Design And Implementation Of Formationcontrol Of Self-balancing Vehicles Based On Monocular Vision

As a typical research object of wheeled robots,the self-balancing vehicle has been concerned by researchers in the field of control engineering all the time.Its instability is not only the criterion for checking the performance of the controller,but also the difficulty in designing the controller.Compared with stable vehicles,self-balancing vehicles have the characteristics of greater flexibility,smaller footprint,and lower power consumption.At the same time,with the increasing of complexity in people's lives and production,more and more tasks have become more complex and arduous,which requires multi robots to cooperate to complete the same complex task.The research of multi-robot formation is the basis for the study of their mutual cooperation and accomplishment.Compared with the stable vehicle formation system,the self-balancing vehicle formation system has increased the complexity of its formation system because of its own instability,so there is currently little research on the development of self-balancing vehicle formations.However,the self-balancing vehicle fleet system not only has the advantages of a single self-balancing vehicle,but also has stronger features such as robustness and flexibility.Therefore,in this paper,the study of multi-robot formation control is the introduction point,and the multi-robot formation system composed of two self-balancing vehicles is taken as the research object.The monocular vision method is adopted to realize the selfbalancing vehicle formation march.On the basis of summarizing the development status of self-balancing vehicles at home and abroad and the current formation control methods of multiple stable smart vehicles,the modeling work of a single self-balancing vehicle and self-balancing vehicle fleet system is completed in this paper.With the Lagrange equation,a single vehicle's motion modeling is carried out to analyze the motion of a two-wheeled self-balancing vehicle and establish its mathematical model.The mathematical model is linearized and decoupled to overcome the difficulties in the study of the nonlinear,strong coupling and variable parameters of the mathematical model of the two-wheeled self-balancing vehicle.After the simulation,the mathematical model is proved to be in line with the actual running state of the two-wheeled self-balancing vehicle,and the mathematical model is established to provide the theoretical basis for the controller design of the self-balancing vehicle.In the other part,through the analysis of the relative distance and relative angle of two vehicles,the differential equation is established,and the mathematical modeling of the formation system composed of two-wheeled self-balancing vehicles is completed.At the same time,because the self-balancing vehicle is in the dynamic balance movement process,the information generated by the camera is affected by the fluctuation.Therefore,the wave model is established and analyzed,which provides a certain basis for the design of the marker and the installation of the camera.In order to obtain and deal with the motion tracking information of the vehicle formation system,this paper adopts the method of monocular vision to achieve two selfbalancing vehicles to maintain a certain distance and relative Angle formation.It is necessary to collect images and deal with them.The processing method includes the transformation,filtering,segmentation,image thresholding of grayscale images,etc.,and information extraction of the marker in the processed images.By using the principle of small hole imaging,the monocular vision fast ranging is realized,and the distance and relative angle of the two vehicles are calculated to track the information needed by the vehicle.In the control of the vehicle formation system,this paper analyzes and tests the coupling of the self-balancing vehicle model and the formation model.The experiment shows that the coupling between them is rather weak,which can be regarded as two independent subsystems.For the balance control of the self-balancing vehicle itself,this paper adopts the PID control and the method of state feedback control to simulate them respectively.The results show that both control methods can achieve a good balance control effect.The mathematical model of formation system is a nonlinear system.In this paper,the input-output feedback linearization method is used to design the controller.The results of Simulink simulation show that the controller is good for the self-balancing vehicle formation and travel control,which proves that the scheme adopted in this paper is feasible.At the same time,the hardware platform of the self-balancing vehicle formation system is established,the overall experimental design scheme and the hardware environment of the system are set up,and the software program is written and debugged.The linear and triangular formation teams of the self-balancing vehicle based on monocular vision are achieved.