Formation Control And Experimental Research For Nonholonomic Wheeled Mobile Robots

Recently,with developments of the artificial intelligence,wheeled mobile robots(WMRs)have been widely applied to the intelligent manufacturing,the intelligent warehousing,the smart home,the intelligent logistics,the modern military operation and so on.Due to the good qualities with high efficiency,high fault tolerance and high performance,formation control of WMRs has become a hot research topic.Therefore,researches and applications of the formation control for WMRs have great values and prospects.In the actual environments,due to external disturbances,ground roughness,load changes and other problems,it is difficult to establish an accurate kinematic model for the WMRs,which brings some difficulties and challenges to the accuracy of the formation control.In addition,there often exists static and dynamic obstacles in actual environments,so the WMRs are not only required to achieve the formation control,but also realize the safe obstacle avoidance,which is one of the hard works in the current researches on the WMRs.In this paper,the WMRs are taken as the research object,and an experimental platform for the WMRs formation is also built.Then,the formation control and obstacle avoidance problems are studied in different environments.Firstly,the experimental platform for the WMRs formation is accomplished,and kinematic models of the wheeled mobile robot(WMR)and WMRs formation are also established respectively.The designment of the experimental platform consists of the platform designing idea,selections of vehicles and sensors,working principle of the platform,hardware structure of the WMR,etc.,which provides the experimental basis for subsequent researches.Moreover,according to the underactuated characteristic and the nonholonomic constraint condition,a kinematic model for a single WMR is established,and then a kinematic model for WMRs formation is also deduced based on the leader-follower formation by using the distance-angle method,which provides the theoretical basis for the following studies.Secondly,the influence of the unmodeled dynamics on the formation control is considered in the environment without obstacles,due to the inaccurate positioning of the WMR barycenter.An error-based extended state observer(EBESO)is designed by applying signals of the error feedback and the control input to observe unmodeled dynamics and complex nonlinear dynamics.Then,the followers’ dependences on the leader’s velocities are avoided and the information interactions are also reduced in the formation system.Besides,based on the dynamic compensation linearization,a distributed formation controller is proposed and the stability of the formation system is proved by the Lyapunov theory.Experimental results show that the proposed algorithm improves the accuracy and convergent rates of the formation control,greatly reduces overshoots,and enhances the stability of the formation system.Thirdly,both the formation control and obstacle avoidance problems are considered for the WMRs in the environment with static obstacles.According to the artificial potential field(APF)method,an adaptive repulsive potential field(ARPF)function is constructed for the static obstacles and neighbor robots.The strength of the repulsive potential field is properly adjusted by the adaptive field strength factor to solve the excessive repulsive force problem of the APF method and then a shorter collision-free path is generated.In addition,a distributed formation controller is designed based on the repulsive forces to achieve the autonomous real-time obstacle avoidance for the static obstacles.Experimental results demonstrate that the ARPF function regulates the repulsive forces reasonably and generates a better collision-free path for the WMRs.Finally,the formation control and obstacle avoidance problems are simultaneously considered for the WMRs in the environment with static and dynamic obstacles.Both realtime distances and relative velocities between robots and obstacles are introduced in the repulsive potential field function.Then,a compound repulsive potential field(CRPF)is proposed by superimposing the position repulsive potential field and the velocity repulsive potential filed.The compound repulsive force is used to correct the formation tracking error to plan a collision-free path in real time.Furthermore,by defining the priority of the formation control task and the formation obstacle avoidance task,a distributed formation controller is designed to realize the autonomous real-time obstacle avoidance for the static and dynamic obstacles.Experimental results show that the designed controller successfully achieves the formation control and the CRPF is suitable for both static and dynamic obstacles.