Design Of Variable Stiffness Actuator And Its Compliance Control

Variable stiffness actuator(VSA)has been widely developed by scholars because of their adjustable stiffness and inherent physical compliance.The rapid response of stiffness regulation is of great significance to improve the performance of complinace control for VSA.At present,the rapid stiffness regulation is only realized from the perspective of motor control.However,this method will add additional burden to the motor when the acceleration or deceleration is large,and even cause damage to the motor and reducer.Based on the lever principle,this thesis designs a VSA which can adjust the position of lever fulcrum and spring module simultaneously,so as to ensure the compactness of the actuator and realize the rapid stiffness regulation,therefore reducing the requirements on the motor.In this thesis,based on the proposed VSA,the key technologies of external force estimation based on current-deflection fusion,adaptive compliance control against sudden change of external force,and compliance control considering the minimum jerk under uncertain operation are studied.The full text mainly includes the following aspects:A novel VSA is developed which is capable of regulating stiffness faster based on the variable lever arm mechanism from the mechanical design perspective,to solve the limited stiffness range and fixed stiffness regulation characteristics.The characteristic of fast stiffness regulation is realized by changing locations of pivot and spring simultaneously with a stiffness adjustment mechanism,where the stiffness changing rate is designable through a cam-based panel.By introducing the linear guide pairs for the location variations of pivot and spring,the internal friction can be greatly reduced to enhance the energy efficiency.To evaluate the performance of the proposed VSA,the point-to-point control strategy is developed which contributes to a higher tracking accuracy and oscillation attenuation at both the start and end points of the trajectory.Additionally,the performance of torque controllability is also verified through experiments.These excellent capabilities enable the proposed VSA to be qualified for constructing a robotic arm towards service applications.In order to improve the performance of external force estimation for VSA under both quasi-static and dynamic stiffness situations,a novel external force estimation method is proposed based on the current-deflection fusion.Firstly,an improved friction model is proposed,which can avoid the discontinuity when the velocity changes its direction.Unified dynamic models are constructed for the position-driven module and stiffness regulation module respectively,and the dynamic parameters are identified to reveal the mapping relation between the current and external force.Such unified dynamic models contain all the frictions of the motor,reducer and other transmission components,further improving the performance of parameter identification.Then the external force estimation based on the current-deflection fusion is developed to enable accurate and reliable estimation quality for both the quasi-static stiffness and dynamic stiffness situations.Finally,an extensive set of experiments are carried out to demonstrate the efficacy of the proposed methods.Considering the safety of physical human-robot interaction,a novel continuous adaptive compliance control method is proposed based on the stiffness-position combined control.The proposed method integrates both modes of robot-in-charge and human-in-charge in the human-robot interaction into a single controller,making the transition between the two modes smooth and stable.The stiffness region control is developed to enhance the compliance in human-robot interaction and avoid transgressing the limits of deformation simultaneously.An online estimation method for the human motion intention is developed to generate the desired trajectory based on the estimated force.Then the adptive controller is constructed with weighted factor to make the transimission smooth.The dynamic stability of the closed-loop system is theoretically proven by using the Lyapunov method.Simulations are presented to illustrate the performance of the proposed controller.Targeting the minimum jerk of the trajectory during the compliance control,a novel trajectory planning methodology considering the minimum jerk is proposed to improve the smoothness of trajectory.The controller involves the stiffness-position combined control,where the stiffness control is consistent with that of Chapter 4.Firstly,the trajectory of physical human-robot interaction is divided into two different stages,namely interaction stage and recovery stage.The former is composed of multiple piecewise curves with fifth-order polynomial while the latter is a curve with fifth polynomial,enabling the trajectory minimum jerk.Then the updated law for interaction stage is introduced to make sure the robot activel y follow the intention of human.An adaptive impedance controller is proposed with consideration of uncertainties in the actuator dynamics and the stability is proved based on Lyapunov theorem.Finally,through the simulations of the proposed VSA,the effectiveness of the proposed trajectory planning algorithm is proved.The experimental setup is established based on the proposed VSA to to evaluate the efficacy of compliance control strageties.Firstly,the experiment of continuous adaptive control method for physical human-robot interaction is carried out.In the experiment,the stiffness region control algorithm and the estimation of human intended motion algorithm are verified.The continuous adaptive control method is validated,comparing the results of hard-swith controller when the actuator changes its motion from robot-in-charge mode to human-in-charge mode.The Head Injury Criterion(HIC)is utilized to evaluate the effectiveness of the stiffness region control for improving the safety during the physical human-robot interaction.Through the experiment of compliance control algorithm in interaction motion mode,the effectiveness of stiffness region control algorithm.The effectiveness of weighting factor in the mode conversion is verified by experiments.Then,the trajectory planning methodology considering the minimum jerk is verified by experiments.The research in this dissertation provides solutions for the design of VSA,the external force estimation and stiffness-position fusion compliance control based on VSA.It has important theoretical and practical significance for improving physical security of the VSA in the process of interaction with the environment.