Random Signal Processing Applied In Engineering

Today, signal analysis and processing have been used in a good many fields. Especially, the random signal processing has been widely applied in various domains such as communication, radar, sonar, image processing, speech processing, seismic (earthquake signal) processing, geological information processing, biomedical engineering and economics etc. Based on the requirement of precise describes of signal models, random model become popular. Plenty of random theories and tools have been quoted into lots of engineering domains in succession, which has achieved good effects and are crucial tools in modern signal processing of radar and communications. The main task of this thesis is to solve two practical problems on the methods of random signal processing, the first problem is the flight conflict detection based on Kalman filter, and the second is the PN codes'blind estimation of DS signals. Two main aspects are included in this paper. One is the detection of fight conflicts based on Kalman filter, another is the blind estimation of PN code of direct-sequence spread-spectrum signals. In the first part, the difficulties of blind estimating PN codes of DS/SS signals modulated by high velocity signals are discussed. Based upon the introduction to existing methods, an eigenvalue method is purposed, which can compute the PN codes when one period PN codes is modulated by several signals, and some useful information can be obtained from eigenvector, such as velocity of signal. The simulating results show that the PN cods can be picked up, even when SNR is –9dB. In the second part, considering measure error and the random effect in fight, a detection method of conflicts based on Kalman filter is presented from random frame. Kalman filter model is constructed to simulate the flight track with the flight plan and radar orientation system, and compute the probability of aircraft in the conflict Region. To avoiding the complexity of analytical algorithm, Monte-Carlo method is used to compute the probability. The model is more approximate to the real process. Both theory analysis and simulation results show that this method not only is efficient to detect the conflict when conflict's posture is obvious, but also it can provide the prediction of the latent conflict generated by random effect.