基于FPGA的改进型最小均方自适应时延估计器

Improved LMS adaptive time delay estimator based on FPGA

提高时延估计器的运算速度和时延分辨力是实现垂直线阵水下被动声目标快速、精确定位的关键。将最小均方自适应时延估计算法中原有的按"滤波-误差更新-权系数更新"严格依次递进的顺序迭代运算,优化为只包含误差更新和权系数更新操作的全并行乘/加运算。改进后的自适应时延估计器结构紧凑,并未改变其误差收敛特性。消声水池缩比试验表明,基于现场可编程逻辑门阵列(field programmable gate array,FPGA)实现的改进型最小均方自适应时延估计器能够在不需要流水线或时序控制单元的前提下提高数据处理速度,且具有一定的抗噪声干扰性能。

Improving the operational rate and time resolution of time delay estimator is critical for vertical liner array to realize fast and precise passive localization of underwater sound targets,the conventional adaptive least mean square time delay estimation（LMSTDE）algorithm,which sequences the iterative operation of filter -updating error-updating coefficient,is optimized as the parallel multiplicative and additive operations of updating error and updating coefficients.The optimized least mean square（LMS）time delay estimator has compact structure and does not change its error convergence characteristic.Similar examination in anechoic tank indicates that the optimized LMS time delay estimator based on field programmable gate array（FPGA） improves the operational rate without pipeline or schedule control unit and has definite anti noise jamming performance.