Research On Actuation And Control System Of Gripper Actuated By Smart Material

In recent years,development and use of smart materials have become one of the most important aspects in modern high technology and new material development.Smart materials are a type of materials which can sense external signal stimulation and actuate motion by self,and they can be applied as novel actuator in robotics field.Smart material actuators become important actuating function device to develop bionic and soft robotics,and the fabrication,modeling,control and application technology have attracted more and more attention from different researchers.The researches of this thesis focus on smart material actuator as end-effector in robotics system application,and the main contents are as follows.(1)A novel ionic polymer metal composite(IPMC)two-finger gripper is fabricated aimed at the end-effector gripping task in micro operation system.One finger is IPMC actuator which is used to generate gripping force,the other finger is piezoresistive film sensor which is used to measure gripping force.Studying on actuation characteristics of gripper in gripping operation process,the gripper reveals obvious hysteresis nonlinear characteristics.This paper applies Kriging method to descript IPMC output force model for the first time.According to the Kriging model,Kriging prediction(KP)control system,Kriging prediction-wavelet filter(KPWF)control system,and simultaneous perturbation stochastic approximation(SPSA)adjusting proportion integration differentiation(PID)parameters with KPWF feed-forward compensator control system are designed and conducted.The physical experimental results demonstrate the effectiveness of SPSA adjusting PID parameters with KPWF feed-forward compensator control system on tracking gripping force of novel IPMC two-finger gripper in gripping foam ball task.(2)An integration of actuating and sensing IPMC gripper is designed aimed at the end-effector fixing task in micro operation system.One finger is fixed metal sheet which has the function of supporting object,the other finger is PF sensor which is used to measure gripping force.There is an IPMC actuator outside of piezoresistive film sensor to generate fixing force,and this force acts on the object through PF sensor.Studying on actuation characteristics of gripper in fixing operation process,the gripper reveals obvious creep nonlinear characteristics.This paper applies Hammerstein model to descript IPMC output displacement model innovatively.According to the Hammerstein model,a state feedback robust control system based on state observer with gripping force compensator is designed and conducted.The physical experimental results demonstrate the effectiveness of above control system on tracking fixing displacement of integration of actuating and sensing IPMC gripper in fixing foam ball task with force compensation.(3)A modular humanoid five-finger hand actuated by shape memory alloy(SMA)actuator is designed.The modular design method realizes combination of SMA actuator,bending sensor,temperature sensor,and transmission mechanism delicately,making the humanoid hand with properties of integration,miniaturization,lightweight and other advantages.The control system of humanoid hand is developed by using reinforcement learning intelligent algorithm,and we select temporal difference strategy to establish stable grasping manipulation control system.The bending value is used as state,the temperature of SMA actuator is used as state to protect the whole system,and the duty cycle of PWM is used as action.?-greedy algorithm completes the action search.The table of reward and punishment values is set up which conform to the law of human hand movements.Online learning Sarsa algorithm and offline learning Q-learning algorithm are adopted to establish the control system,and the physical experimental grasping tests on objects demonstrate the effectiveness of humanoid five-finger hand and its control system.(4)The pneumatic soft gripper(PSG)is fabricated from liquid silicone rubber by shape deposition manufacturing technology.Modeling analysis is done aimed at a type of universal air cavity form of PSG in working deformation process.We select stretch ratio and stiffness two parameters to evaluate model.The experimental results show that stiffness has a positive correlation relationship with load,stretch ratio has a negative correlation relationship with shear elasticity of material and thickness of silicone rubber layer,at the same time,stretch ratio also has the relationship with stretch coefficient,making stretch coefficient and curvature present a certain nonlinearity relationship with inflation pressure.All above results provide theory and reference to fabricate PSG.The bending sensor and pressure sensor are applied in control system of PSG,and bending characteristics and the relationship between bending and spherical object size are studied in working process of PSG,which makes PSG have the functions of spherical object diameter prediction and adaptive stable grasping.Finally,PSG is affixed on the end joint of home-made palletizing robot,and the total system can accomplish sorting of fruits task.