| With the progress of science and technology,intelligent robot has been widely used in more and more fields,such as home service,public service,industry,medical care and so on.Most of these intelligent robots are terrestrial mobile robots,such as food delivery robot in the smart restaurants and smart hotels,airport security robot and so on.These robots often use the target tracking function,so it is a very practical research to realize the target tracking of the terrestrial mobile robot.Compared with the target tracking of fixed camera,the target tracking of mobile camera is easier to lose the target.In order to achieve stable target tracking of mobile camera,this paper will study and experiment on the turtlebot terrestrial mobile platform.The main contents of the study are as follows:1.In this paper,the visual target tracking algorithm is tested and studied.In this paper,Meanshift,TLD(tracking learning detection)and KCF(Kernelized Correlation Filters)algorithm are compared.KCF algorithm meets the real-time and accuracy of target tracking,but it can’t find the target after losing the target.Therefore,this paper aims at this shortcoming,by combining the initial classifier and the current classifier,the algorithm can retain the initial features of the target,recover the target after losing the target,and finally reduce its average center error to 68.8% of the original KCF algorithm in the four experiments in this paper.2.In this paper,the motion control algorithm is tested and studied.In this paper,PID(Proportional Integral Differential)control algorithm is used for motion control,and the control on two degrees of freedom is changed according to the specific situation of the terrestrial mobile platform.PD control is used for the control of angular velocity,and PI control with upper limit is used for the control of linear velocity.3.In this paper,the target tracking flow of the terrestrial mobile platform is proposed.The KCF algorithm and PID control algorithm are combined to realize the real-time target tracking of the terrestrial mobile platform.4.In this paper,the tracking experiments of static objects and moving objects are carried out on the terrestrial mobile platform.The color images and the control response curves of the system are recorded.The experiments show that the terrestrial mobile platform can achieve real-time target tracking. |
