Reliability Evaluation for LTE Based CBTC Train Ground Communication Systems

Nowadays, most communication based train control systems (CBTC) use Wireless Local Networks (WLAN) technology for information transmission. However, because of the high running speed of the train in urban rail transit, WLAN technology cannot fully satisfy the information transmission under the condition of train operation at a high speed. Long term evolution (LTE) technology has a high transmission speed, which can better meet the real-time requirements of train and ground information transmission. Therefore, LTE technology has become an emphasis on the research of communication technology in today's urban rail transit. Although there is a lot of research on the application of LTE technology in urban rail transit, there is little research on the reliability of LTE based train ground communication systems using experimental data. In this paper, the reliability of train ground communication is firstly defined. Then the train ground communication environment was established in the laboratory, and the performance parameters of train ground communication based on LTE technology were obtained in the test. Finally, the reliability is calculated and analyzed according to the experimental results, and a storage method of train data in actual operation is proposed, which can be used to analyze the reliability of train information transmission. The results show that the reliability of train ground communication based on LTE technology meets CBTC requirements.


Introduction
In the process of national economic development, rail transit played a signi cant role. Safe and e cient rail transit has laid a solid foundation for high-tra c passenger transport and high-speed freight transport. However, with the acceleration of China's urbanization process, people have put forward higher requirements for urban rail transit, and its reliability is the most important.
Urban rail transit possesses the characteristics of high speed and high tracing density. Nowadays, urban rail transit adopts CBTC systems as train control systems. e CBTC system can trace the train operation closely according to the situation of the forerunning train and the forward route.
is not conducive to ensuring the safe operation of the rail transit.
Nowadays the time-division-long-term evolution (TD-LTE) technology came into being in the urban rail transit system. For example, the newly opened Beijing Yanfang Line is based on TD-LTE. Possessing a series of advantages such as strong anti-interference capability and continuous coverage capability [7]. TD-LTE is a technical version of LTE that adopts time division duplexing.
LTE is a global common standard developed by the 3GPP (3rd Generation Partnership Project) based on OFDMA (Orthogonal Frequency Division Multiple Access) technology. As the mainstream of the fourth generation of communication technology, LTE can achieve faster data transfer rate and more stable handover. At the same time, LTE can reduce delay and packet loss rate e ectively [8].
Compared with WLAN technology, LTE has a stronger advantage in urban rail transit [9]. ere are main di erences in spectrum resources between LTE and WLAN. LTE works in the licensed frequency band and needs to be employed on licensed spectrum resources. WLAN uses the same frequency band with bluetooth, ZigBee, and other systems without allocating frequency resources, and it is more susceptible to interference [10,11]. LTE supports exible scheduling policies that can distinguish users from time and frequency and ensure the requirements of QoS [12]. However, the WLAN uses the mechanism of preemption and competition among users to schedule users. Resources are exclusively occupied by one user at a time. erefore, collisions occur when the number of users is large enough, and the e ciency of resource utilization is reduced [13]. e introduction of LTE in the CBTC system can reasonably utilize the relatively low 1.8 GHz frequency and the radio frequency leakage cable as the transmission medium, which can e ectively enhance the reliability of the train-ground communication in the CBTC system [14]. Literature [15] introduced two crucial performance indicators of the packet loss rate and delay in the train communication system. Because of packet loss, if the control information is not able to be sent to train, the e ciency of train operation will be a ected severely.
Also, the packet delay must also be controlled in a certain range. Otherwise, the information received by the train is not real-time information which makes a negative impact on the accuracy of train operation. In [16], literature studied the train control problem in CBTC systems of the wireless network.
China will use LTE-based wireless communication technology in a larger range of urban rail transit. erefore, research on LTE communication technology in urban rail transit is particularly important. Although there have been many studies on LTE technology so far, few researches have been done to study the reliability of LTE based CBTC train ground communication systems based on experimental data.
In this paper, the reliability evaluation of the CBTC system based on LTE technology is carried out. e main content of this paper is as follows: (1) We describe the framework and analyze the requirements of CBTC systems. (2) Using the transmission delay and packet loss performance parameters, we de ne the reliability of the CBTC system. (3) A real test environment is set up. We test the transmission delay and packet loss of the communication system, and the system reliability is evaluated. e rest of this paper is organized as follows. Section 2 introduces the principle of CBTC and describes the requirements of train control systems and reliability de nition of train ground communication systems. e testing environment based on the laboratory, which includes the environment of train-ground communication, is presented in Section 3. Section 4 is about the results of testing discussion and prospect.

Safety-Critical Urban Transit Service
e Principle of CBTC. e basic structure of the existing CBTC system is shown in Figure 1. It mainly includes data communication system (DCS), automatic train monitoring (ATS) system, computer interlocking (CI), zone controller (ZC), and vehicle controller (VOBC) [17]. e leading technologies of CBTC include [18]: (1)  may not receive the information of MA in time, causing unnecessary braking and traction [19]. Hence, reliability is an essential index in wireless communication systems. Train running on the track will always receive information from equipment on the ground and send messages to sta s. When the wireless communication link is not available when needed, trains may not be able to successfully obtain the control messages about operation plan from systems such as ATS or CI, and cannot send running state of train or alarm messages to dispatchers. It will have a signi cantly negative impact on train control performance. erefore, the reliability of wireless communication systems needs to be de ned and calculated to evaluate the CBTC system's performance.

e Requirements of CBTC.
In CBTC systems, based on the position of all the trains and obstacles along the rail and the operation plan acquired from ATS, each train computes its MA. e real-time braking curve is calculated using the newest MA by ATP on the train. When the train's speed get to the limited velocity on the braking curve, the train will start decelerating to ensure that the speed of train will never exceed the warning line and travel out of its MA. e ATS sets the trip time between two stations. According to the trip time and some other performance indices, energy savings and passengers comfort, for example, and then the ATO derives an optimized guidance trajectory. e optimal velocity of the train in a speci c position is obtained from the velocitydistance guidance trajectory. e main task of ATO is to make the train run according to this optimized train running curve. Figure 2 shows the impacts of wireless communication reliability on CBTC systems performance. As shown in Figure  2(a), when the communication is available, the updated position of the Train B, whose position is located far from Train A in the gure, will be sent to the following trains in every communication period using the communication network. e following train will operate as the braking curve and will never exceed the MA point unless the MA is updated. However, when the communication interruption or delay in data communication systems is too substantial, the newest MA point may not be able to be sent to the following train by ZC for a certain period. As shown in Figure 2(b), before the updated MA arrives, the old MA is still used by the following train. When the ATP receives the updated position, it will calculate a new breaking curve and the train will come back to the new optimized guidance trajectory controlled by ATO. Compared to the scenario when communication is always available, the last arrived train position caused by communication interruption or delay makes the train travel deviate from the guidance trajectory. Substantial energy will be lost during this process, and it will severely a ect passenger comfort. More importantly, it will increase the trip time between stations, which will a ect the operation of all the trains behind.

Reliability De nition of CBTC.
In the process of data transmission, transmission delay will be bound to happen, and because of channel fading and other adverse environmental factors, packet loss will also take place. In one case, just as shown in Figure 2(a), when Train B is in front of Train A, on account of transmission delay, when Train A gets the messages about operation plan and according to that to calculate MA, the actual MA is further than the MA that Train A has calculated considering that Train B is still running. And in the other case, as shown in Figure 2(b), when Train C get into this line from another one, Train A has got a MA at the end of Train B and it needs to calculate a new MA at Train C. However, because of packet loss, Train A cannot receive the newest information. In Figure 3, the dots of di erent colors in the gure indicate the data packets in the transmission, the blue dots indicate the packets successfully received by the receiving end, and the red dots indicate the data not successfully received by the receiving end. When the time window window takes di erent values, the number of consecutive lost packets has di erent e ects on the reliability of the system. Just like in Figure 3(a), the value of the time window is larger than the time for transmitting two data packets. erefore, the loss of these two data packets does not a ect the reliability of the system, and as de ned, the loss of one data packet does not a ect information transmission either. Moreover, in Figure 3(b), the value of the time window is less than the time for transmitting three data packets, so when three or more data packets are consecutive losses, the transmission of information is deemed to be a ected. e more consecutive packets loss exceeds the time window window , the higher the loss , the lower the reliability of the train communication. As shown in Figure 3(c), the time window takes longer than the time it takes to transmit a packet. Loss of two packets will also a ect the transmission of information.
e reliability of train communication is system , which consists of delay level reliability delay and packet loss level reliability loss . e main parameters needed to measure the reliability of train communication ( system ) are as follows: (i) window : e de ned time window. When transmission delay or consecutive packets loss time exceeds the time window window , it is considered that the transmitted data packet fails to reach the receiving end (ii) delay : e total number of packets whose delay is higher than the time window window (iii) loss : e total number of dropped packets that the consecutive packets loss time exceeds the time window (iv) delay : e total number of packets sent during the delay level reliability test (v) loss : e total number of packets sent during the packet-loss level reliability test (vi) delay : e ratio of the number of packets whose delay do not exceeds the time window window to the total number of transmitted packets (vii) loss : e ratio of the number of consecutive dropped packets not over the time window window to the total number of sent packets (viii) system : e reliability of train communications, consisting of delay level reliability and packet loss level reliability.
By de nition, the delay level reliability is delay = delay / delay and the packet loss level reliability is loss = loss / loss . So, the train communication reliability system consisted of loss and delay , is calculated from the results of the two tests based on the proportion of the total number of packets transmitted by the two tests. erefore, the reliability formula for train communication system can be derived as: If Train A still runs according to old MA, Train A will have a collision with Train C. It is against the requirements of CBTC.
According to the above description, we use transmission delay and packet loss to de ne the reliability of wireless communication system [20]. In the transmitting process, it is desirable to maintain the communication link between two communication terminals available within a speci ed period of . erefore, we de ne the data communication reliability in CBTC systems as the probability when one communication terminal can successfully receive information from the other communication terminal within the time window window . Reliability is divided into two levels, including delay level reliability and packet loss level reliability. At the same time, the reliability of the wireless communication system is calculated from these two levels of reliability.
Delay reliability is to measure the reliability of communication according to the transmission delay of packets, indicated by delay . It is de ned as the percentage of packets that are delayed by no more than the time window and the total number of packets transmitted. When the transmission delay of a packet exceeds the speci ed time window window , it is considered that there is a problem in the information transmission of the packet. In CBTC systems, it is considered that data packets are exchanged at every communication period. When the time between two consecutive received data packets is greater than time window window , the data communication system is considered to be unreliable ł the higher the number of packets whose delay exceed the time window window , the lower the reliability of communication system.
Packet loss reliability refers to the reliability of communications measured by the number of lost packets, represented by loss . It is de ned as the percentage of the number of packets not a ecting the communication and the total number of packets. In this paper, continuous packet loss duration can be obtained based on the data transmission rate and packet size. We believe that when the time taken by consecutive packets loss exceeds the de ned time window window , packet loss will a ect the reliable transmission of system information, and when the consecutive packets loss time is less than the time window window , the standard transmission of system information will not be a ected. performance, and in the range of 20-500 km/h, it provides services of equal or better quality than 3GPPR6. Moreover, LTE technology supports the use of services not only under high-speed mobile conditions but also under low-speed conditions.
In commercial cell networks, 400 users can be activated under 5 MHz bandwidth, but the quality of service for each user will be degraded signi cantly, which cannot be acceptable in the urban rail system. Indeed, in the real system, it is not recommended to activate more than 10 users when the bandwidth is 5 MHz. As is known, the bandwidth can be set as 1.4 MHz in the LTE system. However, because of the integrated tra c, including train control information, PIS, CCTV, and other information, and limited throughput under this bandwidth con guration, the train control information is the only tra c that can be transmitted through the wireless communication system in CBTC.
Here, we built an indoor test environment at the State Key Laboratory of Rail Tra c Control and Safety at Beijing Jiaotong University. In the lab, we prepared the hardware devices needed for the test, including LTE devices for communication, as well as devices such as a programmable attenuator, xed attenuators, and channel emulator. e connection of the device is as shown in Figure 5.
EB Propsim, as a channel simulator, simulates the attenuation, fading, time delay, and Doppler frequency shi characteristics of single-channel wireless transmission during train operation using precomputation les to simulate the transmission environment. A xed attenuator is used to adjust the intensity of the signal received by TAU.

Test Procedure.
Regarding the test, rst, we set the parameters we need. In this lab, we chose to test the CBTC service as a test service and use IxChariot to simulate service performance testing and install it on the server and client to complete the basic performance test of LTE. According to the business requirements of the CBTC in the signal system, the data packet of the service ow adopts the UDP protocol. Moreover, the data packet size is set to 400 bytes. e data transmission rate is 256 kbps downstream, and the uplink is 256 kbps (the statistical interval of the test indicator is 1 second).

Test Environment.
e CBTC system utilizes wireless communication to realize communication between the ground device and the onboard device and two-way large-capacity information transmission so that the ground device can timely transmit the speed limit condition of the front line and vehicle device can obtain the front line information in time. e running train calculates the optimal braking curve in real time and runs according to the running curve, thereby increasing the line capacity, reducing the frequent deceleration braking, improving passenger comfort and signi cantly improving the safety of train operation.
Nowadays, Researchers began to study the use of LTE technology as a communication technology for CBTC systems. LTE is a communication technology with a complete QoS transmission management strategy and higher stability than WLAN. Also, LTE communication technology has been used to transmit messages on some rail transit lines. Not only that, but scholars have begun to research and improve LTE technology for use in next-generation train control systems. Such as, in [21,22], the author studied the use of CBTC system based on LTE-M with FlashLinQ T2T communication.
Just as shown in Figure 4, it adopts a at network structure, including a three-layer architecture (EPC, E-UTRAN, and UE) and use IP transmission. e LTE-based wireless communication technology can realize timely and accurate transmission of information such as CBTC information, train status monitoring information, video surveillance (CCTV), and PIS (including emergency text), providing a solid foundation for the safe and e cient operation of urban rail transit systems.
Compared with GSM-R, LTE wireless communication technology has many advantages in rail transit. Some of the performance bene ts are shown as follows: (1) Latency. In the process of information transmission, the delay time of the UE to user plane of the eNodeB is less than 10 ms, and the delay time of the control plane is less than 100 ms providing better conditions for real-time information transmission between ground control equipment and vehicle equipment. (2) roughput. At a Bandwidth of 20 MHz, the downlink peak data rate reaches 100 Mbps, and the upstream peak rate reaches 50 Mbps.  Journal of Advanced Transportation 6 technology, tested in the laboratory, including the RTT transmission response time and the wireless network packet loss rate, and calculate the reliability of the train-ground communication system. is article will analyze the transmission performance under the 1.4 MHz bandwidth in detail, and calculate the reliability of the train-ground communication system under the experimental environment conditions. In this experiment, Figure 6 is the transmission response time chart of a CBTC service at the near point, and Figure 7 is the transmission response time chart of a CBTC service at far point. As can be seen in Figures 6 and 7, the data transmission time lasts 5 minutes, and the round-trip response time of train-ground communication is less than 300 ms, which indicates that the train-ground communication system based on LTE technology can transmit information stably. It also can be seen from gures that the response time of the near point is less than the response time of the far point. Figure 8 shows packet loss rate charts of CBTC business at 1.4 MHz bandwidth. In this gure, we can see that the packet loss rate test lasts for 55 minutes. During the data transmission process, there is a certain amount of data loss. And during the entire data transmission process, the loss of data packets is a ected by real-time changes of environment and performance of equipment. Figure 9 shows the continuous packet loss rate of the CBTC service. It can be seen that there is a continuous loss of di erent numbers of packets. According to the di erent transmission information, there are di erent in uences on the communication between the train and the ground. In this experiment, whether the lost data packet will a ect trainground communication is determined by value. e higher the value is, the less the impact on train ground communication. In this experiment, a er statistics, there are only two or three consecutive packet losses at the near point. While at far point, there exist more than four consecutive packet loss. Figure 10 shows the reliability of delay-level at the near point and far point. Figure 11 shows the reliability of packet loss level at the near point and far point. According to the de nition, the reliability of delay level and packet loss level are obtained under di erent time windows. Data packets with delays longer than are excluded, and they are considered to have failed to be transmitted successfully. If consecutive packet losses occur, the continuous packet loss time does not exceed the time window at this time, and it is considered that it does not a ect the data transmission. As can be seen from the Figure 10, the reliability of the delay level and the reliability of the packet loss level both reach a certain value, and a er the time window reaches a certain length, it can maintain a relatively stable value. Also, from the gure, we can see that when we take the same time window , the reliability of the near point model is higher than that of the far point model in terms of transmission delay and packet loss rate. Figure 12 shows the reliability of the train-ground communications system at di erent values and di erent static models. According to the data processing method described in the previous section, the reliability of the train-ground communication system is calculated based on the reliability of the transmission delay level and the reliability of the packet loss In an LTE network, di erent bandwidths are set for transmitting CBTC services under di erent conditions. However, in this test, since the CBTC service transmits a relatively small amount of tra c, a transmission bandwidth of 1.4 MHz is selected. Moreover, because the train-ground communication system can better accomplish transmission tasks with higher bandwidth, choosing a smaller bandwidth to analyze its reliability can indirectly indicate that the train-ground communication system is more reliable at greater bandwidth.
At the receiving end, the signal strength of the ground information that the train can receive is di erent during the actual running of the train, and the data transmission is better when the signal strength is large. erefore, the test selected two static signal models for discussion, including near-point and far-point models. We specify that the RSRP of the signal is less than 75 dBm at the near point and the RSRP range of the far point is [−85, −90] dBm.
To simulate the channel environment during train operation and to test LTE-based train communication capabilities in high-speed operating environments, we used the EB Propsim channel simulator. e load channel model has a train speed of 200 km/h, using the ITU-VA channel model and standard Doppler spectrum.
Particularly, in the latency test, the packet size was set to 400 bytes, and 2000 packets were tested. We need to add a timestamp to the analog service data packet sent by the sender and calculate the delay by analyzing the timestamp of the analog service data packet received on the receiving endpoint. In this experiment, the data was processed before the analysis. e response time data obtained in the experiment is the round-trip response time of the communication between the train and the ground. We do not consider the environment, equipment and other problems in the actual scene. Here, our default transmission response time is equal to the received data response time.
e following are the test steps for communication between the train and the ground terminal. Before starting the communication test between the train and the ground terminal, rst connect the device according to the connection method shown in the Figure 5, and con rm that the unit can be set normally, and the terminal can search the network. en we need to load the channel model on the channel simulator and plug in the xed attenuator. Finally, open the IxChariot script on the vehicle analog terminal, correctly load the CBTC service and set the transmission parameters as described above. When we are ready, we can start testing. First record the device model, so ware version, RSSP, SINR, system bandwidth, AMC adaptive parameters and whether ICIC is enabled. A er that, start the service simulation so ware and load the one-way CBTC service. In the delay test, the test lasted for 5 min, while in the packet loss test, the test lasted for 55 min. Test and record the transmission delay and total packet loss rate of the uplink and downlink packets.

Test Results.
In this section, we give the test results of the train-ground communication system based on LTE train-ground communication system in urban rail transit. Also, it has extremely high reliability, indicating that it is su ciently suited for transmitting the information of the event recorder.
During the operation of the train, due to the mutual electromagnetic interference between the equipment in the subway system and the in uence of the natural environment, or the improper operation of human beings, the train may have faults and accidents. At this point, the analysis and evidence collection of the cause of the accident is particularly essential. erefore, a train event recorder capable of recording information in real time is required to record the operating status of a various device, the real-time information of the train operation, and the related operations of the train sta during the running of the train.
In the 1990s, the Institute of Electrical and Electronics Engineers speci cally described the train event recorder in relevant technical documents. Also, in the RAILROAD LOCOMOTIVE SAFETY STANDARDS in the United States, the event recorder is clearly de ned: (CFR 49 Ch II 229.5): "Event recorder means a device, designed to resist tampering, that monitors and records data. " e event recorder will completely record the train-triggered status and various operational information during the continuous change as shown below: level. It can be seen from the gure that the increase of the value can improve the reliability of the train-ground communication system. e reliability can reach over 80%, and the reliability of the train-ground communication system is related to the reliability of the transmission delay level and packet loss level. e reliability of the train-ground communication at the near-point model is higher than that at the far-point model.

Discussion and Prospect.
From the above test results, it can be seen that the transmission delay and packet loss of train-ground communication remain below a certain level regardless of the bandwidth. Also, we can see that even under an unfavorable con guration which are at 1.4 MHz bandwidth and low RSRP, the train ground communication system still has high reliability, indicating that LTE technology can complete the communication of train and ground and meet the requirements of the urban rail transit system. At the same time, due to the limitations of the eld test environment, there is strong signal interference. However, reliability is still good. e result shows that under unfavorable environmental conditions, LTE technology can also complete the data transmission between train and ground in a high velocity and accuracy ensuring the safe and e cient operation of urban rail transit. erefore, we conclude that LTE technology applies to the A er the event recorder stores the information, whether the data can be obtained e ectively is of great signi cance. In the event of a train accident, there is a high probability that the incident recorder will be severely damaged, so that the stored data will not be accessible to accident investigators. erefore, how to save the acquired information well becomes a necessary premise for analyzing the cause of the accident.
According to the characteristics of information transmission in rail transit, we propose an storage scheme based on the existing information storage methods. Using LTE communication technology, all equipment information and operating status related to locomotive operation during train operation are sent to the cloud storage space through the wireless communication system in real time. erefore, the ground sta

Conclusions
In this paper, we have introduced the urban rail transit CBTC systems and train ground communication systems. According to communication performance parameters, the reliability of the urban rail transit CBTC train communication system based on LTE was de ned. en, the reliability of the train ground communication system was calculated using delay and packet loss. Besides, we also proposed a new storage method of information of train using communication network. e results show that the train ground communication system using LTE technology can complete the information transmission of urban rail transit systems and it has high reliability.
Data Availability e data used to support the ndings of this study are owned by a company.

Conflicts of Interest
e authors declare that they have no con icts of interest.
can check the train running condition and equipment status in real time, and analyze whether the train is in a normal state according to the existing experience and existing analysis methods to ensure the safe operation of the train.
Cloud storage is a model of online storage. Information is stored on the Internet, and capacity is provided on demand, enabling fast, persistent, and anytime, anywhere access to data. From the perspective of information storage, cloud storage has substantial advantages, including unlimited storage space, so the information of various states during train operation can be saved, providing researchers with more comprehensive and detailed data parameters, which is conducive to the simulation of train operation process and analysis of train operation fault by employees. On the other hand, in terms of the reliability of information storage, cloud storage adopts a third-party platform, and the stored data will not be lost or damaged due to the failure of the railway system. LTE communication technology can guarantee a high data transmission rate. erefore, it is incredibly reliable to use LTE-based communication technology to transmit the information from the event recorder to the cloud in real time. By using LTE wireless communication technology, the data is uploaded to the cloud in real time, and the sta can access the corresponding modules according to their needs and get the latest data at any time.
As shown in Figure 13, the event recorder obtains the information of equipment in each carriage and relevant data of running state, which is stored in the event recording memory. In the process of train operation, the train is always under the coverage of the LTE communication network. e train transfers the in-memory information to the cloud at any time and anywhere through one or more core networks, to realize high-capacity and highly reliable data storage.