Research On Autonomous Mobile Robot Target Tracking System

Robot is comprehensive research project that contains a number of technology, autonomous mobile robot is a high-tech branch of the robotic., is a concentrated form of artificial intelligence and mechatronics. As the autonomous robot's intelligence and efficiency, as well as the strong adaptability to harsh environments, It has been widely used in the military, scientific research and industrial production and other fields, therefore, the autonomous mobile robot research is particularly important, in this paper, DM-B200-type robot produced by Shenzhen Ou Peng is used as research object, we focus on the image segmentation and targeting algorithms of visual system, and do some introduction about how to make route planning based on visual information., The following is main content:First, start with external structure of the autonomous mobile robot, made an overview of some hardware's functions and features. This paper divide the whole system of autonomous mobile robot into four subsystems, and separately does a brief overview to those subsystems, then according to the contrast of some typical surface mirror imaging principle and characteristics, finally we choose hyperboloid mirror which has small deformation, high resolution and single point view feature, and define angle relations between target points and target pixel.Second, In order to make follow-up algorithm suitable for color image analysis and diagnosis, through the contrast of several color space characteristics, this paper finally selected HSV color space model which is similar with characteristics of the human eye, then we have focused on the segmentation algorithm of seed point growth which is based on region segmentation algorithm, and carried out a simulation experiment for it, results show that seed point growth segmentation algorithm can effectively determine the region of the target object. and we can define offset angle among target object and forward camera's main axis according to visual system imaging principle, immediately make robot be toward target when defining target location, and locate target according to forward camera location algorithm, defining target coordinates and direction of travel from robot. Finally, this paper introduces the basic concepts and theory of genetic algorithms, and does a detailed discussion about what is the fitness function as well as the genetic operator, then designing path planning strategy based on genetic algorithm with specific requirements, and carried out simulation, verified adaptability and effectiveness of genetic algorithms in different environment.