Tracking The Spatial Multiple Targets Based On Array Signal Processing

The spatial multiple target tracking based on array antenna can be widely used in many fields sucn as radar, sonar, navigation and particularly wireless communication etc. This dissertation aims at the effective algorithms of the spatial multiple target tracking based on the array signal processing.The multiple target tracking can be divided two stages:the initial DOA estimation and the dynamic tracking. The first stage estimates the initial DOA parameters which mainly based on the MUSIC and ESPRIT algorithms. The second stage adopts the adaptive algorithms, such as the steepest descent algorithm, the LMS algorithm and the RLS algorithm etc. to track the targets. Currently the stationary estimation algorithms with high complexity for the moving tracking are employed mostly which can not satisfy the tracking process of multiple targets. From the perspectives of DOA adaptive iterative tracking and subspace tracking, we proposed some novel, fast tracking algorithms with low complexity.The main fruits of the current thesis are as follows. First, DOA estimation and tracking algorithms of spatial multiple targets based on beamspace algorithm and direction lock loop from the perspective of direct DOA adaptive iterative was presented. This method estimates the initial DOAs coarsely by the orthogonal beamspace algorithm to ensure fast lock of the DiLL, which possesses low complexity and can be used for the real-time tracking of multiple fast-moving targets.Second, based on the filter theory of sigal processing, the spatial spectrum estimation by FFT algorithm and the DOA tracking algorithm of multiple targets based on adaptive notch and gradient descent were proposed. Taking advantage of the space-time equivalence, the fast change DOAs can be tracked by only one snapshot's FFT. Utilizing the freedom degree information of the array elements, the order of the notch zero can be regulated flexibly to surpress the interference and improve the performance.Furthermore, effecient adaptive algorithms were proposed to track the time variant steer manifold of the array antenna based on the signal subspace tracking. The variable step LMS algorithm and RLS algorithm based on QR decomposition of the Givens rotation were employed respectively to accelerate the convergent rate and improve the numerical computational stability.Based on the eigenvector subspace method, the signal subspace DOA tracking algorithm which combines the Bi-iterative SVD decomposition with ESPRIT algorithm as well as the noise tracking algorithm based on the reduced Toeplitz matrix and Lagrange theorem were presented. Both algorithms possess low complexity with no requirement of spectrum peak searching.