Robot obstacle avoidance trajectory planning method and system based on deep learning

Abstract

The invention provides a robot obstacle avoidance trajectory planning method and system based on deep learning. The robot obstacle avoidance trajectory planning method based on the deep learning comprises the following steps: adding a camera to a simulation environment, taking images from multiple angles and simultaneously inputting the images into a convolutional neural network; obtaining information of a robot arm updated angle according to the input information, and calling a simulation software to update through an interface to obtain a posture; and performing convolutional neural networktraining by means of the deep learning, transferring an obtained characteristic pattern to a one-dimensional vector after convolution operation is performed on the input images, inputting the one-dimensional vector into a subsequent fully connected layer to obtain a q value corresponding to each action, selecting the action with the largest q value and updating the posture, and sending the updatedposture to the simulation environment to obtain a new image input, and executing circularly until the target point is reached. The invention can realize the autonomous obstacle avoidance of an industrial robot and improve the industrial automation production capacity.

Description

technical field [0001] The present invention relates to the field of industrial automation, in particular to a method and system for robot obstacle avoidance trajectory planning based on deep learning. Background technique [0002] In the modern industrial production environment, more and more robots are introduced into the assembly line to participate in the production and manufacturing work. The more common application scenarios include mechanical processing and manufacturing, welding, assembly, spraying, packaging, etc. By replacing manual labor with robots, it can effectively improve work efficiency, improve production yield, and reduce labor costs. [0003] However, common robots have relatively high requirements on the working environment, and it is necessary to ensure that there are no obstacles in the working range of the robot to block the movement of the robot. The existing common treatment method is to set up protective fences in the robot working area, but this ...

AI Technical Summary

Problems solved by technology

[0003] However, common robots have relatively high requirements on the working environment, and it is necessary to ensure that there are no obstacles to block the movement of the robot within the working range of the robot.

The existing common treatment method is to set up protective fences in the robot working area, but this method reduces the space utilization rate of factory production to a certain extent, and is not applicable to temporary obstacles in complex production environments

Images