Balise Up-link Signal Processing Method Based On Cognitive Control

With constant increase of the railway operating mileage and continuous improvement of the train running speed, railway signal equipments need to be more secure and reliable. Balise transmission system is an indispensable key equipment in Chinese Train Control System which means whether it can transmit information correctly has a direct impact on the efficiency and safety of railway traffic. Due to the complexity of the installation environment, balise system may be influenced by serious electromagnetic interference which leads to situations of losing point, ghost balise, telegram can’t decoding and even the train emergency brake when balise is applied practically. Research on telegram has important practical significance on account of lack of relevant domestic and international research on the balise up-link signal processing.Aiming at problems when high-speed railway balise is researched and applied, the balise up-link signal is researched by using the method of cognitive control, kalman filter and so on in this paper. The main work is shown as follows:1) Analyze the fundamental principle of balise system. Research the balise system processing procedure in practical application in which de-nosing method is 4.234MHz band-pass filter and demodulation methods consist of non-coherent demodulation and zero crossing demodulation.2) Clarify the principle of kalman filter algorithm. Present the up-link signal de-noising model based on kalman filter, carefully analyze the principles of every parts of de-nosing model which contain the derivation of kalman filter parameters, the estimation of noise characteristics and so on. Through tests it can be concluded that the SNR(Signal Noise Ratio) can enhance 25dB to 35dB and BER(Bit Error Rate) can decrease two orders of magnitude by using the method present in this paper, which verify the performance advantages of de-nosing method.3) Put forward up-link signal demodulation method based on single-frequency tracking which is realized through kalman filter algorithm or adaptive algorithm. By comparing with the non-coherent demodulation and zero crossing demodulation, it can be obtained that the proposed method can reduce BER greatly, thus verify the superiority of the proposed method.4) Describe the principle of cognitive control and apply it to the up-link signal processing procedure. Propose up-link signal de-noising and demodulation models bases on cognitive control. By introducing Q-learning, the actions of the system can be adjusted continuously with the change of external environment noise, so as to achieve the best system performance and further optimize the up-link signal processing.