Study Of The Control Strategy Of The Upper Limb Rehabilitation Robotic Arm Based On Motion Intention Recognition

As many countries step into the aging society rapidly, combined with the highincidence of the cerebrovascular accident or stroke in the elderly population, more andmore elders cannot enjoy the healthy life due to this disease. A potential consequence ofthe stroke is the impairment of the motor function of the upper limb, and then thedifficulty to perform activities of daily living. Meanwhile, there are also many elders whosuffer from deficits of motor function of the limbs, which are usually caused by lack ofmuscle strength. Many researches on rehabilitation medicine shows that a therapeutictreatment consisting of highly repetitive rehabilitation training of the impaired limb,combined with the guidance of user motion intention, is one of the most effectiveapproaches to partly restore the motion abilities. In this thesis, a cooperative controlstrategy of the upper limb rehabilitation robotic exoskeleton guided by the upper limbmotion intention was proposed to integrate the perception of the human and the intelligentcooperative assistance ability of the robots to improve the effectiveness of rehabilitationtraining.In this thesis, a3-DOF upper limb rehabilitation robotic exoskeleton was designedand a detailed analysis of the kinematics of the exoskeleton was provided. Meanwhile, anew human-robot interface for upper limb robotic exoskeleton based on the FSRs wasproposed to infer the upper limb motion intention by applying the interaction forcesbetween the upper limb and the exoskeleton. The design of the double-shell structure notonly ensures patient comfort and safety wear, while also ensuring the accuracy of thepressure signal measurement. On this basis, we proposed the hardware and softwaredesign of the control system of the rehabilitation exoskeleton. In the hardware part, thecontrol and drive platform was designed, and a new method for accurate measurementof the force signals based on the FSR sensors was proposed. In the softwarepart, the workflow of the control system、communication modules and other modules wasintroduced.On this basis, a method of quantitatively describe of the upper limb motion intention(IRD) was proposed. The state model and the observation model of IRD were built. Basedon these two models, the variable dimension Kalman filter was used to online infer theIRD. Guided by the estimated IRD, an admittance control strategy and a PID controlstrategy is assumed to control the motions of the three DC motors in the joints of robotic exoskeleton. a hypothesis testing experiment、 a basic movement experiment and acomparison experiment were designed to confirm the effectiveness of the proposedapproaches.