| Exoskeleton robot, as one of the representative applications of the robotic technology, isa powered mobile machine worn by a person to assist and protect the wearer. Now manyscientists in china and abroad have focused on this area to design a suitable exoskeleton forsoldier, aged or disable people. Meanwhile, the aging population and increasing physicallydisabled patients bring a significant social problem and have a large demand on theintelligent rehabilitation exoskeleton. However, the successful rehabilitation exoskeletonsystem in clinic is still rare.There are many challenges in the application of exoskeleton, suchas wearable, high reliability, intelligent control. In order to control the exoskeleton by humanintention, it is worth to study the mechanism of human-exoskeleton interaction, especially theskeletal muscle contraction principle to reveal the relationship between EMG signals andmuscle force.The motivation of this thesis is to design and control a lower extremity exoskeleton forpatients with lower limb problems. A human-machine interface (HMI) based on EMG signalsand interaction force between human and exoskeleton is designed to control the exoskeletonnaturally. This thesis studies the force interaction mechanism between human body andexoskeleton and analyzes the force generation principle of skeletal muscle contraction. Adynamical skeletal muscle model based on the collective behavior of myosin II motors isestablished to reflex the relationship between EMG signals and muscle force. This model wasexperimentally verified as correct and valuable. A control strategy of exoskeleton isformulated in the different state of lower extremity patients' rehabilitation process. The maincontants and achievement are as follows:Firstly, this thesis focuses on the coupling mechanism of multi-force interactions in themyosin II during its interaction with actin filament in skeletal muscle. These forces includethe electrostatic force, the van der Waals force and Casimir force in solvent with thermalfluctuations. Based on the Hamaker approach, van der Waals and Casimir forces are calculated between myosin and actin. A Monte Carlo method is developed to simulate thedynamic activity of the molecular motor. It's shown that because of the retardation effect, thevan der Waals force falls into the Casimir force when the distance between the surfaces islarger than3nm. When the distance is smaller than3nm, the electrostatic force and the vander Waals force increase until the myosin becomes attached to the actin. And the electrostaticforce dominates the attractive interactions in the binding process.Secondly, this thesis studies the force generation mechanism of skeletal muscle based onthe collective behavior of myosin motors. The non-equilibrium statistical method is used tostudy the collective characteristics of myosin motors in a sarcomere during its contraction.Combined with Fokker-Planck equation, an active force model for the sarcomere in skeletalmuscle is established. This model has been solved with a numerical algorithm based onexperimental chemical transition rates. The influence of ATP concentration and load oncontraction velocity and maximum active force is discussed respectively. It is shown thatcontraction velocity and maximum isometric active force increase as a result of theincreasing ATP concentration and toward constant when the ATP concentration reachesequilibrium saturation. Contraction velocity reduces gradually as the load force increases.Furthermore, a dynamic skeletal muscle force model is developed to describe its activationkinetics and contraction dynamics based on the collective behavior of myosin motors.According to the structure of sarcomeres arranged in series and in parallel, the mechanicalproperties of skeletal muscle are studied. This model reveals the relations between actionpotential and muscle characteristics. It is shown that calcium concentration in sarcoplasmicincreases linearly with the increasing stimulation frequency and gradually reaches saturation.Active force and contraction velocity follow the trend of calcium concentration and reach apeak value in maximum stimulation frequency.Thirdly, because of the larger angle, more freedom degree and high torque of humanjoint movement, a multi-functional exoskeleton robot and a parallel articulated exoskeletonankle are designed on the bionic princple. The rotation range of exoskeleton knee joint is0~110degree, hip joint is-25~55degree, the parallel ankle can achieve thedorsiflexion/plantarflexion, internal/external rotations in two degrees,which can meet theactual demand of human joint movement. The user in155cm~190cm height can use thisexoskeleton by lifting the height of body weight system. The gravity centre of human body can be controlled to meet the fluctuation characteristics of human gravity in the gait training.Patient lies in the centre of exoskeleton to guarantee the stability and reliability. Themaximum load of this system reaches100kg. Meanwhile, a HMI is designed to control theexoskeleton robot, integrated in the control system of exoskeleton robot, which consists ofthe EMG signal sensor, force sensor, data acquisition and process hardware card. For betterunderstanding the dynamical characteristics of exoskeleton, this paper analyzes the dynamicsof exoskeleton system and forward and inverse dynamics of human knee joint with abiomechanical model that composed of eight muscles and bones, the relationship betweenEMG signal and muscle force is established on the muscle force model.Finally, experimental research has been done for the validation of skeletal muscle modelwith HMI. EMG signals of each muscle, knee angle, interaction force between human andexoskeleton are collected to analyze the force interaction mechanism. EMG signals are usedto represent the activation level of muscle contraction. The nonlinear relationship betweenmuscle force and characteristic frequency of EMG signals is analyzed. Active torque basedon the muscle model with the forward dynamics and the joint torque with reverse dynamicsis calculated, the good agreement between forward torque and inverse torque verifies thevalidity of the muscle model. A control strategy for exoskeleton robot system is proposed.Exoskeleton can help patients with the setting gait and angle on the passive strategy; on theactive strategy, the required joint torque of limb movement is predicted based on the skeletalmuscle model using EMG singals, the exoskeleton is controlled by human intentionsuccessfully.
|
PDF Full Text Request