基于多模态信息融合的水下移动目标定位

Underwater moving target localization based on multi-modal information fusion

移动目标的位置和速度估计是水下传感器网络应用的重要内容。针对到达时间差(TDOA)和到达角(AOA)融合只能进行目标位置估计的局限性以及位置估计中遇到的非线性问题,提出了一种多模态信息融合的三步定位方法。该方法在TDOA/AOA基础上,通过融合到达频差(FDOA)来同时估计运动目标的位置和速度。前两步采用两步加权最小二乘法来估计目标的粗略位置和速度。为了更好地求解非线性定位问题,第三步将定位问题表述为最大似然函数,利用鲸鱼优化算法求解。通过第二步的解,构造鲸鱼优化算法的初始种群,以测量误差方差倒数为算法适应度函数的权重。仿真结果表明,该方法与TDOA定位方法、TDOA/AOA混合定位方法和两步加权最小二乘(TSWLS)算法相比,在位置和速度估计精度和偏差方面均优于上述方法。

Position and velocity estimation of moving target are significant in the application of Under Water Sensor Networks(UWSNs).Considering the limitation that the Time Difference of Arrival(TDOA)and Angle of Arrival(AOA)fusion can only estimate position and the nonlinear problem encountered in position estimation,a three-step multi-modal information fusion method was proposed.On the basis of TDOA/AOA,the position and velocity of the moving target are estimated simultaneously in this method by fusing the Frequency Difference of Arrival(FDOA).In the first two steps,the target rough position and velocity are estimated by two-step weighted least squares method.In order to solve the nonlinear localization problem better,the localization problem is expressed as the maximum likelihood function and solved by Whale Optimization Algorithm(WOA)in the third step.The initial population of the WOA is constructed by the solution of the second step,and the reciprocal of measurement error variance is used as the weight of fitness function.Simulation results have demonstrated that the proposed method compared with TDOA localization method,TDOA/AOA hybrid localization method and Two-Step Weighted Least Squares(TSWLS)algorithm,the proposed method outperforms these methods in terms of position and velocity estimation accuracy and bias.