Research On Deformation Control And Motion Mechanism Of Soft Robot Based On SMA Artificial Muscle

Due to the soft characteristic of the soft robot and its good flexibility and adaptability,the soft robot has attracted the attention of many researchers.The soft robot has great potential in promoting the development of science and technology as a multi-disciplinary field of engineering,material science,mathematics and so on.This dissertation focuses on the deformation control and movement mechanism of the soft robot actuated by shape memory alloy(SMA)artificial muslces,and carries out researches on the bionic design and manufaction technology of the SMA artificial muscle,modeling and deformation control of the SMA artificial muslce,multi-gait locomotion of the soft robot actuated by the SMA artificial muslce,and path planning of the soft robot.The main research contents and results are summarized as follows:1)Through the bionic study of muscle,the SMA artificial muscle actuator is designed and it serves as basis for rapid design of the soft robot.The SMA wire with high energy density,low noise and compact structure as is selected as the skeleton of the actuator,and then two kinds of SMA artificial muslce actuators are fabricated,combined with 3D printing technolgy and intelligent material technology:unilateral SMA artificial muslce actuator and bilateral SMA artificial muscle actuator.M-type configuration of the SMA wire skeleton enables the unilateral SMA artificial muscle actuators to have a good advantage in compact,output strain and external wires.The unilateral SMA artificial muscle actuator can only be in unilateral movement,and its initial state is a curved state.The unilateral SMA artificial muscle actuator is optimezed by the experiments of response frequency and bending amplitude.The end driving force of the unilateral SMA artificial muscle actuator is tested,and the thermodynamic analysis of the unilateral SMA artificial muscle actuator is performed by the multi-physics simulation software COMSOL.The bilateral SMA artificial muscle actuator consists of two pasteable actuator modules with four SMA channels symmetrically attached to both sides of the flexible substrate structure(spring steel plate).The initial state of the bilateral SMA artificial muscle actuator is horizontal,and can move on both sides of the elastic substrate.With the multi-channel SMA wire in parallel,the bilateral SMA artificial muscle actuator can output different bending amplitude.In addition,the end driving force of the bilateral SMA artificial muscle actuator is also tested,and the thermodynamic analysis of the bilateral SMA artificial muscle actuator is carried out by the COMSOL.2)The model of the SMA artificial muscle actuator is established and analyzed.The basic characteristics of the SMA wire are tested by differential scanning calorimetry(DSC)and dynamic mechanical analysis(DMA)experiments:phase transition temperature and mechanical performance.Based on the constitutive equation of SMA,the phase transformation kinetics equation,thermodynamics equation and kinemacits equation of the actuator,the state equations of the unilateral SMA artificial muscle actuator and bilateral SMA artificial muscle actuator are established,respectively.Based on the state equations,the SMA artificial muscle actuators are simulated by Matlab/Simulink.The influence of the PWM signal,distance between the SMA wire and the PVC plate,equivalent stiffness on the motion performance of the SMA artificial muscle actuators is analyzed.This not only provides guidance for the optimization of the SMA artificial muscle actuator,but also plays an important role in the regulation of SMA artificial muscle actuator control parameters.Based on the state equations,the step response of the SMA artificial muscle actuator is also tested.In addition,the characteristics of SMA artificial muscle actuator and DC motor are compared to improve the performance of the SMA artificial muscle actuator.3)Based on the SMA artificial muscle actuator,the soft robot is designed and its motion mechanism is modeled and optimized.According to the SMA resistance self-feedback strategy,the self-sensing feedback method of the SMA artificial muscle actuator is described,and the self-sensing feedback tracking performance of SMA artificial muscle actuator is tested.Based on the unilateral SMA artificial muscle actuator,four actinomorphic soft robots with three to six arms are desgined.Combining the model of the SMA artificial muscle actuator with the pseudo-rigid body model,the motion model of the actinomorphic soft robot is established.Through the optimization experiment of the gait parameters,the three-arm actinomorphic robot with certain gait parameters has the best performance and achieves a maximum stride length of 75 mm.Then,multi-gait locomotion experiments are conducted on the three-arm soft robot,including multi-medium(such as seimi-diving environment,sand,transition environment and so on)gait locomotion,locomotion with inoperable arms,and jump locomotion.The three-arm soft robot can move in multiple environments,and show different movement performances.The three-arm soft robot in four typical scenarios,where some arms are inoperable exhibits different movement characterisitcs,and it can help identify the status of the soft robot in time.By applying instantaneous large voltage to the SMA wire artificial muscle actuator,the three-arm soft robot can achieve jump locomotion(jumping up to 67 mm,about 1.5 times the height of the three-arm soft robot).Combining the 3D printing technology and the bilateral SMA artificial muscle actuator,a soft robot fish is designed.With the reciprocating motion of the bilateral SMA artificial muscle actuator,the robot fish can achieve basic movement,including cruise.The influence of the number of channels,the driving voltage and the swing frequency of the bilateral SMA artificial muscle actuator is analyzed,respectively.It provides guidance in enhancing the performance of the soft robot fish.4)A local path planning strategy based on the obstacle collection is proposed.Based on the information captured by the laser range scanner,two kinds of obstacle models are established:actual obstacle model which is utilized to determine the traversability of the environment,and secure obstacle model which is used to determine the heading angle of the mobile robot.Through a merger strategy,the corresponding actual obstacle group and secure obstacle group are further obtained.The willing gap is selected by combining"the-closest-valley-wins" scheme and motion-context,and then update the heading angle by the safe gap model of the willing gap.Based on two criteria of path length and path bending energy,the proposed path planning strategy and vector field histogram(VFH)are compared and analyzed in various environmental scenarios.The simulation results show that the proposed path planning method can avoid the local minimum and generate a smoother trajectory than VFH method.In addition,a classification strategy of dynamic obstacles based on power spectral density is proposed.By analyzing the power spectrum of the velocity profiles of three typical dynamic obstacles(pedestrian,handcart out of control,and handcart pushed by pedestrian).Then three typical dynamic obstacles can be identified by means of two power spectral properties(the frequency of the second largest component of the power spectrum and ratio of non-zero frequency component of the power spectrum to the total power).