On Improving Data Rates of Users in LTE HetNets

Vanlin, S and Tamma, Bheemarjuna Reddy (2016) On Improving Data Rates of Users in LTE HetNets. PhD thesis, Indian institute of technology Hyderabad.

[img]
Preview
Text
Thesis_Phd_CS_5395.pdf - Submitted Version

Download (5MB) | Preview

Abstract

The proliferation of smartphones and tablets has led to huge demand for data services over cellular networks. Cisco VNI mobile forecast (2014-2019) tells that although only 3.9% of mobile connections were Long Term Evolution (LTE) based they accounted for 40% of the mobile traffic and this will rise to 51% by 2019, by which the mobile data usage will grow 11 fold to over 15 Exabytes per month. Reports by Cisco and Huawei tell that 70% of the traffic is generated in indoor environments such as homes, enterprise buildings and hotspots. Hence, it is very important for mobile operators to improve coverage and capacity of indoor environments. Indoor data demand is partly met by intensifying the deployment of Macro Base Stations (MBSs/eNodeBs) in LTE cellular networks. Owing to many obstacles in the communication path between MBS and users inside the building, radio signals attenuate at a faster rate as the distance increases. Thus, Indoor User Equipments (IUEs) receive still low signal strength ( i.e., Signal-to-Noise Ratio, SNR) compared to Outdoor User Equipments (OUEs). To address this problem, one can deploy a large number of Low Power Nodes (LPNs) a.k.a. small cells (e.g., Picos and Femtos) under an umbrella MBS coverage and thereby form an LTE Heterogeneous Network (HetNet). Small cells are mainly being deployed in homes, enterprise buildings and hotspots like shopping malls and airports to improve indoor coverage and data rates. This is a win-win situation as telecom operators also benefit by reduction in their CAPEX and OPEX. Though the deployment of Femtocells improves indoor data rates, the resulting LTE HetNet may face a host of problems like co-tier and cross-tier interference (due to frequency reuse one in LTE) and frequent handovers (due to short coverage areas of Femtocells). Deployment of Femtos inside a building can lead to signal leakage at the edges/corners of the buildings. This causes cross-tier interference and degrades the performance of OUEs in High Interference Zone (HIZone) around the building area, which are connected to one of the MBSs in the LTE HetNet. Arbitrary placement of Femtos can lead to high co-channel cross-tier interference among Femtos and Macro BSs and coverage holes inside buildings. If Femtos are placed without power control, this leads to high power consumption and high inter-cell interference in large scale deployments. Our goal is to address these problems by developing efficient architecture, Femto placement and power control schemes in LTE HetNets. Random or unplanned placement of the Femtos leads to poor SNR and hence affects achievable data rates of IUEs. Hence, placement of Femtos is important for the cellular operators to perform planned deployment of minimum number of Femtos with no coverage holes and guarantee a good signal quality with no co-tier interference. Once the placement of Femtos is done optimally in enterprise environments, operators need to ensure that traffic load is evenly distributed among neighboring Femtos for improving Quality of Service (QoS) of IUEs by efficiently utilizing the network resources. In traditional cellular networks, the uplink access and downlink access of UEs are coupled to the same (Femto) cell. Suppose a Femto is fully loaded when compared to its neighboring Femtos, the traditional offloading or load balancing algorithms will try offloading some of the UEs for both their uplink and downlink access from the loaded cell to one of less loaded neighboring cells (i.e., target cell) provided that these UEs could get connected to the chosen target cell. This type of offloading is a forced handover to reduce traffic imbalance and trigger for handover is not based on better signal strength from the target cell. But, the offloaded UEs are connected for both their uplink and downlink access to the same target cell. Since UEs are most likely separated by walls and floors from their connected cells in enterprise environments, these offloaded UEs now have to transmit with higher transmit power in the uplink and thereby affects their battery lives. In order to reduce the battery drain for the offloaded UEs while maintaining their QoS, we employ the Decoupled Uplink and Downlink (DUD) access method in such a way that, the uplink of UE is connected to the closest Femto while the downlink is connected to a less loaded neighboring Femto. To maximize the utilization of the limited operating spectrum and provide higher data rate for IUEs, operators can configure Femtos in open access mode with frequency reuse one (i.e., all Femtos and MBSs operates on a same frequency) in LTE HetNets. However, this leads to high co-tier interference and cross-tier interference. Another problem in enterprise buildings having Femtos is frequent handovers, that happens when IUEs move from one room/floor to another room/floor inside the building. This leads to degradation of network performance in terms of increased signaling overhead and low throughputs. In order to reduce this kind of unnecessary handovers in enterprise buildings, Femtos should be placed optimally with handover constraints. Hence, we obtain the optimal coordinates from the OptHO model by adding handover constraints to the Minimize Number of Femtos (MinNF) model which guarantees threshold Signal-to-Interference plus Noise Ratio (SINR) of -2 dB for all IUEs inside the building. Such optimized deployment of Femtos reduces the number of handovers while guaranteeing good SINR to all IUEs. In LTE HetNets, even though planned deployment of Femtos in open access mode boosts the IUEs performance, the power leakage from indoor Femtos create interferix ence to the OUEs in the HIZone in the buildings surrounding areas. We propose an efficient placement and power control SON (Self organizing Network) algorithm which optimally places Femtos and dynamically adjusts the transmit power of Femtos based on the occupancy of Macro connected OUEs in the HIZone. To do this, we use the same MinNF model to place the Femtos optimally and solve Optimal Femto Power (OptFP) allocation problem (Mixed Integer Linear Programming (MILP)) which guarantees threshold SINR of -4 dB for IUEs with the Macro users SINR degradation as lesser than 2 dB. In the OptFP model, Femto’s transmit power is tuned dynamically according to the occupancy of OUEs in the HIZone. But the presence of even a single OUE in the HIZone decreases SINR of numerous IUEs, which is not fair to IUEs. In order to address this issue, we propose two solutions a) On improving SINR in LTE HetNets with D2D relays and b) A novel resource allocation and power control mechanism for Hybrid Access Femtos in LTE HetNets, which we describe in the following two paragraphs. To guarantee certain minimum SINR and fairness to both IUEs and OUEs in HIZone, we consider a system model by applying the concept of Device-to-Device (D2D) communication wherein free/idle IUEs connected to Femto act like UE-relays (i.e., UE-like BS, forwarding downlink data plane traffic for some of the HIZone users connected to MBS). We formulate a Mixed-Integer Linear Programming (MILP) optimization model which efficiently establishes D2D pairs between free/idle celledge IUEs and HIZone users by guaranteeing certain SINRT h for both IUEs and HIZone users. As D2D MILP model takes more computation time, it is not usable in real-world scenarios for establishing D2D pairs on the fly. Hence, we propose a two-step D2D heuristic algorithm for establishing D2D pairs. In above works, we assume that Femtos are configured in open access mode. But Hybrid Access Femtocells (HAFs) are favored by the operators because they ensure the paid Subscribed Group (SG) users certain QoS and then try to maximize the system capacity by serving near-by Non Subscribed Group (NSG) users in a best-effort manner. To reap in the benefits of HAFs, the operators need to employ effective resource sharing and scheduling mechanisms to contain co-tier and cross-tier interference arising out of reuse one in the HetNet system. Towards this, we address various challenges in terms of deployment and operation of HAFs in indoor environments. We propose an Optimal Placement of hybrid access Femtos (OPF) model which ensures a certain SINRT h inside the building and a certain SINRT h in the HIZone of the building. Unlike in previous optimization models, in this model, users in HIZone are connected to HAF s deployed inside the building. Also we propose a decentralized Dynamic Bandwidth Allocation (BWA) mechanism which divides the available HAF bandwidth between the two sets of user groups: SG and NSG. In order to mitigate co-tier and cross-tier interference, we then propose a dynamic Optimal Power Control (OPC) mechanism which adjusts the transmit powers of HAFs whenever the users in the HIZone cannot be served by the HAFs. In such a case, HIZone users connect to an MBS instead. Since the OPC problem is hard to solve in polynomial time, we also present a Sub-Optimal Power Control (SOPC) mechanism. To maintain fair resource allocation between SG and NSG users, we propose an Enhanced Priority (EP) scheduling mechanism which employs two schedulers which are based on the Proportional Fair (PF) and the Priority Set (PS) scheduling mechanisms. In above works, placement of Femtos is optimized to reduce co-channel co-tier interference among neighboring Femtos and transmit power of Femtos is optimized to reduce cross-tier interference between MBSs and Femtos. But, for arbitrary deployed Femtos, Inter Cell Interference Coordination (ICIC) techniques could be employed to address co-tier interference problem among Femtos which are connected with each other over X2 interface. Hence, in this work, we propose an ICIC technique, Variable Radius (VR) algorithm which dynamically increases or decreases the cell edge/non-cell edge regions of Femtos and efficiently allocates radio resources among cell edge/non-cell edge regions of Femtos so that the interference between neighboring Femtos can be avoided. We implement the proposed VR algorithm on top of PF scheduler in NS-3 simulator and find that it significantly improves average network throughput when compared to existing techniques in the literature.

[error in script]
IITH Creators:
IITH CreatorsORCiD
Tamma, Bheemarjuna ReddyUNSPECIFIED
Item Type: Thesis (PhD)
Subjects: Computer science
Divisions: Department of Computer Science & Engineering
Depositing User: Team Library
Date Deposited: 31 May 2019 04:07
Last Modified: 31 May 2019 04:07
URI: http://raiithold.iith.ac.in/id/eprint/5395
Publisher URL:
Related URLs:

Actions (login required)

View Item View Item
Statistics for RAIITH ePrint 5395 Statistics for this ePrint Item