| Summary: | 碩士 === 國立中央大學 === 資訊工程學系在職專班 === 106 === The use of time division multiple access as a burst communication method can improve the efficiency of frequency allocation in limited-frequency bands. However, the weak energy of satellite signals and the bursting feature of burst signals are prone to engender data errors or loss. Such conditions increase incidence levels of communication errors and cause delays in communication. Therefore, this study designed an algorithm for burst-signal detection that integrates multiple methods, such as the F-test, energy detection, and matched filters. In the proposed algorithm, measures of moving average and F-test are implemented on two adjacent signal frames to obtain the similarity of the signals between two adjacent time points, thereby determining whether the signals are changing drastically. The results are then used to predict the starting point of the burst. Subsequently, the maximum value within the prediction time is identified using energy detection and matched-filter detection to acquire the precise opening time of the burst. The experiment presented in this paper was carried out under normal conditions in an environment with strong background noise and with pulsed interference for verification. The detection rate and accuracy rate for the proposed method were more favorable than those for general detection methods, indicating that the proposed algorithm may be used to enhance the quality and stability of burst communications, thereby preventing communication errors and delays.
|
|---|
