Algorithm Research Of Radio Direction Finding Based On Pattern Recognition

As the fast development of modern science and technology and the information, frequency spectrum resource has become the human's main resource. Because the application of radio is more and more extensive, frequency spectrum resource is under great strain at the moment. Correspondingly, contradiction between supply and demand of frequency spectrum resource is becoming increasingly conspicuous, and interference phenomenon becomes increasingly frequent. In order to complete the management of the frequency spectrum resource, the unknown electronic jamming signals must be examined. In this way, we must product high precision finding system in which the direction finding algorithm is the centerpiece. According to the number of the target, radio direction finder technology can generally be divided into two categories: single-target direction finding and multiple-target direction finding. Single-target direction finding usually use the method of correlative interferometer. Multiple-target direction finding usually use the method of spatial spectrum estimation.The principle and technology in pattern recognition has matured a great deal over the last decade, which has been applied in many fields. It has also greatly promoted the development of intelligent system. In this paper, the principle and technology in pattern recognition has been applied in radio direction finding, which has get better performance. We improve some original methods and also give some proposals in single-target direction finding. The author's main production is contained as following:1. Based on the samples'self-learning, we have proposed the principal components analysis method to solve the problem of curse of dimensionality. The main point of the proposed principal components analysis is to select the most reasonable number which represents the most reasonable dimension. We get the satisfied results from the experiment, which prove the exactness of the method.2. Based on the radio direction finding algorithm, we have mainly lay out the vector's congruence based on the least square method. Traditional least square method mainly use phase information which get from antenna elements to find direction. With the conditions of low signal-to-noise ratio, for the noise disturbance, it can't find the direction accurately. However, vector's congruence based on the least square method has overcome this shortage. It chiefly use the antenna elements'phase difference to perform vector accumulation several times. Weighted trigonometric functions direction finding algorithm was proposed in correlative interferometer, in which the idea of weight was introduced and the factor was evaluated differently. In addition, we briefly describe the principle of spatial spectrum estimation.3. In experiment, we compare the result of the original method and the proposed method by matlab simulation. From the comparing experiment, the exactness of these proposed method was demonstrated. Experiment on spatial spectrum estimation by matlab was also given in the last part.In the end, the author summarizes the merits and demerits of the original and the proposed method, and points out different method's limitations of this paper.