A method, device, equipment and storage medium for generating a grasping trajectory of a robotic arm

Abstract

The invention discloses a method, device, device, and computer-readable storage medium for generating a grabbing trajectory of a mechanical arm; the solution includes: determining the basis function weight of a motion model according to the teaching trajectory information, and obtaining a target image through a depth camera; inputting Mask R-CNN network determines the starting point and target point, and generates the grasping trajectory by running the model; it can be seen that in this scheme, the weight of the basis function in the motion model is the same as that of the teaching trajectory, so the shape of the grasping trajectory is the same as that shown The teaching trajectory is similar to achieve the purpose of trajectory learning; this scheme uses the deep learning-based detection method Mask R‑CNN to effectively identify and locate the target in the image, making the robot arm's grasp of the object more intelligent; This solution combines visual perception and dynamic motion primitive framework, endows the motion model with vision, and better interacts with the external environment in terms of robot arm teaching trajectory learning and grasping trajectory planning.

Description

technical field [0001] The present invention relates to the technical field of track generation, and more specifically, to a method, device, equipment and computer-readable storage medium for generating a grabbing track of a robotic arm. Background technique [0002] Teaching and learning is a behavior of human-computer interaction, which makes the robot reproduce the teaching action in a new environment. According to whether the teacher has direct physical contact with the robot, it can be divided into kinesthetic teaching and remote teaching. Kinesthetic teaching means that the operator directly controls the robot to complete the corresponding actions, collects the robot's own information and takes the information as the learning object. This method is not suitable for multi-degree-of-freedom robotic arms. Remote teaching generally controls robot movement through vision, wearable sensors, or other remote control tools. The traditional trajectory planning method represent...

AI Technical Summary

Problems solved by technology

The teaching and learning method of manipulator based on dynamic motion primitives can effectively learn single and multi-degree-of-freedom trajectories, but in practical applications, for example, when the manipulator grasps objects, due to the lack of visual perception of the environment As a result, the research on this technology is too simplistic and cannot realize human-computer interaction well.

However, the artificial potential field trajectory planning method will have the problem of local optimum, and sometimes it is impossible to move the manipulator to the target position, which has limitations.

Images