School of Engineering
1 Version 1 Semester 1 – 2020
School of Engineering
2 Version 1 Semester 1 – 2020
School of Engineering
3 Version 1 Semester 1 – 2020
LTE RF Planning and Optimisation
1. Objective
The objective of the project is to:
1) Expand your knowledge on how to perform wireless access network planning for an indoor
mobile cellular wireless environment based on LTE
2) Understand the difference between noise-limited systems and interference-limited systems
3) Understand the concept of cell-association highest received power
4) Understand the need for interference coordination in a radio access network
2. Prerequisites
This project assumes that students understand the basic concepts of:
1) LTE
2) The components of LTE
3) Matlab basics
3. Equipment and Software
1) Citrix workspace for myDesktop
2) Matlab
3) Computers with Windows® or other operating systems
4) Internet access.
4. Background
We are going to expand on the indoor coverage design of an LTE/LTE-A service for a shopping
centre using several femtocells. We will consider the following scenario:
1. The LTE network performance when the femtocells are not coordinating resource
allocation, i.e. the femto-cells are interfering with each other
2. The LTE network performance when the femto-cells are coordinating their interference
using Inter-cell Interference Coordination (ICIC).
A. Femtocells
Industry and research have responded to the increasing data transmission demands from
smartphones, tablets, and similar devices. As the demand continues to increase, it becomes
increasingly difficult to satisfy this requirement, particularly in densely populated areas and
remote rural areas. An essential component of the LTE / LTE-A strategy for satisfying demand
is the use of femtocells (also known as small cells).
A femtocell is a low-power, short range, self-contained base station. Initially used to describe
consumer units intended for residential homes, the term has expanded to encompass higher
School of Engineering
4 Version 1 Semester 1 – 2020
capacity units for enterprise and busy shopping centres. Key attributes include IP backhaul,
self-optimization, low power consumption, and ease of deployment.
Femtocells are by far the most numerous types of small cells. The term small cell is an umbrella
term for low-powered radio access nodes that operate in licensed and unlicensed spectrum that
have a range of 10-200 m. These contrast with a typical mobile macrocell, which might have a
range of up to several tens of kilometres. Femtocells now outnumber macrocells, and the
proportion of femtocells in 4G networks is expected to rise.
Figure 1. Typical femtocell (simplified) architecture
Figure 1 shows the typical elements in a network that uses femtocells. The femtocell access
point is a small base station, much like a Wi-Fi hot spot base station, placed in a residential,
business, or public setting. It operates in the same frequency band and with the same protocols
as an ordinary cellular network base station. Thus, a 4G smartphone or tablet can connect
wirelessly with a 4G femtocell with no change. The femtocell connects to the Internet, typically
over a DSL, FTTx, or cable landline. Packetized traffic to and from the femtocell connects to
the cellular operator’s core packet network via a femtocell gateway.
B. Raytracing
Raytracing is a powerful method used to obtain accurate simulation of the physical layer
performance in wireless access systems. The calculations are made by shooting rays from the
transmitters and propagating them through the defined geometry. These rays interact with
geometrical features and make their way to receiver locations. Ray interactions include
reflections from feature faces, diffractions around feature edges, and transmissions through
features faces. At each receiver location, contributions from arriving ray paths are combined
and evaluated to determine predicted quantities such as, received power, interference measures,
School of Engineering
5 Version 1 Semester 1 – 2020
path loss, delay spread, direction of arrival, and the impulse response of the channel. Figure 2
shows the different possible ray interactions.
Figure 2. Different types of ray interactions with the geometry of the environment
The received signal at the device will be affected by buildings walls and terrain on the
propagation of electromagnetic waves and the propagation in indoor environments. The
locations of the transmitters and receivers affects signal strength and other physical layer
C. LTE Interference Coordination
LTE is designed to operate with a frequency reuse of one, implying that the same carrier
frequency can be used at neighbouring transmission points.
The second generation cellular system (known as GSM) required that cells should always use
frequencies that are different to those use in neighbouring cells in order to avoid interference.
In 2G the available spectrum was divided based on the number of cells inside a reuse-cluster.
Figure 3 illustrates the concept of frequency re-use clusters with a cluster size of 7. Cluster
sizes may vary.
Figure 3: GSM frequency reuse principle
School of Engineering
6 Version 1 Semester 1 – 2020
However, a major drawback when using the same radio resources in all cells is the possible
interference between cells, thus a careful coordination is required between eNBs. LTE included
explicit support for such coordination, in the 3GPP release-8 context referred to as inter-cell
interference coordination (ICIC). ICIC defines a set of messages that can be exchanged
between eNodeBs using the X2 interface (see Figure 4) controlling the interference between
cells of different eNodeBs.
Figure 4: A simplified LTE architecture showing the interconnection between eNodeBs
5. Simulation scenarios
We are going to construct a simulation environment that comprises of:
• The geometry of the environment (floor plan, doors, windows, etc…)
• 07 femtocells + Antenna
• Receivers (mobile phones)
The Geometry
We have the following floor plan representing a typical shopping mall with wall height of 5m
and the grid width is 10m, see Figure 5.
School of Engineering
7 Version 1 Semester 1 – 2020
Figure 5: Typical shopping mall floor plan
The UEs
The height of the UEs is 1.5m (this is the standard 3GPP simulation height). The moderate
spacing between UEs is about 8-10m as per Figure 6. Each UE has an “isotropic antenna” with
gain 0 dBi for the receiver.
Figure 6: 01 Femtocell with UEs in a shopping mall
The Femtocell
We will start by selecting a location to place a femtocell based on the floor plan (Figure 6),
then you should place an additional 6 femtocells as per Figure 7
Question 1: Select the best locations for the 7 femtocells. Explain your location selection.
School of Engineering
8 Version 1 Semester 1 – 2020
Figure 7: Coverage example with 7 femtocells
Question 2: Explain what the different colours mean. Why do the cells overlap in some
Processing received signal data using MATLAB
MATLAB is used for simulation and analysis. We will use MATLAB to perform postprocessing of the “walking test” results to interpret the network performance. The excel folder
“lab4P2Exel” contains “walking test” results around the shopping mall for 07 Femtocell. The
following Figure gives an example of a walking test result with respect to a given Femtocell
Figure 8: An example of a walking test result with 01 Femtocell
School of Engineering
9 Version 1 Semester 1 – 2020
Importing Data Into Matlab
In order to perform this analysis, we would need to import the measurements into Matlab. Note
that each transmitter will have a separate “.p2m” file, which includes the received parameter
power readings of each eNB.
Question 3: Your task now is to find a way to import the 7 “.p2m” files into Matlab and to
organize the received power into a matrix format according to the following example:
Suppose you have hundreds of driving test files to import, think of an automated method to
import the files (using a loop function).
You can use dir function in Matlab to read the file names within a certain folder. Also you
can use readtable function to read .p2m files.
Finding the Serving Cell
The next step is to find the serving cell, we can use a simple method by searching for the
highest received power and consider the corresponding eNB as the “serving-eNB”.
Using the “RXpower” table we can sort the elements in column-wise according to the
following Matlab function:
[RXpower_sorted, Ind]=sort(RXpower,2,’descend’);
Where RXpower_sorted is a 2D matrix with the first column representing the highest
received power.
“Ind” is also a 2D matrix describing the element arrangement, with the first column
representing the ID of the dominant eNB as per the following figure:
School of Engineering
10 Version 1 Semester 1 – 2020
Question 4: Plot the cell dominance according to your walking test results, the figure would
be like the following example figure:
Hint: An easy method to plot the cell dominance is use the scatter function in MATLAB
MarkerSize =100;
h=scatter(x,y, MarkerSize, Ind(:,1),’s’,’fill’);
Also do not forget to use the command
axis equal
such that both x-axis and y-axis will have the same scale
School of Engineering
11 Version 1 Semester 1 – 2020
You also need to use more advanced method to perform the plot using pcolor function.
Interference Coordination
We will observe the resulting SINR in two different cases:
• Case 1: when femto-eNBs are not coordinating their resource allocation (no
interference coordination)
• Case 2: when inter-cell interference coordination is enabled.
The SINR is given by:
𝐼 + 𝑁
Where S is the received signal of the “serving” eNB
I is the total interfering power and.
N is the noise power. The noise power, 𝑁 = 𝑘 𝑇 𝐵𝑛𝑜𝑖𝑠𝑒 𝐹
You can assume a fixed noise power across the receivers and calculate it based on the following
Noise temperature 𝑇 = 300𝑜 K
UE noise figure 𝐹 = 8 𝑑𝐵
UE noise bandwidth 𝐵𝑛𝑜𝑖𝑠𝑒 = 10 𝑀𝐻𝑧
UE data bandwidth 𝐵 = 1 𝑀𝐻𝑧
Case 1: No interference coordination between the eNBs
A poor situation can occasionally occur in practical 4G networks when neighbouring cells
(eNBs) are not “logically” defined as neighbours. In this case there will be no interference
coordination between these eNBs, and the eNBs will cause mutual degradation in the service.
In our simulation we will start with worst-case-scenario assuming that all eNBs are not
coordinating the interference. Thus, the interference power summed at a certain receiver can
be calculated from:
𝐼 = ∑ 𝐼𝑖
Where the interferers set is all eNB except the serving one
Question 5: Calculate the SINR at each receiving UE and plot the SINR as a heatmap. Refer
to the following figure as a reference. Add the locations of the eNB on the heatmap
• Add the interference using linear power. A common mistake is to sum the powers
using dBm
• Remember that you need to use the signal of the “serving” eNB
School of Engineering
12 Version 1 Semester 1 – 2020
Question 6: Plot the histogram of the SINR as received by the UEs. Discuss your results
Case 2: inter-cell interference coordination is enabled.
In the second case we assume that eNBs have ideal interference coordination, i.e. no
interference between eNBs. In this case the SINR will reduce back to the SNR formula:
Question 7: Calculate the SNR at each receiving UE and plot the SNR as a heatmap, use the
same colors as in the first case. Add the locations of the eNB on the heatmap. Refer to the
following figure as a reference:
Hint: To control the colorbar limits you can use the following Matlab function
caxis ([cmin,cmax]);
Question 8: Plot the histogram of the SINR as received by the UEs and compare with results
of case 1.