Mechanical arm grabbing method

Abstract

The invention provides a mechanical arm grabbing method. The mechanical arm grabbing method comprises the following steps: a robot operation system builds communication mechanisms between a central controller and a visual sensor and between a mechanical arm and a manipulator; objects to be grabbed are placed in a visual range of the visual sensor; the visual sensor acquires surface information ofthe objects to be grabbed; the central controller processes data acquired by the visual sensor to obtain a grabbable coordinate point of the manipulator; according to a position relation of the visualsensor with the mechanical arm and the manipulator, the central controller can convert coordinates of the grabbable point to control instructions to send to the mechanical arm and the manipulator; and the mechanical arm and the manipulator move to an appointed position in succession to grab the objects to be grabbed according to the control instructions. Linear combination kernel functions are applied to reconstruct the surfaces of objects with uncertain shapes; and the method is low in needed data quantity, and only needs an observation camera, so that the cost is lower, and excellent practicability is achieved.

Description

technical field [0001] The invention relates to the field of robot grasping, in particular to a method for grasping a mechanical arm. Background technique [0002] With the rise of the wave of artificial intelligence, robots are playing an increasingly important role in various industries. For robots, grasping is an essential skill for robots to enter the real world, such as sorting objects in the logistics industry and assembling parts on industrial production lines. However, there are still many uncertainties in the robot's grasping task that need further research. Therefore, how to deal with uncertainty so as to improve the success rate of crawling is a problem worthy of research. Generally, the uncertainty in the grasping process mainly includes the shape uncertainty of the object to be grasped, the uncertainty of the attitude of the object to be grasped, the uncertainty of the contact point of the manipulator, and the uncertainty of the quality of the object. [0003...

AI Technical Summary

Problems solved by technology

[0004] However, the disadvantages of these two methods are obvious. Even if the first method obtains more object information by adding sensors, it is finally matched with a multi-degree-of-freedom manipulator, which greatly increases the cost and is not suitable for industrial use. production and daily...

Images