WEEK TWO

Summary of the week

In the second week of our project, we further focused on developing the core computational code for cellular network capacity, with an emphasis on channel allocation, traffic load analysis, and user capacity evaluation.


Methodology

In capacity planning for cellular networks, our calculations are divided into three main parts:

1. Channel Allocation

2. Traffic Calculation

3. User Capacity Calculation

Channel Allocation

First, we need to calculate the number of available channels per cluster, cell and sector, which is the basis for the entire cellular network planning.

1. Channels per Cluster

2. Channels per Cell

Since each cell within a cluster uses a different frequency:

3. Channels per Sector

In the case of sectorization, each cell can be further divided into multiple sectors, each using a different channel:



Traffic Calculation

Traffic is a key parameter for measuring the carrying capacity of a network. We use the Erlang B formula to calculate network capacity:

1. Traffic per User

This indicates the average Erlang traffic per user per hour.

2. Erlang B formula to calculate traffic volume

In our project, the Erlang B formula is used to calculate the relationship between network capacity and blocking probability.


The basic assumptions of the Erlang B formula are:

- A user initiates a call randomly;

- When all channels are busy, new calls are rejected outright and are not queued;

- We usually want the blocking probability to be below a certain level, say 2%.


 The Erlang B formula is as follows:

- B(E, C) denotes the blocking probability;

- E denotes the call volume in Erlang;

- C denotes the number of channels.

User Capacity Calculation

After calculating the traffic capacity, we can estimate the maximum number of users that the network can support.This calculation method is the same as that for Traffic per Cluster, ensuring that the network meets the requirements of the target blocking probability.Since each user will occupy a certain amount of traffic, we can get the total number of users by dividing the total available traffic by the single-user traffic:

Problems

Although our initial implementation appeared to be on track, we discovered that the Traffic per Cluster (Erlangs) calculation yielded results that deviated from theoretical expectations. This prompted us to conduct a thorough investigation, during which we also encountered several other issues, such as debugging errors, data inconsistencies, and algorithm efficiency concerns. 

To overcome these challenges, we engaged in collaborative troubleshooting, tested multiple approaches, and carefully reviewed and adjusted the underlying formulas. After multiple iterations of optimization and code refinement, we successfully produced a functional version of the program, marking a significant step toward accurately analyzing cellular network performance.







Comments

Post a Comment

Popular posts from this blog

WEEK ONE

Image
          Summary of the week In the first week’s group meeting, we discussed the division of tasks and the next steps for our project.  It was decided that Fanghao and Aljory would be responsible for writing the core computation code, while Helal and Huseen would focus on designing and implementing the graphical user interface (GUI).  We also established a weekly work plan, outlining the tasks and goals for each stage to ensure steady project progress. In the following weeks, we will provide regular updates and address any issues in a timely manner to keep the project on track. Background In the modern era of wireless communication, optimizing network capacity is crucial for efficient service delivery. Our project focuses on Capacity Dimensioning of Cellular Networks , where we develop a MATLAB GUI application to compute the capacity of a cellular network based on various input parameters. By leveraging key telecommunications principles such as freque...
Read more

WEEK THREE

Image
Summary of the week In the third week of our project, we made critical modifications to our code, focusing primarily on traffic calculations and user count estimation. Initially, we used multiplication in these calculations, but after reviewing our approach, we realized that this method was not only prone to confusion but also differed from the channel calculation approach.  To improve the clarity of our code and ensure consistency with the channel estimation method, we decided to switch to division. This adjustment not only enhanced the accuracy of our calculations but also made the overall code structure more intuitive and easier to understand. As a result, we updated the relevant equations to ensure a more accurate representation of network traffic distribution and user load. Additionally, we introduced a new input parameter: cell radius range, which plays a key role in our graph generation. By incorporating this input, we could analyze how network performance varies with differ...
Read more