机车信号从钢轨提取轨道电路信号作为行车凭证,其译码输出性能对列控系统的可靠性和安全性有直接影响.但列车运行过程中,机车信号不可避免地混入大量噪声和干扰,译码前需要降噪以提高准确性.为此,提出一种基于全卷积神经网络(fully convolutional networks,FCN)的机车信号降噪方法,该方法利用基于原始波形"端到端"处理方式的FCN,直接从时域对机车信号进行降噪处理,以提高信噪比(signal-to-noise ratio,SNR);并利用仿真和实测数据对本方法进行了实验.结果表明:相较于传统基于频谱的滤波方法,本方法对带内干扰有更显著的效果,采用FCN能使机车信号信噪比提高8～14 dB,可有效降低带内噪声.

This paper presents the track circuit signal acquisition procedure and the structure of the receiving antenna. A new track circuit signal noise extraction method by adding a center tap of the receiving antenna is proposed. Furthermore, the Q-learning algorithm is applied to process the track circuit signal based on the extracted noise. By employing the Q-learning algorithm, the strategy of the denoising system can be adjusted continuously with the change of environmental noise. Our method automatically learns different types of noise during normal operation. Experimental results indicate the signal-to-noise ratio (SNR) of signal could be improved by 15-30 dB.Compared with the traditional filtering method, this denoising method has a better effect on the noise in the signal working band.

针对国内外高速铁路的快速发展,道岔故障严重影响行车安全及运输效率,本文提出一种基于隐马尔科夫模型的道岔故障诊断方法,通过增加道岔设备的潜在故障状态,将道岔设备的状态进行多状态细分.采用基于Fisher准则函数和主成分分析的方法进行特征提取,矢量量化处理后,建立不同故障模式下的H M M模型,通过比较测试数据与训练得到的不同HMM模型的匹配值进行故障诊断.利用京广铁路长沙南某型号道岔连续动作功率数据,对模型的性能进行测试,完成了故障诊断的实现与验证.仿真结果表明,采用四维特征信息时,其训练时间相对于其他机器学习方法有了较大提高,正确率达到90％ 以上,且该方法将道岔状态进行细分,通过分析每种状态之间的状态转移,可以预测道岔故障,从而进行道岔设备健康状态监测.

According to the transportation organization mode of Daqin Line, this paper proposes a heavy-haul railway dispatching system based on fuzzy expert system. This paper analyzes the conditions and transportation organization modes of Daqin Line, sums up and summarizes the actual experience of dispatchers, scheduling rules and strategies, establishes an intelligent dispatching system, realizes the real-time dispatching and adjustment of trains online, ensures the safety of trains, and improves the production efficiency of railway transportation.

Due to the limited inner space of the dual-channel balise,the distance between the two adjacent an-tennas is close.The multi-antennas of the balise suffer from coupling,caused by magnetic field.The magnetic coupling,called mutual inductive coupling,introduces mutual inductance in the near field region between the two closely-packed antennas.This mutual inductance causes a change in the impedance for the sensitive anten-na,which is unexpected interference.The mutual coupling between the antennas reduces the independence of the antennas and affects the performance of the antennas,which increases the difficulty of antenna tuning.The parameters influencing the mutual coupling of the multi-antennas include but are not limited to the distance be-tween the antennas,their overlapping area,and their arrangement order.In order to decrease the coupling coef-ficient of the multi-antennas,the paper proposed to decrease the overlapping area,increase the distance be-tween the two adjacent antennas and re-sequence the antenna order.The simulation results indicated that these proposed changes increased the isolation of the multi-antennas and decreased their coupling coefficient.

The mutual coupling among 4 antennas with different frequencies in the dual-channel balise destroys antenna tuning status and affects the normal operation of antennas.Based on the series and parallel equivalent circuit of antenna,the formula for calculating antenna bandwidth is derived.The three-dimensional simulation model is built for the multi-antenna in dual-channel balise to simulate and analyse the series equivalent circuit.Results indicate that the mutual coupling among multiple antennas makes the loss resistance increase,the equivalent inductance and self-resonance frequency decrease.Simultaneously,the transfer impedance is generated among antennas,which results in tuning frequency shift.The transmission coefficients from 4.234 MHz FSK transmitting antenna to 27.095 MHz receiving antenna,from 9.032 MHz PSK transmitting antenna to 27.095 MHz receiving antenna,from 9.032 MHz PSK transmitting antenna to 4.234 MHz FSK transmitting antenna,from 4.234 MHz FSK transmitting antenna to 9.032 MHz PSK transmitting antenna,are all greater than-10 dB.After low pass filters are added in the signal output terminals of 4.234 and 9.032 MHz transmitting antennas,all the transmission coefficients are reduced to below-10 dB.It indicates that adding low pass filter can effectively reduce the transmission coefficients among multiple antennas so as to realize the decoupling among multiple antennas.

The side lobe problem is a typical problem in engineering application of Balise Transmission System (BTS).Side lobe generation mechanism is analyzed based on the electromagnetic theoretical models of BTS.The influences of several factors on side lobe performance,such as Balise I/O feature,mounting modes,mounting heights and Balise Transmission Module (BTM) receiving threshold are discussed.Some suggestions for side lobe suppression methods are given.Related analysis methods and research findings have certain guiding significances to system optimization and engineering application of BTSs.

Engineering data are the important basis for balise telegram compilation . Existing engineering data verification rules are generated based on balise application principles and engineering data compilation specifica-tion of CTCS-2 .These rules ,described by natural languages ,are characterized by lack of completeness ,am-biguity and data validation insufficiency ,which may directly lead to data safety problems . To address the a-bove issues , data logic verification rules were derived firstly based on the deep data mining of constraint rela-tionship between various kinds of data , to extract data logic verification rules . Then , the satisfiability feature of the logic rules was proved through SAT modeling , ROBDD constructing and depth-first search algorithm . A data verification tool based on the derived logic rules was developed finally .Based on the engineering data sheet of an intercity mass transit line , the engineering data were verified with this tool , as well as another tra-ditional data check software . The results showed that , the new tool can identify data hazards which cannot be found by the traditional software , thus eliminating the incompleteness of primitive verification rules and greatly improving the engineering data safety .

Data , w hich has a direct bearing on the correct realization of computer interlocking function , is relat‐ed to the safety integrity level of the entire railway signaling system . The verification of data safety is a critical task in the development and application of computer interlocking in urban railway signaling .Starting from the role of data safety in system safety ,the idea of data safety verification was elaborated first .Based on the mod‐eling of computer interlocking data structures in urban railway signaling , the key rules of the safety restrictions between the data were derived . The interlocking data safety verification algorithms based on the point‐point , point‐line and line‐line relations were proposed and proved in this paper . A typical Nanjing Metro Line was used to verify the correctness of this method as a case study . Compared to the traditional manual verification process , the verification efficiency of this method significantly improved .

Chinese existing railway upgrading plan was proposed that the ground Balise could provide the data to the Train Control Record Device (TCRD). Because of the huge data requirement for the TCRD, the Balise data capacity should be expanded. Five technical proposals were mentioned and compared with each other. And the Dual-channel proposal, which consists of the Frequency-Shift Keying (FSK) channel and the Phase-Shift-Keying (PSK) channel, was considered as the better one preliminarily. Channel simulations were operated including frequency spectrum, bit error rate and spurious interference. Experiments were carried out in lab. Test results agreed well with the simulation analyses indicating the Dual-channel proposal is reasonable and feasible for Chinese existing railway upgrading.

In order to manage the hazards in the phase of train control system design and safety assessment , the quantitative safety analysis of train control system is crucial . T he results of the quantitative analysis can be used to judge and compare the pros and cons of the prototype designs , to evaluate the probabilistic risks of haz‐ards and to determine w hether hidden dangers can be controlled within the acceptable range with the risk miti‐gation measures taken . In this paper , a Markov Decision Process based modelling method was proposed to build system behavior model of the two consecutive trains in the train control system , to integrate the normal behaviours and failure behaviours of the system and to put forward Comprehensive Behaviour Model (CBM ) . The dangerous failure probability of the hazard was calculated under the probabilistic model checking tool PRISM . The methodology of quantitative safety analysis for train control system was presented .

When right side failures and wrong side failures of the track circuits occur ,the train control center cannot judge the exact states of track occupancy in sections according to the states of track circuit relays ,cau-sing the so-called"red light strip" and"train occupancy loss" on the screen of the train control center ,posing potential safety risks . .In order to reduce the operation risks ,four logic states of track occupancy were pro-posed first ,and an occupancy checking logic layer was added into the software design of the train control cen-ter ,to deal with the unsafe state information of track circuit relays ,and to transfer the safety track occupancy logic states to the centralized train control center .Markov Decision Processes were used to formally present the logic state processing models ,and four colors were employed to present the track occupation states .This re-search enhanced the inapplicable situations of the restrictions in the design specifications . Finally , through quantitative calculation ,the derived dangerous failure probability of single trip was 2 .120 × 10-10 ,which proved the acceptable range of the hazard risks of the design .

In image boundary detection, the common method is the threshold based on image segmentation. Traditional fixed grey threshold based on segmentation results in high false alarm and misses alarm rate. A framework of switch gaps detection based on CMOS plane is constructed, which focuses on the location and monitoring of switch gaps in high-speed railways. By the information theory, adaptive threshold selection and edge detection algorithm based on Shannon entropy are proposed. This algorithm shows low computation complexity and is adaptive to various contrast ratio of image with noise. It improves the detection precision and reliability. The efficiency of the proposed framework and the edge detection algorithm is verified by field data from Changsha Station in Beijing-Guangzhou high speed railway. The experiment results show that the efficiency of the proposed framework and edge detection algorithm based on Shannon entropy is improved.

Engineering data safety pervades the balise telegram compilation, which affects the railway traffic safety. The existing method to ensure data safety is to verify that whether the data conform to the specific verification rules generated based on the balise application principles and engineering data compilation specifications. However, these rules are described by natural language, which are characterized by ambiguity and scarcity. The situation often arises where insufficient verifications of engineering data may lead to data safety problems in most cases. To address this problem above, implicit logic rules are explored ground on the constraints between every pair of data, and modelled as a Boolean function firstly. Then, the Reduced Ordered Binary Decision Diagram (ROBDD) construction and depth-first searching algorithm are performed to verify the satisfiability of the rules. Compared with the original ones, the extracted authenticated logic rules have promising results in the accuracy of data checking, and can further improve the safety of engineering data.

Dual-channel balise consists of four coil antennas--27.095 MHz tele-powering receiving antenna, 4.23 MHz FSK uplink transmission antenna, 9.032 MHz PSK uplink transmission antenna and programming antenna. The dominant interaction between two closely-packed antennas is the mutual coupling. It causes difficulty in mass production. Mutual coupling introduces the impedance alteration. The tuning frequency is deviated from the operation frequency so that the antenna could not transmit the maximum power. If the deviated tuning frequency exceeds the antenna bandwidth, it cannot work normally. When the sensitive antenna receives in-band higher harmonic wave introduced by source antenna, it induces voltage which is unexpected interference. This paper analyzes the mutual coupling coefficient of multi-antenna. Some mutual coupling coefficient is more than -10 dB. In order to reduce the mutual coupling, a low-pass filter is proposed for 4.23MHz antenna and 9.032MHz antenna. The low pass filter decreases the higher harmonic wave interference. The simulation results indicate that the filter declines the coupling coefficient between two close antennas, which is less than -10 dB. It increases the antenna isolation and independence.

The existing portable balise programming device is operated manually in close distance ,which is inef-ficient and could not fully meet the field application requirements .An onboard balise programming device with automatic and long distance operation was proposed in this paper .Based on induction coupling theory ,firstly , the relationship between the shape of transmitting antenna ,side length of antenna ,transmitting current ampli-tude ,lateral displacement and inductive voltage was analyzed .The calculation results showed that the pro-gramming receiving antenna obtained sufficient energy when onboard programming transmitting antenna and balise programming receiving antenna both adopt 8-shape design .With the rectangular design of tele-powering transmitting antenna ,the balise programming circuit was not activated when the train passed a balise .The nor-mal safety data transmission was not affected .The optimization of antenna shape ,antenna size and transmitting current amplitude was studied .Experiments were carried out on the test platform .Correctness of the proposed method and applicability of antenna prototype were verified .

Balise is the indispensable component used for ground-train information transmission within the Train Control System.In this paper,on the basis of analyzing up-link signal features and establishing up-link signal process procedure,AR model was used to estimate the characteristic of noise,then kalman filter was adopt to process the up-link signal.Related parameters of kalman filter were derived from up-link signal features and the characteristic of noise.Finally,processed up-link signal could be got.The real balise telegrams of Beijing-Tian-jin intercity train were used to verify the validity and usefulness of this method.Through extensive simulations, we know that signal noise ratio of up-link signal could be improved by 25~35 dB and BER (Bit Error Ratio)re-duced by two orders of magnitude.In summary,this process method has great significance about assuring the reliability of up-link signal transmission and improving the efficiency of train operation.