Two Algorithms Suitable For Pulse Signal Adaptive Beamforming Performance Analysis

The submarine as the main military of every state’s navy has a strong penetrationability and good concealment. Submarine activities underwater, early detection of theprobe pulse of the enemy, and the estimating target position and other parametersaccording to the pulse signal parameters, have great significance to improve thejudgment of the submarine threat, attack and defense conversion, and to improve thedefense capability and defense effects of the submarine. So adaptive beamformingmethods which are applied to acoustic positioning system should focus on resolvingproblems how to complete a stable target detection and DOA estimation in time of asingle-pulse. This article hopes to propose an algorithm which can take advantage ofthe data length as short as possible to achieve stable and optimized estimation of theweight vector.The paper firstly reviews the conventional beamforming algorithm and twoalgorithms suitable for short-pulse signal estimation: the Direct Data Domain (DDD)algorithm and time-domain analysis MVDR algorithm, and comparatively analysis oftheir performance. Theoretical analysis and simulation experiments show thatConventional beamforming method which is still widely used in underwaterpositioning systems has lower performance in target detection and azimuth resolution.DDD has higher sidelobe and a wider beam width, and constructs the noise matrix, atthe cost of reducing the number of weight vectors; its beam width widens furtherly.Time domain Analysis MVDR (TAMVDR,) completes the optimized estimationof the weight vectorin the time domain with only about2000data points, andachieved better beam performance, higher array gain, and higher resolution, whichpromotes the MVDR technology applied to short pulses estimation.To realize the stable optimal estimation of weight vectors, on the basis of DDDalgorithms and TA MVDR algorithm, a novel approach for temporal and spatial unioncomputation of adaptive weight vectors is proposed---Temporal and Spatio union MVDR adaptive beamforming(TSMVDR).The new approach combines TAMVDR’sconstruction of Time-domain signal with DDD’s spatial sliding motion. At the cost ofreducing half of weight vectors, it can eatimate the covariance matrix and weightvectors with shorter data length than the existing method optimally and stably. It isshown by simulation and experimental data on the sea processing results that,compared with the CBF, and DDD, the new approach has the following advantages:TSMVDR employs Spatial Temporal way to estimate the covariance matrix andweight vector and has better stability. The objective function of the TSMVDR issignal power, TSMVDR has better beam performance--the main lobe is narrower, theside lobe is lower, the array gain is higher.TSMVDR can estimate the adaptive weight vector optimization optimally andstably with only about300data points. The data length it needs is shorter thanTAMVDR, which makes it more suitable for short pulse adaptive beamforming.However, due to TSMVDR pays to the cost of reducing half of the weight vector, itsbeam performance than TAMVDR be reduced.