Integrated Algorithms and Multiple Antenna Techniques for Direction of Arrival (DOA) Estimation
In this dissertation, we design and develop a novel direction-of-arrival (DOA) finding system. We investigate the problems of DOA finding using canonical and crystallographic antenna array structures, develop a novel integrated algorithm consisting of combined multiple signal classification (MUSIC) algorithm, Kalman Filter and Kent Distribution to improve the accuracy and robustness of DOA estimation, design and conduct the real time testing of DOA and verify the accuracy and efficiency of the designed DOA finding system.
We first examine the ability of mitigating the aliasing and enhancing the DOA estimation of different antenna structures, including canonical and crystallographic antenna structures. Our results show that the crystallographic antenna array has a better performance of overcoming aliasing in many circumstances, improving the estimation accuracy and covering more spatial region of DOA estimation.
Then we propose a novel integrated algorithm to achieve a more robust DOA finding with higher accuracy. We show that the DOA estimation using MUSIC algorithm can be strongly influenced by the size, spacing and distributions of elements of the receiving antenna array as well as noise and mutual coupling. We propose a combined MUSIC and Kalman Filter algorithm to reduce the noise and enhance the robustness of the DOA estimation. Further more we map the DOA estimation onto the sphere and use Kent distribution to characterize the spread of DOA points on the sphere. We calculate the mean direction of Kent distribution to present the DOA vector, which further improves the accuracy of DOA finding.
At last, we design and build a multi-channel and real time automated measurement system to validate the proposed antenna structure and integrated algorithms. Our testing results indicate that the designed DOA finding system can work practically and efficiently, with higher accuracy and stronger robustness.