In this chapter, considering realistic constraints, we have provided an MILP model for the optimal placement of Femtos based on user occupant probabilities inside an enterprise building scenario to achieve desirable signal strength for all the users. We established DuD connections based on the shortest-path loss Femto for the uplink
access and a less loaded neighboring Femtos for the downlink access. We conducted extensive experiments in MATLAB based LTE system simulator to demonstrate the benefits of proposed optimal placement model. On average we observed 70% energy savings in decoupled access system when compared to the traditional coupled access system.
Chapter 3
Handover and SINR Optimized
deployment of LTE Femtocells
3.1
Introduction
In the previous chapter, we proposed an optimal Femto placement model which did not consider the impact of co-tier and cross-tier interference. Also we assumed the users are static in nature. In this chapter, we consider co-tier interference among neighboring Femtos but no cross-tier interference (i.e., Macro and Femtos are assumed operating in different frequency bands) and formulate two ILP optimization models for Femto placement: first, Minimize the Number of Femtos (MinNF) model and second, Optimal Handover (OptHO) model. MinNF model guarantees a certain minimum SINR for each region inside the building while minimizing the number of Femtos needed for coverage of the entire enterprise building.
A major issue in enterprise building is frequent or unnecessary handovers (i.e., ping pong effect [49]), which may happen when user moves from one room to another room or within the same room and similarly in the corridors of the building. This leads to degradation of performance like service interruption during signaling overhead [50], decrease in throughput and increase in number of handovers [51]. Hence, in order to avoid the unnecessary handovers (i.e., handovers happening within the same room or in corridors) in enterprise buildings, Femtos should be placed by considering an additional constraint. Hence, we add Handover (HO) constraint to the MinNF model which reduces the number of handovers and at the same time guarantees good SINR to all UEs inside the building. This MinNF model along with HO constraint is referred as OptHO model in this chapter.
3.1.1
Organization of this chapter
Rest of the chapter has been organized as follows. Section 3.2 describes the notable research literature relevant to our study. Section 3.3 discusses the system model and proposed ILP formulation for MinNF. Section 3.4 presents experimental setup and numerical results for MinNF. Section 3.5 explains the user mobility model and proposed ILP formulation for OptHO. Section 3.6 demonstrates the experimental setup and numerical results for OptHO. Finally, Section 3.7 summarizes the work.
3.2
Related Work
Considering MBS and Femto-to-Femto interference plays a key role in small cell plan- ning, in [52], Femtos are placed in a multi-storey enterprise building by not consider- ing co-tier and cross-tier interferences. Authors of [34,53], considered the interference among MBSs to achieve better throughput in the system. In [32], the authors investi- gated a joint Femto placement and power control optimization problem in enterprise buildings with the aim to prolong UEs’ battery life. They proposed a novel two-step reformulation approach to convert the original Mixed Integer Non-Convex Problem (MINCP) into a MILP and then devised a global optimization algorithm by utilizing the MILP. But their system model did not consider co-tier and cross-tier interferences. In [53], the locations of the pico cells are moved iteratively. This way they reduce the interference to maximize the network throughput of the users in outdoor and indoor environments. However, in that work the placement of pico cells in indoor environ- ments and the traffic pattern are not considered. In [34, 35], Guo et. al. suggested an automated small cell deployment model which attempts to find the optimal location of a new cell, subject to knowledge about the locations of existing cells, UEs and the building environment. A closed-form equation is given for the new cell’s deployment location which is a function of transmit power, transmission scheme and path loss parameters.
In [54], the transmission power of randomly placed Femtos is optimized. This is to avoid interference and guarantee certain minimum SINR threshold for indoor UEs. But as the available bandwidth gets split into three parts (i.e., because the reuse factor is three), there could be inefficient usage of the spectrum. And because Femtos are placed randomly, the number of Femtos to obtain threshold SINR may not be minimal. Path loss factors such as obstructions (walls) inside the building are also not considered in the model while solving the optimization problem.
In [55], a joint Femto placement and power control model was proposed for guaran- teeing SINR threshold depending on varying user occupancy in each region inside the building. The efficiency of the spectrum usage is improved by considering reuse factor one and the number of Femto needed for enterprise deployments also got reduced, which would reduce the overall cost. User density, interference among Femtos and MBSs and building obstructions are also given as inputs to the system model. Then authors studied the joint optimal placement and power control of Femtos to maintain a certain minimum downlink SIN RT h. The joint placement and power control will
help to reduce the energy/power by Femtos and minimize the Femto count. However, doing so decreases SINR of indoors when compared to the case wherein Femtos are transmitting at the maximum power.
Large scale deployment of Femtos in enterprise environments could lead to unnec- essary handovers, [26, 51] which decrease the network throughput [56], increase signal overhead [50, 57] and cause delay [56]. Our current literature survey includes various unnecessary handover reducing mechanisms [58–60]. The authors in [58], proposed an optimized handover algorithm based on UE mobility state for reducing unnecessary handovers and also to improve the performance of LTE femtocell network. The au- thors in [59], proposed a dynamic handover hysteresis margin calculation based on the UE position within cell coverage region. In [60], the authors proposed a Fuzzy Logic Controller (FLC) that adaptively modifies handover hysteresis margins for reducing handovers. The authors in [50] proposed a simplified handover algorithm based on UE mobility state for reducing unnecessary handovers and signaling overhead in two-tier LTE femtocell network. In [57], a mobility management scheme is proposed where the control point of mobility in user plane is shifted from the S-GW to the Femto-GW so that it will make the handover decision between Femtos. In high density of Femtos and two-tier LTE HeNB network, UE mobility will produce lot of signaling overhead. In [56], a SON [61] model was proposed to mitigate unnecessary handovers inside enterprise environments with the help of building information and estimated user position information. In [62], the handover decision at FBS is based on energy effi- ciency and knowledge of interference. Thus handover plays a major role in enterprise buildings.
3.2.1
Contribution of this Chapter
Suggesting power saving by Femtos is not an appreciable contribution to the research community. Moreover, it also leads to decrease in average indoor SINR. We place
the Femtos optimally with all the constraints as mentioned in the system model (explained later) for MinNF and the Femtos are transmitting at full power to boost the average indoor SINR. We extend this MinNF model to bring in HO constraint thereby minimizing unnecessary handovers and boosting the throughput of the users in the building even while they are moving. To our knowledge, this is the first work that endeavors at reducing the number of handovers by placing the HO constraint as one of the input factors for the problem of placement of Femtos.