sliding phase was analyzed.Thirdly,works including slip detection method based on wavelet transform theory,surface texture information extraction method based on Phase Delay Algorithm and surface material classification research based on the neural network algorithm were carried out.Corresponding experimental platform was established.Through simulation analyses and experiment verifications,the validity of the proposed detection methods was demonstrated.In Chapter 1,the background as well as the significance of this dissertation was introduced.The research status at home and abroad,including tactile sensor types,their mechanical modeling methods and tactile information recognition method,was elaborated.By analyzing the problems that need to be further studied in the current research work,the main research content and the overall framework of this dissertation were determined.In Chapter 2,the static simplification of the force condition of the flexible tactile sensor unit during the initial slippage phase was conducted firstly.Then the analytical model for contact surface Localized Displacement Phenomenon analyzing was established based on Beam-Bundle model and finite element method.The result was discussed with the help of the FEA software.Based on the iterative algorithm,the revised methed for calculating the tangential force of the beam element was proposed,which improved the accurary of the model output.The established mechanical model could theoretically analyze the time sequence of the displacement happened on any point on the contact surface as well as its size.In Chapter 3,the surface reconstruction of surface with different texture was conducted by using the W-M function and basic statistical methods,which based on the statistical self-similar fractal theory.Combining with the simplified 3-D Beam-Bundle model of the tactile sensor unit,the calculation and analyzation of the effects of surface texture on the mechanical behaviors of the sensing unit during the gross sliding phase was carried out.The characteristics of the signal are analyzed in the time domain as well as the frequency domain.In Chapter 4,the finite element model of the flexible tactile sensor array was built in ABAQUS software environment,based on which the mechanical behaviors of the sensor array during the initial slippage phase was analyzed,and the slip detection method based on wavelet transform theory was verified.The integrated experimental platform for sensor array performance testing and surface recognition was then established.Based on corresponding experimental studies,the effects of the number of decomposition layers,wavelet basis functions as well as the detection areas of sensitive materials on sensing array slip detection were discussed.In Chapter 5,researches on surface recognition application based on detection signals from flexible tactile sensor array were carried out.The Phase Delay Algorithm for regular texture surface information extraction was proposed based on theoretical analysis and FEA results;the surface texture classification method based on the neural network algorithm was also proposed.Corresponding verification experiments were conducted on the established integrated platform,which validated the practicability of the above methods.In Chapter 6,the main research contents of this dissertation were summarized,and the future research work was prospected.