Implementation Of Interferometer Direction Finding Algorithm Based On Parallel Baseline

With the development of radio direction detecting technology, interferometer is widely used in many military and civil fields. According to the difference of the phase for expected signal which were received by different antenna elements, interferometer system determines those phase differences to estimate the direction of arrival. In this dissertation, the algorithms using in interferometer system is studied, a parallel baseline based phase interferometer direction finding algorithm is proposed and an implementation of the algorithm based on GPU is provided.The contradiction between array aperture and accuracy always lingers in interferometer direction finding methods. The bigger the array aperture is, the higher the precision is. However, the extension of aperture will normally make direction finding algorithms face the problem of phase ambiguity. How to solve this problem proves to be the key of the matter. In this paper, we devote ourselves to find a simple, high probability of correctly solving ambiguity and speedy interferometer direction finding algorithm.Firstly, in consideration of the property that some solving ambiguity methods are sensitive to noise, this paper proposes an algorithm based on expanding baseline to solve the phase ambiguity. Result of simulation indicates that this algorithm has a high probability of correctly solving ambiguity, even though the SNR is low. At the same time, parallel baselines are used to solve the phase ambiguity. This reduces the calculated quantity rapidly, resolves the instantaneity problem with no influence on performance. Using the pre-measured samples, correlation interferometer effectively reduces the influences of imperfect introduced by array elements. The parallel baseline algorithm based on sample library has an advantage over running time and the capacity of Resistance to system error, which is proved by the simulation test.Then, this thesis statistics the direction finding performance of the parallel baseline algorithm based on sample library systematically, including precision, sensitivity and the capacity of Resistance to system error. This paper also provides an experimental result of the algorithm used on the interferometer direction finding system experiment platform, and proposes tow processing methods aiming at some problem in the experiment for actual measurement.In order to improve the instantaneity performance, at the last of the paper, we propose an implementation of the parallel baseline algorithm based on sample library which is based on GPU. The parallelism of this algorithm is investigated before its GPU based implementation is designed and accomplished using CUDA programming platform. Finally, running time test results based on both GPU and CPU are given to improve the good performance of GPU implementation.