| A common sequela of stroke is hemiplegia.The patient loses complete control of the limbs due to damage to the motor control function of the brain nerves.The connection between the cranial nerves and the limbs can be reestablished by repetitive exercise stimulation of the affected limbs.Using rehabilitation robot for rehabilitation can not only make full use of the robot’s advantages in repeatability and motion quality,but also record the position and results of the rehabilitation process.The upper limb rehabilitation robots use position,force sensors and other multi-sensor fusion methods to ensure safety which is not safe when the system fails.To solve this problem,an upper limb rehabilitation robot based on compliant joints is proposed,and the inherent compliance of the robot joints can be used to implement force and safety control.A compliant joint based on dual-torsion-spring is proposed.For the proposed torsion spring based on involute curve,the stiffness model of torsion spring is established by using pseudo-rigid-body model,and the relationship between the design parameters and stiffness is obtained.The joint uses two identical torsion springs during two-way transmission,which can not only avoid the inconsistent bidirectional stiffness of the torsion spring,but also eliminate the rebound process when the torsion spring is reversing which can cause jitter and motion error.The joint can utilize inherent flexibility to absorb impact energy,ensuring the fail-safety of the upper limb rehabilitation robot.An upper limb rehabilitation robot based on compliant joint is proposed.The robot adopts planar hinge link structure design which can realize a large range of motion and simulate daily life movements.The structure size of the robot is calculated through the feature points of the trajectory.The structure size of the robot is optimized by nonlinear optimization method,which improves the comfort of rehabilitation.Calculations from multiple aspects such as energy and force prove the safety of the proposed upper limb rehabilitation robot.The dynamics model of the upper limb rehabilitation robot was established,and the relationship between the internal force of each hinge and the force of the robot end and the active joint are obtained.The response of the joint is obtained through the dynamic model.In order to realize the active rehabilitation function,the admittance control model based on compliant joints is established,and the parameters of the admittance controller are selected by simulation.The collision safety control is realized without using force sensors which can absorbs collision energy and protects the patient during a collision.An experimental platform is built,and joint stiffness,admittance control and collision safety control experiments were carried out.The experimental results show that the established torsion spring stiffness model can effectively predict the actual stiffness of torsion spring.It is feasible and effective to regard flexible joints and dual encoders as a low-cost torque sensor.And can realize measurement of torque and perception of collision.The upper limb rehabilitation robot based on flexible joints has certain application value and can protect the patient from injury. |