Deep neural network based flexible arm intelligent sensing and control method and system

Abstract

The invention provides a deep neural network based flexible arm intelligent sensing and control method and system. The method comprises the following steps of training a deep neural network by collecting training samples to obtain a training network; and using the training network to perform flexible arm posture learning on collected target pictures, and controlling a flexible arm according to anobtained flexible arm control result. Aiming at complex environmental information, a target positioning deep neural network is designed. According to a complex kinematics model of the flexible arm, akinematic model is designed to calculate the neural network. By adopting the deep neural network based flexible arm intelligent sensing and control method and system, data is collected to perform training and learning on the network to obtain an effective calculation model for a control task of the flexible arm, the precision and the control capability of a space maneuvering platform to non-cooperative target tracking and recognition can be improved, the real-time sensing capability of a monitoring region is achieved, and the deep neural network based flexible arm intelligent sensing and control method and system can be widely applied to the fields of space rail service and unmanned monitoring systems.

Description

technical field [0001] The invention relates to the technical field of computer vision and deep learning, in particular to a method and system for intelligent perception and control of a flexible arm based on a deep neural network. Background technique [0002] In order to assist or replace humans to complete various tasks, intelligent robots have received extensive attention and development in recent years. The significance of using robots is particularly important in high-risk activities such as disaster relief and space on-orbit services. Because in such tasks, the use of robots can effectively reduce the danger to the operator. There are many types of intelligent robots, and robotic arms are widely used in space on-orbit service tasks. For example, the Canadian Space Station Manipulator System (SSRMS), the Japan Aerospace Exploration Agency (JAXA) Engineering Test Satellite ETS-VII, and the U.S. Defense Advanced Research Projects Agency (DARPA) Orbital Express project,...

AI Technical Summary

Problems solved by technology

At the same time, the configuration of the flexible arm brings difficulties to the control, and the traditional effective control methods for the rigid manipulator will not be applicable.

At present, the control methods of most of the research work are relatively preliminary. Generally, open-loop methods are used to control the movement of the flexible arm by artificially setting motion commands. However, such control methods are very limited in application, because in disaster rescue and In the space on-orbit service, the autonomy of the flexible arm is high, and it needs to find the target independently and operate the target so that it can automatically complete various functional requirements

Images