If a user is moving among base stations (eNBs), handover
is performed to maintain user's connection and its Quality of
Service (QoS). To determine a proper time instant of handover
initiation, signal quality of a serving cell and all cells in user's
neighborhood is measured by a User Equipment (UE) and
reported back to the network. Cells considered for scanning are
included in so called Neighbor Cell List (NCL) [1]. The NCL
is defined for each cell after its deployment to the network.
Once the NCL of the new cell is defined, NCLs of all
neighbors must be updated as well. Besides the definition of
the initial NCL for the new cells, the NCL’s content should be
optimized as large amount of cells in the NCL leads to an
increase in overhead. This is due to the fact that the NCL must
be delivered to all UEs to keep them aware of the
neighborhood for scanning. Furthermore, scanning itself
reduces a time for data transmission and more cells in the NCL
prolong a scanning time [2][3]. Contrary, if the NCL is
incomplete, i.e., some real target cells are missing, the
handover cannot be performed properly and a connection of the
UE can be dropped.
An optimization of the NCL size can be done either
manually based on an estimation of a signal propagation (see,
e.g., [4][5]) or dynamically using knowledge of real network
parameters (e.g., [6]-[10]). The first approach utilized, for
example, in GSM networks, is very inefficient in terms of
accuracy and cost of manpower required for such optimization.
Therefore, we focus on the second way of the NCL
optimization, which exploits real network parameters. Most of
the work in the area of NCL optimization focuses on networks
with macrocells only (see e.g., [1][2][6]-[11])]. However, in
the future mobile networks, femtocells (in 3GPP denoted as
Home eNodeBs – HeNBs) are considered as a very important
part of the network architecture [12]. The problem of
employing proposals focusing only on macrocells to the
environment with HeNBs consists in large NCLs. This is
notable especially if HeNBs are deployed densely [3].
In this paper, we propose algorithm for a self-optimizing
NCL in networks with HeNBs. We focus on reduction of the
number of scanned cells after an initial NCL is defined. For the
initial NCL built-up, we consider sensing with hidden node
discovery algorithm defined in [13]. An innovation of our
approach consists in further reduction of the NCL by exploiting
knowledge of user's behavior and by dynamic adaptation of a
threshold for scanning according to the quality of the signal
received from the serving cell and interference.
The rest of the paper is organized as follows. The next
section gives an overview on the state of the art in the area of
NCL optimization. Then, our proposed approach is introduced
in Section III. Simulation models, including new mobility
model, and scenario are described in Section IV. Simulation
results are presented and discussed in Section V. Last section
summarizes major conclusions and defines potential future
improvements of our work.