Grasp and clamp is one of the most important operation forms of a robot. The research on how to achieve fast, accurate and stable grasping has great value to industrial robot, service robot, planetary exploration robot and newly intelligence robot. Base on that, this paper mainly conducts following study. (1) Workpiece deformation must be controlled in the manufacturing process and other engineering application. Fixture configuration (position), clamping force and temperature are main aspects that influence the degree and distribution of Workpiece deformation. This paper takes large optical glass as an example, develop a new multiple kernel learning method to discuss the optimal fixture design. The proposed method uses two layers regressions to group and order the data sources by the weights of the kernels and the factors of the layers. Since that, the influences of the clamps and the temperature can be evaluated by grouping them into different layers. Then, based on the proposed model, the optimal magnitude and positions of clamping forces can be obtained. The experiments show is effective for the optical element clamping optimization analysis. (2) This paper takes three-finger gripper and equilateral triangle for instance, analyses their fault tolerant space on the grasping of plane polygon and find the corresponding fault-tolerant area. That is, explore the original fault-tolerant constraint set of three-finger grippers when they are grasping plane polygon. So, every original grasping status could achieve corresponding stable status at last and realize stable grasping. Besides that, this paper puts forward a design plan of flexible visual manipulator and outlines application process. (3) In order to solve the problem of redundancy and coordination of human’s upper limb, this paper has done a series of work on the dynamics and kinematics simulation of the human’s upper limb and then gets stimulus signals of upper limb’s neuro-muscular and furtherly, analyses human’s upper limb coordinated motion control. Besides that, this paper makes a study of mapping relationship between neuromuscular stimulus signals and upper limb’s pose. Mapping relationship of continuous bilinear interpolation is also put forward and research hypothesis is verified through simulation experiment. This paper provides the basis of human’s upper limb coordinated motion control and achieve robot’s coordinated, flexible, fast and smooth grab operation.
修改评论