摘要
针对卫星信号受阻,无预设基础设施(定位基站、地标等)环境下多机器人间的相对定位问题,提出了一种基于单个超宽带(ultra-wideband, UWB)融合里程计的多机器人相对定位方法。该方法利用滑动窗口截取历史时刻的多组机器人间测距信息与里程计预测的机器人位姿,构建非线性最小二乘问题,实现机器人间的相对位姿估计;利用扩展卡尔曼滤波算法估计里程计协方差,并将其以加权的方式运用于非线性优化,抑制滑动窗口内里程计累积误差对定位结果的影响;最后,利用图优化算法融合里程计与非线性优化获得的相对位姿作进一步优化,抑制UWB测量误差影响,以获得稳定的相对定位结果。实验结果表明,在6 m×12 m的真实测试环境中,所提方法能够获得0.32 m的相对位置精度和4.16°的相对角度精度,相比于现有多机器人相对定位方案,该方法具有高精度、低成本、部署简单以及定位稳定的优点。
Aiming at the problem of relative positioning among multiple robots in GPS-denied, no pre-set infrastructure(positioning base stations, landmarks, etc.) environment, this paper proposed a multi robot relative positioning method based on a single ultra-wideband(UWB) fusion odometry.To obtain relative pose estimates between robots, the method constructed a nonlinear least squares problem by utilizing sliding window to intercept range information between robots and poses predicted by odometry over a period of time.In addition, the method estimated the odometry covariance by utilizing extended Kalman filter and weighted it in a nonlinear optimization to suppress the cumulative odometry error in a sliding window.Finally, it used the general graph optimization(g~2o) algorithm to further optimize the relative position and attitude obtained by integrating odometry and nonlinear optimization, which effectively suppressed the influence of UWB measurement error, so as to obtain stable relative positioning results.The experimental results show that the proposed method can achieve an average relative position accuracy of 0.32 m and relative angle accuracy of 4.16° in an indoor environment with the size of 6 m×12 m.Compared with the existing multiple robot relative positioning schemes, the proposed method has the advantages of high accuracy, low cost, simple deployment and strong robustness.
作者
邓忠元
刘冉
曹志强
肖宇峰
Deng Zhongyuan;Liu Ran;Cao Zhiqiang;Xiao Yufeng(School of Information Engineering,Southwest University of Science&Technology,Mianyang Sichuan 621000,China;Robot Technology Used for Special Environment Key Laboratory of Sichuan Province,Southwest University of Science&Technology,Mianyang Sichuan 621000,China)
出处
《计算机应用研究》
CSCD
北大核心
2023年第3期839-844,共6页
Application Research of Computers
基金
国家自然科学基金资助项目(12205245,12175187)
国家重点研发计划资助项目(2019YFB1310805)。
关键词
相对定位
非线性优化
滑动窗口
扩展卡尔曼滤波
图优化
relative localization
nonlinear optimization
sliding window
extended Kalman filter
general graph optimization