On The Precision Compensation Technology Of 6-DOF Measurement System Based On Extreme Learning Machine

With the development of science and technology,six-degree-of-freedom position and orientation measurement system has an urgent demand in the fields of robot,electromechanical equipment,large electronic equipment,etc.Effective real-time measurement technology of six-degree-of-freedom position and orientation is crucial to the measurement effect.In this thesis,based on the principle of parallel mechanism,a sixdegree-of-freedom measurement system is built by using cable displacement sensor.Aiming at this measurement system,the key technologies such as structure optimization,establishment of pose error compensation model are carried out.The validity of the proposed position and orientation error compensation technique is verified by the experimental results.The main work of this thesis is as follows:(1)The design scheme and system composition of the six-degree-of-freedom position and orientation measurement system is introduced,including mechanical structure,measuring element and data acquisition and processing module.In addition,the mathematical model of this system is established,including forward and inverse kinematics algorithm of 6-UPS parallel robot.(2)In the process of the calibration based on inverse kinematics,there is ill-posed problem of coefficient matrix involved with optimization model of structural parameter is proposed.So,the condition number of orthogonal of coefficient matrix is considered as the objective function,and the limited length of each leg and limited angles of universal and spherical joints are taken as constraints in the optimization model of structural parameter.The standard immune algorithm and the metropolis adaptive immune algorithm are selected to solve the model,and the effectiveness of the proposed algorithm is verified by comparison.Based on the optimal structural parameters,the measuring range of the 6-DOF measurement system is analyzed.(3)In this thesis,the physical prototype is built based on the cable displacement sensor.When the cable has a certain tilt angle,the structural characteristics of the outlet inevitably lead to the measurement error of the cable length.In order to compensate this error and weaken its influence on the final measurement accuracy,the outlet error compensation model of the cable sensor is built in this thesis.Secondly,this thesis established the calibration model based on inverse,BP neural network position and orientation error compensation model,and the adaptive immune algorithm based on metropolis to optimize extreme learning machine position and orientation error compensation model.In addition,the selection of key parameters such as hidden layer node is studied.(4)Finally,experimental design of the three models is carried out in this thesis,and summarizes the advantages and disadvantages of the three algorithms.The experimental results show that the proposed position and orientation error compensation model has favorable capability than the other two error compensation models.