Increasing the LTE-Advanced Network Capacity Using Inter-band Carrier Aggregation (Downlink Side) Method

According to the identification of the Operating Support System (OSS) by the Smartfren cellular operator in the Central Bandung area, six sites are found to have high traffic capacity with the physical resource block (PRb) percentage of 82.6 %. The use of PRb > 80 % is included in the warning indicator 2 based on the operator’s standards. It is also strengthened by the condition of the existing sites with the average Reference Signal Receive Power (RSRP) of -103.3 dBm, Signal to Interference Noise Ratio (SINR) of 6.28 dB, and throughput of 27.78 Mbps, thus resulting in non-optimal network performance in the area. Therefore, in this study, the inter-band Carrier Aggregation (CA) was applied by combining the 40 Time Division Duplex (TDD) band (2300 MHz) and band 5 Frequency Division Duplex (FDD) (850 MHz). One of the advantages of applying this method is that it can increase the user network capacity by maximizing the resources owned by the operator. The predetermined scenario taking into account the initial network condition indicated a decrease in the PRb percentage by 44.50 % and an increase in the average RSRP value by 12.8 dBm, SINR by 5.14 dB, and throughput by 34.59 Mbps.


INTRODUCTION
Cellular service users tend to measure a network based on the speed of uploading and downloading data. The faster the upload and download, the better the service of the cellular operator. To provide high-speed data services, a wide bandwidth is necessary [1], [2]. This leads to the development of 4G LTE-Advanced cellular technology that provides very high data speeds with high capacity and high mobility. However, the use of wide bandwidth takes up the frequency spectrum, which is a problem for cellular operators who have limited frequency resources, and if the cellular operator has excess frequency resources, the use of Long-Term Evolution (LTE) networks can only use a maximum bandwidth of 20 MHz for LTE release 8 [3], [4]. This is the background of an international standardization body, the 3rd Generation Partnership Project (3GPP), to serve as the LTE network developer. In March 2008, the 3GPP began a study to further develop LTE towards LTE-Advanced by targeting the IMT-A requirements set by the International Telecommunication Union (ITU). The study results produce a new set of radio features such as carrier aggregation in relea se 10, which is a solution to the limitation of frequency use [5], [6].
Based on the results of the survey and identification of OSS conducted by a cellular operator in the Central Bandung area, there are six sites with high traffic capacity, i.e., the physical resource block percentage of 82.6 % with an average throughput of 27.78 Mbps, the RF parameter (RSRP) value of -103.3 dBm, and the SINR value of 6.28 dB. Two of the six sites are in a warning condition, including ZBDG_4393, with the physical resource block percentage of 88 % and ZBDG_4417 with the physical resource block percentage of 72.21 %. A percentage of use of physical resource block (PRb) > 70 % is included in the warning indicator 1, > 80 % is included in the warning indicator 2, and > 90 % is included in the warning indicator 3, so the LTE network performance in the Central Bandung area is not maximum. The high percentage of PRb usage is the background of the use of the inter-band CA method to increase the user capacity in the planning 53 Jurnal Infotel Vol. 12  area by maximizing the bandwidth owned by the cellular operator [7], [8].
In this study, the application of inter-band CA was done to increase network capacity by optimizing bandwidth usage owned by the cellular operator, which combined the Band 5 FDD (850 MHz) and the Band 40 TDD (2300 MHz). There are several parameters to be analyzed, namely RSRP, SINR, throughput, and the percentage of PRb usage [9], [10].
For a better understanding, the rest of this paper is organized as follows. Section II discusses the research methodology, while the results of the inter-band carrier aggregation application are discussed in section III. Lastly, the conclusion is presented in section IV.

A. Carrier Aggregation Method
Carrier aggregation is a technology that allows 4G networks to run on two different frequencies by combining several Component Carriers (CC) to achieve a peak data rate. The component carriers can have a bandwidth of 1.4, 3, 5, 10, 15, or 20 MHz, and a maximum of five CC can be combined with a maximum bandwidth of 100 MHz [11]. This CA is also able to maintain the compatibility between UE release 8 and UE release 9 or what is called "backward compatibility". The LTE release 8 ca n only be a maximum of 20 Mhz [12].
In the 3GPP release 8 specification, this criterion has not been fulfilled because it only reaches 300 Mbps, so LTE release 8 can be called 3.9 G, whereas in the 3GPP release 10 the LTE-A has reached a peak data rate of above 1 Gbps. The LTE-Advanced uses the CA, which can combine up to five CCs with each bandwidth reaching 20 MHz, depending on the spectrum availability and the EU capability.

B. Carrier Aggregation Spectrum Scenario
The CA can be used on both FDD and TDD technologies. The carrier aggregation arrangement can be implemented either on the same or different band and bandwidth frequencies. In genera l, CA has three different features, as shown in Fig. 1 Table 1 illustrates the CA bandwidth class that shows the combination of the maximum Aggregated Transmission Bandwidth Configuration (ATBC) and the maximum number of CC, where ATBC is the combined bandwidth configuration that is equal to the total PRb aggregate. The application of three CCs is already available; meanwhile, the four CCs is expected to subsequently available since it is currently in the status of For Further Study (FFS). Figure 2 indicates an area that is the object of research on the application of inter-band CA. Six sites have quite high traffic in the area and have a good potential market. The focus zone (green line) shows the boundary area for the application of inter-band CA, while the computation zone (red line) shows the area around the site coverage.

E. Carrier Aggregation Configuration
In this article, CA is implemented with CA configuration using a combination of the operating frequency band owned by the cellular operator, namely band 40 (2300 MHz) with a bandwidth of 30 MHz and band 5 (850 MHz) with a bandwidth of 10 MHz. This configuration affects the LTE network itself. Table 2 shows the CA configurations applied in this study. Table 2 shows the four CA configurations in two scenarios. Scenario 1 indicates the configuration of CA_40A-5A with 2 CCs and CA_40C-5A with three CCs. The scenario 1 uses a CA configuration based on frequency band 40 (2300 MHz) as the Primary CC (PCC) or main carrier and band 5 (850 MHz) as the Secondary CC (SCC) or second carrier. Then, scenario 2 indicates the CA_5A-40A configuration with two CCs and CA_5A_40C with three CCs. Scenario 2 uses CA configuration based on frequency band 5 (850 MHz) as the PCC or main carrier and band 40 (2300 MHz) as the SCC. The CC from CA_5A has one carrier with a total NRb <100 so that it belongs to CA Bandwidth Class A with a bandwidth of 10 MHz and a total NRb of 50. The CC from CA_40A has one carrier with a total NRb <100 so it belongs to CA Bandwidth Class A with a bandwidth of 20 MHz and a total NRb of 100. Meanwhile, the CC from CA_40C has two carriers with a total NRb between >100 to <200 so that it belongs to CA Bandwidth Class C with 10 MHz and 20 MHz bandwidths and a total NRb of 150. The distribution of the CA Bandwidth Class can be seen in Table 1.
In this inter-band CA, different resources between TDD Frame and FDD Frame are important things to consider for the CA application. The cellular operator can allocate resources on request where the TDD resources are asymmetrical. It is an advantage for the application of inter-band CA, which can allocate uplink (UL) and downlink (DL) resources according to the operator. In this study, the Resource Block (RB) allocation is based on duplex and bandwidth, as shown in Table 3. The value used in this process is the fixed value of the smartfren operator, depending on the use of bandwidth for the calculation of UL & DL PRb. FDD is 1:1 symmetrical between DL & UL, while asymmetric TDD depends on the configuration it uses [5]. Inter-Band CA TDD -FDD CA_40C-5A 40 3 Inter-Band CA FDD -TDD CA_5A-40A 30 2 Inter-Band CA FDD -TDD CA_5A-40C 40 3 Next, a resource block calculation is performed for the application of inter-band CA, which can be seen in Table 4.

F. Coverage Dimensioning
At this stage, the calculation of the link budget and cell radius is done. Table 5 shows the Maximum Allowed Path Loss (MAPL) value based on the morphology of urban areas with propagation modeling, i.e., the Cost-231 Model. We use this empirical model because it has frequency specifications from 800 MHz to 2000 MHz. The next step is to calculate the cell radius, which is determined by the value of the link budget's results, i.e., the smallest MAPL between the downlink and uplink. The calculation results can be seen in Table 6. In this calculation, it is known that the antenna height is based on the existing conditions assuming the user height is 1.5 meters, and the working frequency is 2300 MHz. The calculation employs (1) and (2).
Where PL is a pathloss (dB), hb is a height of eNode B (m), a(hm) is an UE antenna correction factor, and d is a radius (km).

a(h m )urban=(1.1 log f -0.7)h m -(1.56 log f -0.8) (2)
Where PL is a pathloss (dB), f is a frequency (MHz), hb is a height of eNode B (m), hm is a height of UE (m), d is a radius (km), and a(hm)urban is an UE antenna correction factor for urban areas.
The next step is to calculate the forecasting user. The calculation employs (3).

Future Population = po [(1+GF)] n (3)
Where Po is a current population, GF is a grow factor, and n is a number of forcasting years.
The next process is to calculate the forecasting user number of LTE-Advanced operator with the following steps. The calculation results can be seen in the following Table 7.

III. RESULT
The inter-band CA simulation was done by taking into account: the initial network condition and two predefined scenarios. First scenario: performed according to the configuration schemes of CA_40A-5A with two CCs and CA_40C-5A with three CCs. The second scenario: performed according to the configuration schemes of CA_5A-40A with two CCs and CA_5A-40C with three CCs.
The simulation is based on a research area limited by the focus zone (green line) and the computation zone (red line), which is the site coverage area around the CA application. After that, the evaluation is performed by taking into account the RSRP parameters, SINR, throughput, and the percentage of PRb based on cellular provider standards. Fig. 4 -9 indicates the results of predictions of the application of inter-band CA from both scenarios. Table 10 illustrates the summary of planning results by comparing the initial conditions of the network using scenarios 1 and 2. Table 10 shows the comparison of the forecasting results of the initial network condition, scenario 1, and scenario 2. The optimal result is found in scenario 1 with the inter-band CA configuration of CA_40C_5A with 3 CCs.  Table 11 shows a decreased traffic load percentage of the physical resource block usage in the application of inter-band CA compared to the initial network condition that has high traffic.  Table 11 shows the percentage of PRb usage where the testing is performed under the same traffic load (DL) condition. In the initial column, (%) shows a high percentage of PRb at the initial network condition, while in the final column, (%) shows a lower percentage of PRb after application of the inter-band CA method. The application of inter-band CA can maintain high traffic to remain stable under the cellular operator's warning with an increase in the network capacity side. Table 12 shows the percentage of each target parameter according to the operator's Key Performance Indicator (KPI) standard. Table 12 shows the final simulation results of the inter-band carrier aggregation application. It indicates an increase in the percentage of KPI targets in the focus zone and an increase in the computation zone. An optimal forecasting result can be found in scenario 1 with an inter-band CA configuration of CA_40C_5A with 3 CCs.
Based on the results of capacity increase simulations, this study recommends the applica tion of inter-band CA using CA_40C_5A configuration with three CCs to be implemented by the operator in the Central Bandung area. This is due to the consideration of initial network condition and efficiency in the use of operator resources, as well as better performance in terms of coverage with an average RSRP of -90.5 dBm, better network quality with an average SINR of 11.42 dB. In addition, in the term of capacity, it increases with an average throughput of 62.37 Mbps and at a percentage of PRb of 44.50 %.

V. CONCLUSION
According to the simulation results, scenario 1 with the CA-40A-5A configuration with two CCs can increase the average throughput capacity of 38.12 Mbps with the traffic load percentage of PRb usage of 72.80 %. This value improves the traffic percentage of high PRb usage in the initial network condition, which is 82.6 %, and improves the quality of RF parameters such as RSRP with an average of -90.5 dBm, and SINR with an average of 13.12 dB. Meanwhile, the use of 3 CCs increases the average capacity more on the throughput wherein scenario 1 with CA-40C-5A configuration, it can be more optimally obtained with an average throughput of 62.37 Mbps with the load traffic percentage of PRb usage of 44.50 %. This value improves the percentage of high PRb usage traffic by 82.6 % at the initial network condition, that at SINR by 11.42 dB, and that at RSRP by -90.5 dBm.
The inter-band CA application of CA-40C_5A configuration with 40 MHz bandwidth is more optimal to be implemented because it can increase capacity by maximizing available resources.