摘要
传统的室内机器人导航系统都是依靠开发者单独设计的控制系统及软件框架,在功能模块上不同产品之间的通用性极其不好。针对此问题,基于机器人操作系统提出了一种室内机器人导航方法,采用支持向量机实现摄像头定标和坐标转换,并运用卡尔曼滤波器进行定位导航。为了实现对多目标的检测和跟踪,提出优化增长型神经网络,使之聚类速度得到提升。最后,在Rovio机器人平台上实现导航算法。实验所得数据和结果,论证研究方法的有效性和算法的优化思路。
The control system designed separately is used in the traditional indoor robot navigation system,but its versatility is poor.For this,an indoor robot navigation method based on ROS(robot operating system)is proposed.SVM(support vector machine)is used in the program to realize camera calibration and coordinate transformations,and Kalman filter is used to complete the navigation and location.To achieve multi-target detection and tracking,this paper proposes optimizing the growth neural gas(GNG)which is used to improve the clustering speed.The experimental results show that the proposed method for Rovio robot platform is effective and practical.
作者
刘海平
战强
LIU Haiping;ZHAN Qiang(Sino-French Engineer School,Beijing University of Aeronautics and Astronautics,Beijing 100191,China;School of Mechanical Engineering and Automation, Beijing University of Aeronautics and Astronautics,Beijing 100191,China)
出处
《机械制造与自动化》
2018年第6期158-161,共4页
Machine Building & Automation
