Research On Wearable Parallel Assistive Robot For Hip Joint Power Assist

Among the assistive robots, with supplementingthe function of human limb or replacing it completely, wearable assistive robot has become an attractive method of power assisting, since it can be used for a wide range of application. Wearable assistive robot, which is person-oriented robots, can be worn by human operator through operating alongside human limbs.With the assistance provided by the wearable assistive robot, the operator feels scaled-down loads as part of the load is being assisted by the wearable assistive robot.In our research, we focused on human hip joint power assist, which is also part of the work of the Research of Intelligent Robot for the Old and Disabled Power Assist (supported by the National High Technology Research and Development Program863). The aim is to design and manufacture a human hip joint3-DOF power assist robot.Based on the analytical architecture of hip joint and combined with robotics, bionics and human engineering, we have designed a robot for human hip joint3-DOF power assist. Different than other hip joint assistive robots, the proposed robot is a3UPS parallel mechanism. This robot can keep hip joint’s own3-DOF, and then move with human thigh avoiding interference. In order to obtain the parameters of hip joint, we have designed a stereophotogrammetry localization system. This system can measure the position of hip joint center. Then we can build the kinematical model for human thigh and assistive mechanism.For the human limb and assistive robot, we consider them as a whole parallel mechanism. Through the serial robotics, we build the kinematics model for human thigh and UPS chain, and then we build the inverse Jacobian for the parallel mechanism. In order to build the direct Jacobian for parallel mechanism, we proposed a method based on differing the geometry constraint equations. After differentiation, we can obtain the velocity relation between the passive joints and active joints, and then we can use this relation to build the direct Jacobian for parallel mechanism.For the mechanical singularity, we proposed a quick algorithm based on stochastic approximation (SA) which is used to seek mechanical singularity. Different from conventional methods, the proposed SA algorithm is quicker and much more effective. Integrated with mechanical kinematic Jacobian matrix, the SA algorithm is based on the strict and integrated illation process. During the illation process, the restrictive qualifications for this algorithm are also outlined. Through the applications on several mechanisms, it is found that the algorithm can be used for those mechanisms which satisfy the restrictive qualifications to seek out singular points near the initial working points in workspace quickly and effectively.The purpose of the design optimization for assistive mechanism is ensuring that assistive mechanism is able to satisfy assistive feasibility and realizes better assistive effect in the whole expected workspace. For this purpose, based on manipulability comparison between assisted limb and slave-active-assistive mechanism, Manipulability Inclusive Principle (MIP) is proposed as an effective and simple method to consider assistive feasibility and assistive effect in terms of kinematics. Then MIP evaluation criterions are proposed to evaluate assistive feasibility and assistive effect in mathematical ways. Concretely, weak inclusive judgment algorithm and strong inclusive judgment algorithm are proposed for judging inclusive cases to evaluate assistive feasibility, and evaluation criterion proposed for evaluating assistive effect considers in terms of assistive efficiency, assistive ability and assistive isotropy. The applications show design optimization based on MIP can effectively optimize the parallel assistive mechanism to realize better assistance.The control stratgy also has influence on assistive result. The basic principle for the control of assist robot rests on the notion that the assistive robot needs to shadow the wearer’s movements quickly, and without delay. This requires an effective control to minimal the response between human and the exoskeleton quickly, where in our research pseudo-compliance control is adopted. Through the simulation, experiment on artificial limb and experiment on human limb, we have proved that the control stragety is available for hip joint assistive robot. Additionally, the results also show that the prposed method for building direct parallel Jacobian is available.