Font Size: a A A

Research On Manipulator Catching Technique Based On Moving Target State Prediction

Posted on:2018-05-31Degree:MasterType:Thesis
Country:ChinaCandidate:X ChenFull Text:PDF
GTID:2348330533469265Subject:Mechanical and electrical engineering
Abstract/Summary:PDF Full Text Request
Traditional industrial robot manufacturing process mainly relies on offline programming method.There are problems such as using in single task only,the weak ability to adapt to complex environment,etc.The use of visual feedback to guide the robot can work effectively to compensate for these shortcomings.The method in which using visual information to control robots to catch non-cooperative targets,has pretty practically-significant significance and broad application prospects.As for catching a moving target,the task involves many technical aspects and difficulties,Such as the difficulty in designing control strategy,visual feedback information more susceptible to noise and other issues.In this project,a visual servo-guided grabbing system is designed,and the problems such as the mentioned difficulties above are better resolved as well.Then draw the conclusion by simulation,which was verified by experimentation in the end.Based on the complex scenes in reality,a set of visual guidance and grasping system is realized.The overall scheme of the capturing task is designed.For the acquisition of visual feedback information,a binocular vision syste m with greater depth information and larger operating space is selected.The robot part uses a six-degree-of-freedom manipulator.The control strategy is used to estimate the target motion by means of the state observer.In order to deal with the influence of the equipment performance,noise,controller delay and other factors,a predictive control strategy is used to acquire the parameter to instruct the robot to perform a grab operation.The particle filter is selected in the filtering algorithm,and its superiority to the widely used extended Kalman filter in the strong nonlinear system is verified by MATLAB simulation.At the same time,the edge particle filter algorithm is used to further improve the accuracy of the parameters such as the target positio n,velocity or acceleration needed in the manipulator grasping operation.Finally,the simulation results are verified by experiments.
Keywords/Search Tags:state estimation, particle filter, object grasping, non-cooperative target
PDF Full Text Request
Related items