Session 4: Applications and Infrastructure
HotWireless’17, October 16, 2017, Snowbird, UT, USA
Optically Backhauled Moving Network for Local Trains Yu Nakayama
Kazuki Maruta
University of Tokyo 4–6–1 Komaba, Meguro–ku Tokyo, Japan
[email protected]
Chiba University 1–33, Yayoicho, Inage-ku, Chiba-shi Chiba, Japan
[email protected]
Takuya Tsutsumi
Kaoru Sezaki
neko 9 Laboratories 1–9–7–1015, Kitashinagawa, Shinagawa–ku Tokyo, Japan
[email protected]
University of Tokyo 4–6–1 Komaba, Meguro–ku Tokyo, Japan
[email protected]
ABSTRACT
1
The concept of moving cell in cellular systems has been discussed for 5G group mobility where rapidly moving platforms such as trains carry a large number of user terminals. It has been considered to employ wireless backhaul for moving cell, the problem of which is its limited and unstable bandwidth. To realize high bandwidth with wireless backhaul, a large number of base stations (BSs) are required along the railway. This paper proposes the concept of optically backhauled moving network (OBMN) for local trains to efficiently provide backhaul links for local trains. In the OBMN, an autonomous BS (ABS) is set on the top of a train, and is connected to a gateway via optical backhaul. To follow the movement of the train, the length of optical fiber is adequately adjusted with a high-speed reel, which is located along the railway and spins to reel up and send out the fiber. When the train arrives at a transfer point, another ABS is set on the train and the demand is handovered to the newly set ABS. The number of required BSs and deployment cost will be drastically reduced with the proposed OBMN compared with the existing static wireless backhauling deployment.
Wireless access services have been expanded rapidly and mobile traffic is increasing exponentially [3]. Supporting demand users via moving hot spots have attracted considerable attentions toward 5th generation mobile communications (5G) networks [12]. Moving hot spots are generated by the tremendous increase in data traffic of mobile users using smartphones, tablets, and laptops via mobile networks onboard. Since the number of mobile users onboard has increased dramatically these years, the moving traffic demands have been densely distributed on public transit vehicles such as buses, trums, and trains. It becomes a significant research topic for network operators to improve onboard mobile coverage and capacity to satisfy the moving demands. There have been various technologies developed for different scenarios of moving demands. One of the most promising solution is to deploy the mobile/moving relay nodes (MRNs) on the vehicles [23][22][1][2][14]. The MRNs consist of outdoor and indoor antenna units connected with cables. An MRN moves along with the mobile users within the vehicle, and communicates with the donor evolved nodeB (DeNB). The users onboard keep their network connections by properly placing DeNBs and MRNs. Mobile relay offers three advantages; high rate communications with good channel condition between users and the MRN, small signaling overhead for group handovers, and low transmit power of user equipments enabled by the spatial closeness to the MRN [1]. The performance of MRNs was compared with the single-hop direct transmission and fixed relay nodes in [22], and they showed that the MRNs deployed on top of public transportation vehicles bring significant enhancement to the quality of service (QoS) experienced by the users in the case of moderate to high vehicular penetration loss (VPL). The challenges for MRNs are pointed as the VPL, user mobility management, and handovers [23]. In particular for ultra-dense urban scenarios, the major challenge in deploying moving networks inside vehicles was the inter-cell interference, which is exacerbated by the urban canyon effects and can be reduced with interference management approaches such as multi-antenna interference suppression techniques [20][21].
CCS CONCEPTS
Networks Network design principles; Mobile networks; Computer systems organization Robotics;
KEYWORDS moving network; moving cell; optical backhaul; autonomous base stations
Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit o r c ommercial advantage and that copies bear this notice and the full citation on the first page. Copyrights for components of this work owned by others than ACM must be honored. Abstracting with credit is permitted. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior specific p ermission a nd/or a f ee. R equest permissions from
[email protected]. HotWireless’17, October 16, 2017, Snowbird, UT, USA. © 2017 Association for Computing Machinery. ACM ISBN 978-1-4503-5140-9/17/10. . . $15.00 https://doi.org/10.1145/3127882.3127885
31
INTRODUCTION
Session 4: Applications and Infrastructure
HotWireless’17, October 16, 2017, Snowbird, UT, USA
As regards efficient small cell deployment, the concept of moving small cells was presented [4][9]. The performance of moving small cells was evaluated in [4] for the scenario where mobile small cells are mounted in public buses, aggregate user traffic, and communicate with macro cells through wireless backhaul links. In [9], it was supposed that a small cell is deployed on the top of public transportation circulating in crowded streets, i.e. a bus or a taxi, and it allows to carry traffic generated by the passengers in addition to the one coming from its vicinity. This scenario was indicated to be efficient for deploying small cells in crowded areas in the presence of stationary traffic hotspot inside a macro cell. The mobile femtocell is another concept that has been recently proposed for moving demands [16][18][17][15][7]. This is a scenario for deploying heterogeneous and small cell networks (HetSNets) which provide a good throughput using small cells close to mobile users [8]. The essence of mobile femtocell is to adopt the femtocell technology inside vehicles. Femtocells are low-range and energy-efficient mobile base stations (BSs) that is suitable for improving coverage inside homes and buildings. With mobile femtocells, the small cells move around and dynamically change their connections to core networks. The performance of mobile femtocell was investigated in [16][18][17], and they proved that mobile femtocells users enjoyed better QoS than fixed femtocells users. The implementation of mobile femtocells can also reduce the signaling overhead and enhances system capacity [7]. A related concept of a vehicular network with cellular infrastructure as a backbone was proposed in [15]. For this purpose, mobile femto access points are used as relays in place of road side units (RSUs). Throughput and delay performances of the proposed concept were better compared with those of the conventional IEEE 802.11p vehicular networks. As stated above, there have been huge research efforts related to moving networks. This paper focuses on moving networks for local trains in urban and suburban areas. To provide broadband Internet access in trains has been significant research topics [5], and a major one of which is vehicular communications systems for high speed trains (HSTs) at speeds of up to 350 km/hr, 500 km/hr, or higher [26][24], for various subjects such as cooperative MRN [19][13], multiple-input multiple-output (MIMO) channel model [6], millimeter-wave beamforming [10], and distributed antenna systems [25]. However, to the best of our knowledge there is no architecture specialized for local trains in urban and suburban areas. They go and return on railways that range from several kilometers to several tens of kilometers at speed up to around 120 km/hr. The moving network technologies for HSTs are relatively costly for local trains, because such technologies are optimized for high-speed moving objects. Thus, in this paper we propose a novel concept of moving networks optimized for local trains utilizing the characteristics of them. This paper proposes the optically backhauled moving network (OBMN) architecture for local trains. The idea behind the proposed architecture is that optically backhauled BS can provide high and stable bandwidth for the moving demands,
compared with the wireless backhaul. In the OBMN, a BS is set on the train and connected to a gateway node through an optical backhaul, which consists of fiber optic networks. The traffic at moving demands is forwarded to and from the BS with a radio access technology (RAT). While the demands onboard move as the train moves, the BS communicates with the gateway through high-bandwidth optical fiber. The optical fibers can be laid along overhead lines without interruptions in local trains. To follow the movement of the train, the length of optical fiber between the BS and the gateway is adjusted with a high-speed reel. If the train reaches certain points, the BS on it is autonomously unloaded and another BS is autonomously loaded, which is called transfer. This transfer process executes the handover in the OBMN, and the fiber length can be limited with it; thus the OBMN can ensure the certain range of propagation delay in the optical fibers. The proposed architecture can efficiently provide higher bandwidth than the existing moving cell architectures that employ wireless backhaul. In the following, a BS that provides autonomous movement functions is referred to as an autonomous BS (ABS), and the adjustable optical backhaul for a moving ABS is called the optical reflex backhaul (ORB). The rest of the paper is organized as follows. Section 2 introduces the proposed OBMN architecture. The contributions and challenges for the proposed architecture is summarized in section 3. The conclusion is provided in section 4.
2 OBMN ARCHITECTURE 2.1 Concept The concept for the proposed OBMN is described in the following. The architecture of OBMN is depicted in Fig. 1. An ABS is set on the top of a train, and is connected to a gateway via optical backhaul. Access point to be connected to user equipments shall be equipped on ceiling inside the train. To follow the movement of the train, the length of optical fiber between the ABS and the gateway is adequately adjusted with a high-speed reel, which is located on telecommunications facilities such as telegraph poles, or stations. The structure of the high-speed reel is similar to a fishing reel. It spins to reel up and send out the optical fiber to adjust the length of fiber. Although MRNs can be used in the OBMN, we explain the proposed architecture taking the simple moving antenna as an example. The advantage of the proposed OBMN is introduced with Fig. 2 as regards the efficiency for demand satisfaction. Consider a scenario where the mobile traffic demands onboard move with the train. Fig. 2a shows the conventional moving cell architecture, which we call the static architecture. In this scenario, a moving antenna is set on the top of the train, and communicates with static ground BSs deployed along the railway via wireless backhaul. To achieve high throughput in this backhaul connections comparable to optical fiber connections, the cell coverage of each BS is limited. Thus, the ground BS spacing cannot exceed certain distance, around 100 m [27]. The problem of this architecture was that the number of required static BSs increases in proportion to the
32
Session 4: Applications and Infrastructure
HotWireless’17, October 16, 2017, Snowbird, UT, USA
(a) Static architecture
Figure 1: Concept of OBMN architecture.
length of railway lines. Fig. 2b shows the conceptual deployment model of the proposed OBMN architecture. An ABS is set on the top of the train, and moves with the train. The traffic demands onboard can always achieve high throughput, because they are always included in the cell coverage of the ABS which is directly connected to a gateway through optical backhaul. In this scenario, the number of required BSs is independent of the length of railway and is drastically reduced compared with the static model.
2.2
(b) Proposed OBMN architecture
Figure 2: Demand satisfaction for railways.
System architecture
2.2.1 Overview. As shown in Fig. 1, an ABS is set on the top of the train. It is connected to a gateway node through ORB which consists of fiber optic networks. The physical topology of the ORB network can be point to point (P2P) or a star such as passive optical network (PON). The ABS forwards traffic to and from moving demand onboard using an arbitrary RAT in the same way as a static BS. The ABS functions are introduced in 2.2.2. As the train moves, the ABS follows the demand movement with the optical backhaul connected, The optical fiber length between the ABS and the gateway is adequately adjusted to follow the movement. The key ideas for realizing this concept are high-speed reels and fiber lanes, which are explained in 2.2.3 and 2.2.4. The optical fiber cable is held with fiber lanes and adjusted with high-speed reels. ABS with this architecture always covers the traffic demands onboard. The ABS can continuously forward traffic to the gateway via fiber optic networks during the movement. As a consequence, high-bandwidth and stable optical backhaul is always available for the moving demands.
technologies [11], e.g., sensing, navigation, and motion planning. Unlike the generic ABSORB, the required autonomous movement function for an ABS in the OBMN is limited to the functions related to getting on and off a train. This is because the movement to a new location is performed by the train which the ABS is set on. The autonomous getting on and off function can be implemented in two ways: self-movement and supported-movement. The self-movement represents that robot vehicle, i.e. drone, technologies are attached to a BS. A robot BS autonomously moves and rides on a train based on the schedule. A possible implementation of the supportedmovement is that a robot arm grabs a BS, moves it, and puts it on and off a train. Either type of BS is referred to as an ABS in the following description for simplicity. 2.2.3 High-speed reel. The key idea for deploying OBMN is a high-speed reel. The conceptual structure is depicted in Fig. 3a. It is placed along the railway on telecommunications facilities such as telegraph poles. The structure of a highspeed reel is similar to a fishing reel, and it spins to reel up and send out the optical fiber to adjust the length of fiber. When the connected ABS is approaching, it rotates to reel up the fiber. When the ABS is going away, it spins to send out the fiber. The basic rotation schedule is programmed corresponding to the ABS movement schedule by the controller. In
2.2.2 ABS. An ABS moves to a new location according to a schedule computed by a controller, which is a remote server connected via a control channel. The control channel can be established using ORB or macro-cell communication. The relocation is realized with autonomous robot vehicle
33
Session 4: Applications and Infrastructure
HotWireless’17, October 16, 2017, Snowbird, UT, USA
(a) High-speed reel
(b) Fiber lane
Figure 3: Architecture for OBMN components. addition, a high-speed reel should provide a sensing function to keep the tension of the fiber constant. 2.2.4 Fiber lane. Fig. 3b shows the fiber lanes and corresponding structure of an ABS. Fiber lanes are paired structure to hold the optical fiber cable. They are installed with a certain interval in parallel to the overhead contact wire of the railway. The track of cable in the train movement is controlled with fiber lanes. Without the fiber lanes, the high-speed reel and the ABS are linearly connected irrespective of the railway track. Thus, more number of fiber lanes are required for curves and the interval can be set longer in straight tracks. As shown in Fig. 3b, the fiber cable is outputted from the ABS via the upright. The ABS passes through fiber lanes using the hanger. When the hanger arrives, fiber lanes slide to pass it. It is essential to reduce the frictional resistance of fiber lanes to realize smooth movement of the hanger and cable. In addition, fiber lanes can help the fiber cable movement by spinning.
(a) Arrival
(b) Handover
(c) Departure
2.3
Transfer Figure 4: Sequence of transfer.
2.3.1 Sequence. Transfer is a concept for executing handover in the proposed OBMN architecture. This is one of the key technology for realizing OBMN, because the length of optical fiber is physically limited, and it cannot be infinitely extended due to the propagation delay. Transfer is performed at a certain point, which is called a transfer point. The sequence of transfer is the following.
3. Departure The train departs from the transfer point (Fig. 4c). It is possible to perform the transfer when the train stops or moves at a low-speed, and thus the transfer points are mainly assumed to be stations.
1. Arrival A train arrives at a transfer point. An ABS is set on the train (Fig. 4a). 2. Handover The ABS gets the train off, and another ABS is set on the same train and the demand is handovered to the newly set ABS (Fig. 4b).
2.3.2 Rideshare. Rideshare is an idea of transporting ABSs for their efficient use. This concept is depicted in Fig. 4. Surplus ABSs are loaded on a train at a transfer point, and unloaded at another transfer point. Accumulation of ABSs at a certain transfer point can be avoided by rideshare, even if
34
Session 4: Applications and Infrastructure
HotWireless’17, October 16, 2017, Snowbird, UT, USA
the train timetables are unsymmetrical, e.g. commuter trains between midtown and suburbs.
3
[10] Junhyeong Kim, Hee-Sang Chung, Il Gyu Kim, Hoon Lee, and Myong Sik Lee. 2015. A study on millimeter-wave beamforming for high-speed train communication. In International Conference on Information and Communication Technology Convergence (ICTC). IEEE, 1190–1193. [11] Tom´ as Lozano-Perez, Ingemar J Cox, and Gordon T Wilfong. 2012. Autonomous robot vehicles. Springer Science & Business Media. [12] NGMN Alliance. 2015. 5G white paper. Next generation mobile networks, white paper (2015). [13] Meng-Shiuan Pan, Tzu-Ming Lin, and Wen-Tsuen Chen. 2015. An enhanced handover scheme for mobile relays in LTE-A high-speed rail networks. IEEE Transactions on Vehicular Technology 64, 2 (2015), 743–756. [14] Agisilaos Papadogiannis, Michael Farber, Ahmed Saadani, Muhammad Danish Nisar, Petra Weitkemper, Thiago Martins de Moraes, Jacek Gora, Nicolas Cassiau, Dimitri Ktenas, Jaakko Vihriala, and others. 2014. Pass it on: advanced relaying concepts and challenges for networks beyond 4G. IEEE Vehicular Technology Magazine 9, 2 (2014), 29–37. [15] Moumita Patra, Rahul Thakur, and C Siva Ram Murthy. 2016. Improving Delay and Energy Efficiency of Vehicular Networks using Mobile Femto Access Points. IEEE Transactions on Vehicular Technology (2016). [16] Rand Raheem, Aboubaker Lasebae, Mahdi Aiash, and Jonathan Loo. 2013. From fixed to mobile femtocells in LTE systems: issues and challenges. In International Conference on Future Generation Communication Technology (FGCT). IEEE, 207– 212. [17] Rand Raheem, Aboubaker Lasebae, Mahdi Aiash, and Jonathan Loo. 2016. Performance Evaluation of Mobile Users Served by Fixed and Mobile Femtocells in LTE Networks. Journal of Networking Technology 7, 1 (2016), 17. [18] Rand Raheem, Aboubaker Lasebae, and Jonathan Loo. 2014. Performance evaluation of LTE network via using Fixed/Mobile Femtocells. In International Conference on Advanced Information Networking and Applications Workshops (WAINA). IEEE, 255–260. [19] Simon Scott, Jouko Leinonen, Pekka Pirinen, Jaakko Vihriala, Vinh Van Phan, and Matti Latva-aho. 2013. A cooperative moving relay node system deployment in a high speed train. In IEEE 77th Vehicular Technology Conference (VTC Spring). IEEE, 1–5. [20] Yutao Sui, Ismail Guvenc, and Tommy Svensson. 2014. On the deployment of moving networks in ultra-dense urban scenarios. In International Conference on 5G for Ubiquitous Connectivity (5GU). IEEE, 240–245. [21] Yutao Sui, Ismail Guvenc, and Tommy Svensson. 2015. Interference management for moving networks in ultra-dense urban scenarios. EURASIP Journal on Wireless Communications and Networking 2015, 1 (2015), 1–32. [22] Yutao Sui, Zhe Ren, Wanlu Sun, Tommy Svensson, and Peter Fertl. 2013. Performance study of fixed and moving relays for vehicular users with multi-cell handover under co-channel interference. In International Conference on Connected Vehicles and Expo (ICCVE). IEEE, 514–520. [23] Yutao Sui, Jaakko Vihriala, Agisilaos Papadogiannis, Mikael Sternad, Wei Yang, and Tommy Svensson. 2013. Moving cells: a promising solution to boost performance for vehicular users. IEEE Communications Magazine 51, 6 (2013), 62–68. [24] Cheng-Xiang Wang, Ammar Ghazal, Bo Ai, Yu Liu, and Pingzhi Fan. 2016. Channel measurements and models for high-speed train communication systems: a survey. IEEE Communications Surveys & Tutorials 18, 2 (2016), 974–987. [25] Jiangzhou Wang, Huiling Zhu, and Nathan J Gomes. 2012. Distributed antenna systems for mobile communications in high speed trains. IEEE Journal on Selected Areas in Communications 30, 4 (2012), 675–683. [26] Jingxian Wu and Pingzhi Fan. 2016. A survey on high mobility wireless communications: Challenges, opportunities and solutions. IEEE Access 4 (2016), 450–476. [27] Hiroto Yasuda, Akira Kishida, Jiyun Shen, Yoshifumi Morihiro, Yasufumi Morioka, Satoshi Suyama, Akira Yamada, Yukihiko Okumura, and Takahiro Asai. 2015. A study on moving cell in 5G cellular system. In IEEE 82nd Vehicular Technology Conference (VTC Fall). IEEE, 1–5.
CONTRIBUTIONS AND CHALLENGES
The contribution of the proposed OBMN is providing stable and high-bandwidth optical backhaul for the moving demands on local trains in urban and suburban areas, unlike any other moving cell architecture. Further advantages of the OBMN is reducing handovers and cell-edge users, because the small cells follow the moving demand and handovers only occur at transfer points. The novelty of the OBMN originates from the specialization for the local trains characteristics: fibers can be laid along overhead lines without interruptions, the movable range is limited from several kilometers to several tens of kilometers along railways, and the moving speed is up to around 120 km/hr. The main technical challenges for OBMN are constructing the high-speed reel and fiber lanes which can follow the train movement.
4
CONCLUSION
In this paper we proposed the concept of OBMN architecture for efficiently deploying moving cell for local trains. In the OBMN, an ABS which is connected to a gateway via optical backhaul is set on the top of a train. While the train moves, the ABS always forward traffic of moving demand onboard through high-bandwidth optical backhaul. The proposed architecture can efficiently provide high and stable bandwidth using small number of BSs, unlike existing moving cell architectures with wireless backhaul.
REFERENCES [1] Lin Chen, Ying Huang, Feng Xie, Yin Gao, Li Chu, Haigang He, Yunfeng Li, Feng Liang, and Yifei Yuan. 2013. Mobile relay in LTE-advanced systems. IEEE Communications Magazine 51, 11 (2013), 144–151. [2] Yangyang Chen, Philippe Martins, Laurent Decreusefond, Feng Yan, and Xavier Lagrange. 2014. Stochastic analysis of a cellular network with mobile relays. In IEEE Global Communications Conference (GLOBECOM). IEEE, 4758–4763. [3] Cisco. 2017. Cisco Visual Networking Index: Global Mobile Data Traffic Forecast Update, 2016–2021. (2017). [4] Mohamed F Feteiha, Mahmoud H Qutqut, and Hossam S Hassanein. 2014. Outage probability analysis of mobile small cells over LTE-A networks. In International Wireless Communications and Mobile Computing Conference (IWCMC). IEEE, 1045–1050. [5] Daniel T Fokum and Victor S Frost. 2010. A survey on methods for broadband internet access on trains. IEEE communications surveys & tutorials 12, 2 (2010), 171–185. [6] Ammar Ghazal, Cheng-Xiang Wang, Bo Ai, Dongfeng Yuan, and Harald Haas. 2015. A nonstationary wideband MIMO channel model for high-mobility intelligent transportation systems. IEEE Transactions on Intelligent Transportation Systems 16, 2 (2015), 885–897. [7] Fourat Haider, Cheng-Xiang Wang, Bo Ai, Harald Haas, and Erol Hepsaydir. 2016. Spectral/energy efficiency tradeoff of cellular systems with mobile femtocell deployment. IEEE Transactions on Vehicular Technology 65, 5 (2016), 3389–3400. [8] Insoo Hwang, Bongyong Song, and Samir S Soliman. 2013. A holistic view on hyper-dense heterogeneous and small cell networks. IEEE Communications Magazine 51, 6 (2013), 20–27. [9] Aymen Jaziri, Ridha Nasri, and Tijani Chahed. 2016. Offloading traffic hotspots using moving small cells. In IEEE International Conference on Communications (ICC). IEEE, 1–6.
35
