Research On Precise Positioning And Motion Control Of Hydraulic Serial Manipulator Of Demolition Robot

The demolition robot is a type of special engineering machinery and equipment that is used in the fields of metallurgy,mining,rescue and disaster relief,and hazardous chemical and nuclear industries.This robot was developed to work in unstructured environments,and its working object and load vary significantly.The mathematical model and the control method directly influence the motion dynamic characteristics and the positioning accuracy of its serial manipulator.Therefore,the research on a laser-guided demolition manipulator with accurate positioning would help to improve the ability of autonomous and accurate operation.However,the current researches on positioning control of serial manipulators mainly focuses on industrial manipulators,while the related research on the engineering hydraulic manipulators is rare.This doctoral dissertation analyzes and summarizes the key technology of the serial manipulator positioning and the research status at home and abroad.Further,the hardware and software system of computer measurement and control of laser-guided demolition manipulator motion was constructed.Furthermore,the strategy of uncalibrated localization is proposed for the first time,which is to implement local accurate approaching based on the global close to the target point.In this dissertation,the key technical issues focused on are the global inverse calculation model,positioning error elimination,and the joint variable load nonlinear control.The main work and achievements are as follows:(1)The inverse mapping model of the dynamic gravity differential evolution neural network is established.To avoid the problem of dependency of the precision of inverse calculations of large manipulator on the parameters and calibration accuracy and multi-solution problems,a direct mapping model from the laser measurement to the joint hydraulic cylinder driving space is constructed using the artificial neural network(ANN).Based on the gravity-driven differential evolution(GDE)which adjusts the convergence speed of the population by gravity to optimize the weights of ANN,the concept of expectation curve of population average distance is proposed for the first time.Further,the gravity constant is adjusted dynamically by the difference between the expected value and the actual value of the population average distance in order to realize the convergence of the differential evolution population according to the expected curve.From the study,it was found that the dynamic gravity-driven differential evolution(DGDE)could significantly accelerate the convergence speed while avoiding the local convergence of DE,and provide an effective inverse solution of initial positioning for the autonomous positioning of the manipulator.(2)A neighborhood differential motion approaching method for accurate positioning of demolition manipulator is proposed.In the case that the positioning error of the end of the manipulator still exists after the model calculation and joint motion control,a successive approximation method is proposed to eliminate the positioning error,and a first-order differential linear motion equation of the manipulator is derived based on the position error.Further,the neighborhood space motion is converted into linear equations to solve the problem,and the position error is eliminated.Furthermore,the error and influencing factors of the method are analyzed.The simulation results demonstrated that this method could effectively reduce the positioning error of the manipulator through successive approaching,and approach the target point with accuracy.(3)A sliding mode controller for a variable load based on the linear extended state observer is designed.In order to satisfy the high-precision position control of the joint hydraulic cylinder,a theoretical formula and variation range of the natural frequency of the joint hydraulic cylinder that was changed with the attitude of the manipulator are obtained based on the kinetic energy principle.Further,a fifth-order nonlinear dynamic model of the hydraulic control system with variable parameters is derived.A sliding mode controller for variable load using higher-order linear extended state observer(LESO)to estimate state variables is proposed,which avoids the problems of physical measurement of high-order differential variables and jitter in high-order sliding mode control.Moreover,the stability of the control algorithm is proved by using the Lyapunov method,and the simulation results presented that the controller could meet the requirements of robustness and dynamic characteristics of joint control.(4)The experiment of position control of the joint hydraulic cylinder and positioning of the manipulator is carried out.The joint hydraulic cylinder position control and laser-guided positioning of the manipulator are tested.The result further proved that the effectiveness of the idea of uncalibrated autonomous positioning of the manipulator.