Pose Estimation And Tracking Control Of Wheeled Robots Based On Xtion

The multi-robot cooperation system can effectively accomplish the work single robot can't complete.Compared with the single robot,multi-robot system has some natural advantages,such as strong robustness and adaptability,reasonable distribution,effective coordination and high performance.Research of multi-robot system contains identification,information fusion,topology communication,cooperative work,tracking control and so on.The recognition and localization are the foundation of the multi-robot cooperation system,and the tracking control is also a hot research direction of the system.The multi-object pose estimation and trajectory tracking control are studied in this dissertation.On the basis of the achievement of multi-target pose estimation,the trajectory tracking control is realized.From the aspect of multi-agent,different robots have different perception and execution abilities.Similarly,agents in pose recognition system are divided into the observers and the observed objects,and those in trajectory tracking system are divided into the leaders and followers.Based on the Xtion camera,a multi-robot relative pose estimation method is proposed.In this method,the four degree of freedom relative pose estimation steps are designed.The method discards the disadvantages of the global sensor which depends on environment.The airborne sensor used in this method is cheap and the designed color target is robust and the method is simple and effective.First,the objects' features are more obvious by designing the target,and airborne camera avoids environmental dependence by slanting installation.Second,the technology of computer vision is used to complete recognition and localization.Moreover,the pose estimation and visual information are generated by perspective transformation.This dissertation presents a multi-robot trajectory tracking strategy based on the estimation of pose,which takes advantage of the wireless communication.This strategy allows using the communication mechanism to ensure that the follower completes the trajectory tracking task on the condition of the lack of environmental perception.The main features of the strategy are as follows.The strategy based on TCP/IP protocol uses the Client/Server Model.Besides,the navigator client combined with the odometer information collects and transmits path characteristics on real-time,and the recipient server deals with the data receiving,filtering,storage,local stabilization and global trajectory tracking control.