Optimal sensor placement for underwater threat detection

The optimization framework for optimal sensor placement for underwater threat detection has been developed. It considers single‐period and multiperiod detection models, each of which includes two components: detection algorithm and optimization problem for sensor placement. The detection algorithms for single‐period and multiperiod models are based on likelihood ratio and sequential testing, respectively. For the both models, the optimization problems use the principle of superadditive coverage, which is closely related to energy‐based and information‐based approaches. An algorithm for quasi‐regular sensor placement approximating solutions to the optimization problems has been developed based on corresponding continuous relaxations and a criterion for its applicability has been obtained. Numerical experiments have demonstrated that the algorithm consistently outperforms existing optimization techniques for optimal sensor placement.© 2008 Wiley Periodicals, Inc. Naval Research Logistics, 2008     

光纤陀螺随机漂移分析与建模方法

随机漂移是影响光纤陀螺精度的重要因素。根据光纤陀螺静态漂移测试信号,利用Allan方差法对光纤陀螺随机漂移误差进行了分离和辨识,并分析了系数拟合过程中最小二乘法的病态矩阵问题。最后探讨了利用AIC准则和长自回归法建立光纤陀螺随机漂移ARMA模型的方法,为进一步对光纤陀螺输出信号进行滤波处理奠定了基础。     

多传感器任务分派的快速启发式规划新算法

在多目标多传感器管理中经常采用的线性规划算法中,随着传感器个数和目标个数的增加,计算量会爆炸式增长,使得跟踪系统不能实时计算,为此,根据传感器管理中线性规划的特点,提出了一种快速启发式算法,考虑组合中的传感器个数将组合的分配效用转化为权重,递推分配权重最大的组合,逐步减小组合和目标的个数.证明了权重最大的组合分派能实现组合中的传感器的最大效用.仿真结果表明该算法在与采用线性规划方法的跟踪精度相当的情况下,能有效地减小计算量.