Touch servo control method used for mechanical hand grabbing task

Abstract

The invention relates to a touch servo control method used for a mechanical hand grabbing task. The touch servo control method includes a first grabbing strategy planning stage and a second grabbing control stage. In the grabbing strategy planning stage, firstly, the mapping relation between the shape of a target object and the grabbing manner of a mechanical hand is established; then a collecteddata training convolution neural network is used for obtaining the touch feature of each finger in the single-time grabbing process; and then the above touch features are corrected by means of collected rotating angles near the finger joints, and corresponding controllers are designed according to the corrected touch features and the set drive motor voltage and expected touch force of all the fingers. In the grabbing control stage, in the process that the mechanical hand grabs the target object, the mechanical hand is automatically switched into the different controllers according to the real-time touch features of the fingers, and the touch servo control over the mechanical hand grabbing task is achieved. By means of the touch servo control method used for the mechanical hand grabbing task, the mechanical hand can better complete more elaborate operation and complex tasks, and the blank of the touch servo technology in the field of mechanical hand operation control application is filled up.

Description

technical field [0001] The invention relates to a tactile servo control method for a manipulator grasping task, and belongs to the technical field of tactile perception, tactile servo and operation control of a manipulator. Background technique [0002] In the manipulator operating system, tactile perception is an important sensory function when the manipulator is in direct contact with the external environment, and it is one of the key technologies for the development of manipulators. Tactile information can provide more accurate positioning information, information on the physical characteristics of objects, and information on the state of contact between the manipulator and the object. [0003] At present, in the field of manipulator control, the tactile information is mostly used to extract the characteristics of the target object such as shape, size and stiffness for classification and target recognition, that is, the tactile information is mainly used to complete the r...

AI Technical Summary

Problems solved by technology

However, it is rare to use tactile information to realize the detection function, that is, lack of tactile feedback. Therefore, it is impossible to accurately judge whether the manipulator grasps the object tightly or whether it can be damaged. Unnecessary contact caused by the error between the actual position and the command position The force and moment cannot be clearly sensed, and it is difficult to guarantee the safety and reliability of the manipulator's grasping task

[0004] To sum up, in order to make manipulator operation tasks have better intelligence, stability, and flexibility, traditional manipulator control methods based on dynamics / kinematics can no longer meet the needs.

Images