Model-free Adaptive Smooth Sliding Mode Impedance Control Of Parallel Robot For Cluster Fruit Clamping

During recent years,fruit production has increased rapidly in China.The structural shortage of agricultural labors appears and the cost of traditional manual sorting operation is rising constantly under the tendency of urbanization and population aging.It is the trend of modern agricultural development to realize automatic and nondestructive fruit sorting effectively,reduce labor intensity and improve economic benefits through robots.In the research of automatic nondestructive sorting of fruit,compared with independent fruit such as apples,pears and pineapples,cluster fruit such as grapes,longan and lychee own irregular shapes,uneven weight distribution and fruit grain stacking.So it is difficult to meet the requirements of high clamping stability and accuracy for automatic nondestructive sorting of cluster fruit which makes the research of automatic nondestructive sorting of cluster fruit more challenging.The parallel mechanism owns the advantages of stable structure,high stiffness,high precision and good dynamic performance.It is expected to better realize the nondestructive sorting operation of cluster fruit which requires high clamping stability and accuracy.A cluster fruit clamping parallel robot is developed by our research group based on the superiorities of parallel mechanism to implement nondestructive clamping of cluster fruit.Achieving high performance control of the clamping mechanism of parallel robots is one of the key problems to realize nondestructive clamping of cluster fruit.In order to realize the nondestructive clamping of cluster fruit by parallel robots,a position-based impedance control method is introduced in this paper.This method adopts the outer loop of impedance controller and the internal loop of position controller to form a cascade control structure.The contact clamping force errors between expected clamping force value and actual contact clamping force value acquired from the force sensor are used as the input of the outer loop of impedance controller and its output can be realized by the reference trajectory correction to the internal position control loop to realize the clamping force control.The control method is easy to implement,but its control performance is directly related to the position control performance.Therefore,it is necessary to study the position control method of the clamping mechanism to improve the control performance of the parallel robot for the clamping of cluster fruit.The existing control strategies for position control of clamping mechanism are mainly divided into controls based on the kinematic model and controls based on the dynamic model.The latter control strategy takes dynamic features of hybrid robots into consideration and its non-linear dynamic factors can be compensated by the controller,which can theoretically achieve better control performance.Thus,this paper focuses on the dynamic control of the clamping mechanism of parallel robots.Compared with the rotary clamping finger structure,the complexity of the dynamic model for the parallel mechanism makes it difficult to establish accurate dynamic model and to realize real-time dynamic control due to its translational clamping finger structure.Therefore,the time-delay estimation method for dynamic modeling of clamping mechanism of parallel robot is proposed to obtain the system model online.The time-delay estimation errors are the inherent performance limitation caused by time delayed signals and the complex uncertainties like friction,external interference and chattering of traditional sliding mode control methods exist in the practical engineering application system.So a smooth sliding mode control method based on time-delay estimation dynamic model of parallel robot for clamping mechanism is proposed.It is necessary to select the maximum switching gain to overcome the system uncertainties under the most serious situations to ensure the robustness of sliding mode control method.However,excessive switching gain usually leads to high energy output and actuator saturation which will decrease the performance of the control system.Considering the difficulty of measuring the inertial gain directly which has a great influence on the dynamic characteristics of the system and changes with the movement state and load of the robot,the existing adaptive methods of dynamic parameters own more adjustment parameters.For solving the above problems,an adaptive rule with less adjustment parameters is further designed based on the smooth sliding mode control of the parallel robot clamping mechanism of the time-delay estimation dynamic model and a model-free adaptive smooth sliding mode impedance control method is proposed.The adaptability of the system to the clamping speed and the change of the clamping load is improved and the large switching gain of sliding mode control may be avoided simultaneously through the adaptive adjustment of the controller's inertia gain.Then the high-performance control of the parallel robot clamping mechanism and the nondestructive clamping of the cluster fruit of parallel robot could be realized.Firstly,the development and application of parallel robots are summarized,and the research status of clamping mechanism and its control methods are systematically analyzed.The structure characteristics and working principle of the clamping mechanism of the parallel robot are introduced in detail.Then,the kinematics analysis of the clamping mechanism of the parallel robot is carried out,and the relationship between the rotation angle of the active joint screw and the position of the clamping finger of the clamping mechanism is obtained,and the desired trajectory of the clamping mechanism is determined.Afterwards,the dynamic analysis of the parallel manipulator clamping mechanism is carried out.The dynamic model of the parallel robot clamping mechanism is established by Lagrange method as the controlled object of the simulation system.The dynamic model of the parallel robot clamping mechanism is acquired by using the time-delay estimation technology online,which lays a foundation for the subsequent controller design based on the time-delay estimation dynamic model.On this basis,in order to solve the time-delay estimation errors and overcome the complex uncertainties like friction,external interference and chattering of traditional sliding mode control methods exist in the practical engineering application system,a smooth sliding mode control method based on time-delay estimation dynamic model of parallel robot for clamping mechanism is presented.For avoiding the large switching gain of sliding mode control which is used to ensure system robustness,a model-free adaptive smooth sliding mode impedance control method is proposed by designing an adaptive rule with less adjustment parameters.And the stability of the proposed control algorithm is proved by Lyapunov stability theorem.Under the same impedance parameters,the model-free adaptive smooth sliding mode impedance control method is compared with the smooth sliding mode impedance control method by MATLAB simulation.Finally,the construction of the control system and the experimental platform of the cluster fruit clamping parallel robot system are completed.And the proposed control method is further applied to parallel robot prototype to carry out cluster grape clamping control experiment.The simulation and experimental results verify the effectiveness and superiority of the proposed model-free adaptive smooth sliding mode impedance control method.