Font Size: a A A

Design And Application Of Flexible Robot Skin Based On Airbag Array

Posted on:2021-03-18Degree:MasterType:Thesis
Country:ChinaCandidate:Z R YanFull Text:PDF
GTID:2518306107474664Subject:Engineering
Abstract/Summary:PDF Full Text Request
Currently,the deployment of robotics on industrial production lines has greatly increased production efficiency.However,the safety problem of human-robot collaboration has not been solved so far.Exploring the application research of emerging collaborative robots in various occasions is an important development trend of industrial robot technology.To realize human-machine collaboration,an important prerequisite is that the robot must have a certain tactile perception ability,and the robot skin with flexible features can mimic human tactile perception,recognize objects and uncertain dynamic interaction scene information,and ensure the intrinsic safety of human-machine collaboration,and it has a very considerable application prospect.Therefore,it is of great significance to carry out research on the skin tactile perception of flexible airbag array robots for intelligent human-computer interaction applications.The main research contents of this paper are as follows:(1)A flexible robot skin tactile sensor based on airbag array was designed.Inspired by human hand tactile bionics,combined with the skin surface-nerve system tactile sensing features,the structural design and working mechanism of airbag skin sensors were studied;and the sensor structure and the force under different working conditions were simulated through finite element analysis software,While laying the data foundation for pattern recognition applications,verified the rationality of the skin sensor structure design and size.(2)A new CRXL fusion algorithm is proposed,which can complete the recognition application of airbag skin sensors under different working conditions.The tactile information obtained by simulation is used to extract geometric feature information to complete the training and prediction of machine learning network models.The algorithm is a fusion of decision tree,random forest,XGboost,and Light GBM.It integrates XGBoost model complexity control,robustness of noise data,optimization of Light GBM memory and parallel computing,simple and stable advantages of random forest,and new CRXL training The pattern recognition method has higher accuracy,reliability and scope of application,and solves the problem that traditional algorithms have limited prediction capabilities and cannot accurately identify objects.After synthesizing the performance evaluation results of the four single models and comparing the CRXL fusion algorithm used in this paper,it is found that the fused algorithm model can effectively improve the success rate of detecting the type of contact objects.(3)According to the structural characteristics of the airbag skin sensor and the tactile sensing principle,the relevant materials selection,packaging and characteristic analysis of the skin tactile sensor were carried out.Based on the simulation results,the structural dimensions of the airbag skin sensor were determined,and the fabrication and packaging of each component of the airbag skin sensor were studied.Secondly,a static characteristic test platform for skin sensors was established,and a static characteristic test experiment based on an airbag array flexible robot skin sensor was carried out to explore the characteristics of force-displacement and variable stiffness.(4)An airbag array-based flexible robot skin sensor tactile information sensing system is designed to conduct experimental research on pattern recognition based on human-machine collaboration mode.The Lab VIEW host computer programming software is used to build the experimental test platform of the tactile sensing system,and perform operations such as signal acquisition and processing.In order to verify the effectiveness and accuracy of the airbag skin sensing system,experiments were performed with different shapes of contact and different human-machine collaboration objects.The objects and objects with different geometric characteristics are used to knock,touch,and squeeze the airbag skin sensors respectively to obtain experimental data.Finally,the skin sensor recognizes and detects the pressure signal of the contact object based on the CRXL algorithm.The experimental results show that the skin sensor based on the airbag array proposed in this paper can effectively complete target recognition in human-machine cooperation mode,and demonstrates the feasibility of the application of airbag skin sensors in tactile perception.
Keywords/Search Tags:electronic skin, software robot, bionics, human-machine collaboration
PDF Full Text Request
Related items