Computer Engineering and Applications ›› 2021, Vol. 57 ›› Issue (23): 129-136.DOI: 10.3778/j.issn.1002-8331.2104-0161

• Network, Communication and Security • Previous Articles     Next Articles

Alternative Nonnegative Constrained Framework-Based Cooperative Localization Algorithm in Ocean Sensor Networks

CHENG Shuai, WU Huafeng, MEI Xiaojun   

  1. Merchant Marine College, Shanghai Maritime University, Shanghai 201306, China
  • Online:2021-12-01 Published:2021-12-02

交替非负约束框架的海洋传感网协同定位

程帅,吴华锋,梅骁峻   

  1. 上海海事大学 商船学院,上海 201306

Abstract:

To improve the localization accuracy and efficiency when using non-cooperative algorithm to sequentially locate multiple surface targets in Ocean Sensor Networks(OSNs), an Active Set Method-based Re-Estimation Cooperative Localization(ASM-RECL) algorithm is presented. The original non-convex and non-linear localization problem is transformed into an Alternative Nonnegative Constrained Least Squares(ANCLS) framework, in which an Active Set Method (ASM) is utilized to figure out the solution via both the inner loop and the outer loop. However, the solution may drop to the local optimum. In this case, the first-order Taylor series linear expansion based on the solution obtained by ASM is used to re-construct the optimization problem to minimize the localization error and improve the quality of the solution. Besides, the study also derives the Cooperative Localization-based Cramer-Rao Low Bound (CRLB-CL) to be the calibration to evaluate the proposed approach. Simulations demonstrate the effectiveness of ASM-RECL compared with other state-of-the-art approached in various scenarios.

Key words: Ocean Sensor Networks(OSNs), cooperative localization, Alternative Nonnegative Constrained Least Squares (ANCLS) framework, least squares, Cramer-Rao Low Bound (CRLB)

摘要:

针对海洋传感网(Ocean Sensor Networks,OSNs)中采用非协同算法单一循环地对多个水面目标节点依次定位导致的定位效率低、定位精度差等问题,提出一种基于有效集的再优化协同定位(Active Set Method based Re-Estimation Cooperative Localization,ASM-RECL)算法。研究将原定位的非凸非线性问题转化为基于交替非负约束最小二乘(Alternative Nonnegative Constrained Least Squares,ANCLS)的优化问题,利用有效集法(Active Set Method,ASM)通过内外循环寻求优化问题的可行解。但ASM算法易陷入局部最优,为进一步提升解的质量,改进定位精度,基于ASM得出的可行解,应用一阶泰勒级数线性展开再次构造优化方程,最小化定位误差。此外,研究还推导得到基于协同定位的克劳美罗下界(Cooperative Localization-based Cramer-Rao Low Bound,CRLB-CL),以此作为评价标准评估提出的定位算法的有效性。仿真实验表明,在不同的条件下,ASM-RECL的定位精度较高于其他算法。

关键词: 海洋传感网(OSNs), 协同定位, 交替非负约束最小二乘框架, 最小二乘法, 克拉美罗下界(CRLB)