Ⅰ. Introduction
Recently, the subway running in metro environments has a problem in securing the safety of passengers due to the failure of announcement in emergency situations such as breakdown, train accidents and power outage in underground tunnels as shown in <Fig. 1>. Thus, there is a need to develop an EBS(Emergency Broadcasting System, <Fig. 2>) that can provide the announcement to all passenger cars in any emergency situations on the railway route.
Recently, some of metro transportation operators in South Korea are deploying the radio-connected EBS based on zigbee and BT(Bluetooth). But, underground tunnels make it difficult to achieve good communication performance. In addition, there are many frequencies and services depicted in <Fig. 3> for the safe operation of trains and for the internet services of passenger. In particular, More than 70% of tracks operated by Seoul Metro, South Korea's biggest urban railway operator, are located in the underground tunnel. These poor radio environments make it difficult to deploy radio-based EBS to ensure the safety of passengers from train accidents (Liu et al., 2017 and Fraga-Lamas et al., 2017).
Therefore, metro transportation operators in South Korea define the following technical requirements for the radio-connected EBS. First, all EBEs(Emergency Broadcasting Equipment) have to provide passengers with emergency guidance(voice and light) in any accident or emergency situation. Also, each EBE equipped in all passenger cars must have emergency calls with the train cab or command office. In addition, it should not interfere with other radio-connected equipment installed on other trains. In particular, robust communication performance should be ensured in tunnel areas and CSZ(Communication Shadowing Zone). Finally, even if some of the passenger cars are lost due to a train accident, the EBEs of all the remaining passenger cars must operate normally. For this, the direct communication distance between the EBEs in the tunnel should be at least 200m without relay.
Considering these requirements, it seems that either LTE-R(Long Term Evolution-Railway) or WAVE using a dedicated band described in <Table 1> can be applied for the radio-connected EBS. However, since the LTE-R communication technology does not support direct communication between terminals, its application to EBS is limited(Kim et al., 2016;IEEE Computer Society, 2012). On the other hand, the WAVE(is also named the ITS-G5 in ETSI) communication technology based on IEEE 802.11p standard using 5.9GHz dedicated frequency band enables direct communication between terminals more than 200m without additional relay infrastructure (IEEE Computer Society, 2012). Especially, it was showing robust communication performance even in the metro environments such as underground tunnel, therefore, it is considered to be a suitable communication method for the EBS of the urban subway. In this paper, we investigate the performance of WAVE communication system for real urban railway environment. And then, based on the results of analysis, it will be examined the applicability as a radio communication technology for EBS. In addition, a SUB-1 communication technology, which is widely used in IoT(Internet of Thing) and smart metering in 920MHz ISM band, will be also compared and analyzed. Although it is not a fair comparison with WAVE and SUB-1 communication technology, we also evaluated the performance of SUB-1 compared to WAVE in metro environment.
The rest of this paper is organized as follows. An overview of the measurement campaign is given in Section Ⅱ. Also, devices used in the measurement as well as the measurement setup are presented together in Section Ⅱ. And then, measurement results are discussed and presented in Section Ⅲ. Finally, conclusions is summarized in Section Ⅳ.
Ⅱ. Measurement Campaign
1. Railway Network and Environments
Seoul metropolitan subway in South Korea is the most widely used rapid railway transport system in the world, featuring ten subway lines. The system serves nearly ten million inhabitants of the capital city, Seoul, and the provinces of Gyeonggi, Incheon and northern Chungnam. The total length of the subway line is about 287km (179.4 miles) of which 70% is underground. The subway has 291 stations. As shown in <Fig. 4>, this measurement campaign was conducted for the lines 2 and 3 of Seoul metro, and it was performed from the subway car depot to the main stations of lines 2 and 3 while going through the round trip.
<Table 2> shows the typical environment of the Seoul metropolitan subway Line 2 and Line 3. Both routes have ground and underground sections with over 70% underground tunnels. The Line 3 has many sections with a high bending degree. Especially, the Line 2 has the high bending degree of tunnel, and there are many the CSZ such as City Hall Station and Shindorim Station.
The maximum speed of train is up to 80 km/h, and the measurement campaign runs continuously, including all test scenarios defined in <Table 3>. Here, the evaluation criteria targets throughput, coverage, PER(Packet Error Rate), RSSI(Received Signal Strength Indicator), and the like (Unterhuber et al., 2016).
2) Measurement Setup and Test Scenarios
As shown in <Table 3>, the measurement campaign is divided into two scenarios. First, the performance evaluation is conducted for communication links depicted in <Fig. 5> which are divided into intra-link and inter-link according to the mounted location of the TX/RX antennas on the train. As shown in <Fig. 6> and <Fig. 7>, the TX/RX antennas for the intra-link are installed in the train, and the TX/RX antennas for inter-link are mounted on the roof of train. Here, the setting information of both WAVE and SUB-1 modules used in the measurement campaign are given in <Table 4>. Both modules were connected to laptops to operate the modules and to analyse received packets, which consist of packet sequence number, the GPS locations, throughput, PER, received signal power as well as other logging information. Especially, the performance evaluation for WAVE method carried out in diverse environment conditions, such as ground/underground tunnels, directional/ omni-directional antenna types, CSZ, the various bending degree of tunnel, and the movement of train.
Ⅲ. Data Analysis
In the case of intra-link, the WAVE method was satisfied with the communication coverage (over 200m) defined in the requirement when the connection doors which consist of two glass sliding door between each consecutive cars were kept open. Whereas, the communication coverage for the SUB-1 method between the terminals was about 100m despite using the hyper-directional antenna with high gain. Here, the maximum communication coverage between terminals were derived from the link distance operating within acceptable PER (10%) values. On the other hand, when the connection doors are closed, both methods cannot secure the coverage requirement. This is probably because of the shadowing effect caused by the metallic structure and the closed connecting doors inside train. Furthermore, when passengers actually ride on train, the communication coverage is expected to be further shortened due to the abrupt fluctuation of the radio-wave level and the frequency interference.
Next, the results of performance evaluation for the inter-link will be described. In the case of WAVE method, all evaluation indexes in <Table 5> were satisfied even in case of running at maximum speed (up to 80 Km/h) in the CSZ such as City-hall station and Shindorim station. In addition, the WAVE method showed an excellent communication performance results for the poor radio environment of underground tunnels such as the propagation interference caused by the train moving in the opposite direction as well as the tunnel having the high bending degree. <Fig. 8> shows the throughput value in Mbps when the data rate was set to 3 Mbps. It shows that the throughput value maintains the 2.4 Mbps throughout the entire measurement routes. Here, the throughput value in 2.4 Mbps means the maximum throughput of 3 Mbps transmission mode in MAC layer. The only noticeable behavior is the temporary throughout value drops occurring in the CSZ such as the Shindorim station. This temporary drop taken place when the train was entering the some station, and then it recovered immediately. In <Table 5>, The RSSI was always in the range between −61 and −82 dBm. This RSSI range is sufficient to operate normally considering that a minimum sensitivity for the WAVE method is to -99dBm. Also, the PER values where measured for each second by taking the error rate for each 1000 packets sent in one second. For this measurement campaign, the PER could not be measured for values less than 10−2. Therefore, the PER values are considered high for safety critical applications but we noticed that there were no major outages throughout the measurements.
However, in the case of SUB-1 method, it does not satisfy with all the evaluation indexes in <Table 5>. Especially, we confirmed that the RSSI value drops drastically in the spot of CSZ or when entering the tunnel. Also, even though the minimum sensitivity of SUB-1 method is -109dBm, the received signal with a higher signal strength than this value was not recovered normally. As described prior, the communication distance of SUB-1 method for the inter-link does not increase even though the directional antenna with a high antenna gain is applied. It is assumed to be due to the limitation of the SUB-1 technique using the single carrier based modulation scheme vulnerable to multi-path channel environment.
For investigating the influence of the antenna directivity, the directional and omni-directional antennas are considered under the metro underground environments and CSZ. The results depicted in <Fig. 9> show that the omni-directional antennas can offer better throughput in metro tunnel environment, while the directional antennas can dramatically reduce the throughput when it starts entering curved tunnel with high bending rate. One of the reasons of this drop is the beam pattern of directional antennas. It is too narrow to secure the LOS(Line of Sight) channel path in curved NLOS tunnel environment. Therefore, the use of omni-directional antennas is expected to support more suitable performance in urban railway environment. Table 6
Ⅳ. Conclusion
In this paper, the applicability of various wireless communication technologies for the EBS through the measurement campaign was examined in Seoul metropolitan subway. The WAVE is a communication technology that can use 5.9GHz dedicated frequency band without charge and it is possible to communicate between terminals without the help of additional relay infrastructure. Especially, it confirms robust communication performance in urban metro environment. Therefore, it is considered to be suitable as a communication method of the radio-connected EBS for an urban subway. It is also expected that it will be possible to develop the EBS capable of safe evacuation and emergency guidance to passengers using urban subway in any emergencies.
