Movement Of The Robot In A Dynamic Environment In Real-time Obstacle Avoidance

One of the most important robotics research content for mobile robot obstacle avoidance automatically when,In this paper,based on monocular vision technology for automatic mobile robot obstacle avoidance of contents.Monocular vision based around mobile robot obstacle avoidance in three key aspects of a new algorithm,Traditional optimization algorithm improvements and Optimization method.Specific studies and in the following order:First,For obstacle detection and identification step the traditional method of image segmentation based on the identified problem of poor results,This paper improves the traditional obstacle recognition method based on image segmentation,It uses a color image segmentation obstacle recognition,The use of image segmentation algorithm based on HSI color space sequence to complete segmentation image segmentation.In the meantime,For ambient light conditions less likely to have an impact on the problem of image segmentation results.I use the space for the brightness before thresholding to conduct a nonlinear optimization approach,While solving the same image segmentation results affect uneven illumination problems.Traditional practices step in positioning and obstacles that exist in sectors ranging tedious computationally intensive problems,Firstly,the establishment of a target ranging model based on monocular vision,Parameter estimation performed(camera calibration)camcorder,Research on the camera pitch angle calculation problem,In this paper the camera calibration problem with a classic camera calibration Zhang Zhengyou,Plus Levenberg-Marquardt nonlinear parameter optimization method for a more accurate estimation of camera parameters.On the camera pitch angle calculation presented herein innovative geometric relationship based on positive ground after the round and oval imaging fast and accurate calculation of the pitch angle of the camera.Then,Traditional artificial potential field obstacle avoidance algorithm in the presence of links in the path planning and local minimum point and destination unreachable problem,This paper presents artificial potential field algorithm improvements.That is,when the robot into local minimum point,A repulsive force to the variable angle.It is possible to reach the goal before taking into account the robot will repeatedly fall into local minimum around the same obstacle,So also proposedsetting up a virtual local minimum point and then around the local minimum area concept.Finally,the use of genetic algorithms to optimize the value of the parameter,the parameter values in order to avoid arbitrariness.At the end,T It describes the results of test and obstacle avoidance comprehensive analysis based on monocular vision technology for mobile robot,T Test hardware platform is the smart car wheel,Combined with Open CV vision library and code package,Coding herein correlation algorithm.Not only proved by testing monocular vision technology for mobile robot obstacle avoidance feasible.And to demonstrate the effectiveness of the algorithm or design algorithm improved.