Font Size: a A A

The Design And Implementation Of Networked Multiple Robots Interception System

Posted on:2018-09-11Degree:MasterType:Thesis
Country:ChinaCandidate:Z J LinFull Text:PDF
GTID:2348330533466828Subject:Control theory and control engineering
Abstract/Summary:PDF Full Text Request
Moving target tracking problem of multi-robots interceptinon is one of the hottest researches of multiple robots control.Based on the prediction,planning,execution algorithm of interception,the interception problem is divided into three parts,which are the trajectory prediction base on the motion state of the moving target,the path planning based on the distiance between virtual leader and the predicted position of moving target,and the trajectory tracking controller design law of moving robots.In order to verify the interception algorithm,a networked multi-robots interception platform is designed.The main work of this paper includes as follows:Firstly,a networked multiple robot platform is built.According to the function,the platform can be divided into three parts,the detection part,the communication part and the object part,which achieve the function of robot localization,information transmission and algorithm implementator.This paper will make a detailed introduction of the hardware and the working process of these three parts.As a typical realtime system,the clock control mechanism of this planform is also introduced.Secondly,the trajectory prediction of the moving target is the premise of the interception algorithm.And the estimation of motion state is the key to the trajectory prediction.By analysising the kinematic model and the dynamical model of the moving target,this article raise a trajectory predict algorithm based on the extended Kalman filter,which estimates the motion state of the moving target and then make the multistep prediction.By simulaiton,the paper analysises the factors which influence the presicion of multistep prediction.At last,a comparision simulation shows that the extended Kalman filter excels the common Kalman filter when the moving target has varying veloccity.Thirdly,this paper puts forward an arc formation of multiple robots with virtual leader mode.In dynamic condition,an iterative path planning algorithm based on the planning period is raised.The interception is divided in three parts: tracking,convergence and capture,which respectively corresponds three formations.The stage switching is according to the relative position between the virtual leader and the predicted position of the moving target.For the collision problem and the escape behavior of the moving target,some solution is also discussed.Finally,according to the pose error modeling,a trajectory tracking error modeling is built by coordinate transformation.Then the error function is linearized and a linear quandratic controller called LQR is designed.Moreover,a trajectory tracking controller designed on Lyapunov method is introduced.Comparision simulation between these two controllers on line and circle trajectory is made and according to the simulation results and the feature of the interception algorithm,LQR controller is chosen as the trajectory tracking controller.
Keywords/Search Tags:Networked control, Extended Kalman filer, Dynamic path planning, LQR controller
PDF Full Text Request
Related items