Research And Implementation Of Multi-robot Formation Control

With the development of industrial automation, computer technology, andwireless communication, research of robotics is further and further. The intelligenceof robot is improving, which promotes the development of artificial intelligence. Asan important branch of robotics, mobile robot can perceive the environment, planand learn by itself, bring a certain impact on production and life of human. As animportant research field of robot and wireless sensor networks, problem of mobilerobot formation and deployment are paid close attention by researchers.Firstly, the issue studies hardware architecture and composition methods ofintelligent mobile robot based on ARM, in the background of robot formationcontrol. With the “Creative Star” modular robot kits, we develop the softwaremodule of mobile robot’s common components. We design mobile robots withcapability of communication, which laid the foundation for mobile robot formationand graduate innovative experimental platform of mobile robot.Secondly, we collect images through webcam on the top, and process the imageinformation into the location information. Then, we transform the locationinformation into the mobile robot sensor system, and send the information to eachmobile agent robot through wireless communication module, for the purpose ofcollecting location information of obstacles and targets around the robots. Weimprove the performance of single robot system by designing the virtual sensor.Finally, we propose a control algorithm for multi-robot formation based oncluster consistency, which is consolidated on the basis of previous studies. Bymethods of MATLAB theory simulation and transplanting the algorithm into robots,we complete the distributed robot formation. We not only prove the effectiveness ofour formation control algorithm, but also analyze and modify problems during theprocess of simulation.In this paper, we develop mobile robots in embedded Linux system, andimplement motion control of mobile agent robots. We design and implement aplatform of image processing, which can perform operations of image collection,image stitching, contour extraction, and contour matching, to achieve the functionof “virtual variety of sensors”. We propose a distributed algorithm for multi-robotformation to detect and capture surrounded the targets. We complete algorithmsimulation and experiments of real objects, which confirm the effectiveness of theformation algorithm. The algorithm can ensure not only the sensors distribute around the target uniformly, but also the speed of all nodes is same with the target.