When cellular service providers build their networks, their networks are designed to provide coverage to the area of desire with the expectation of possible increase in population in the near future. For example, a company may design a cellular network to cover a city of area 1000 km2 with population of 1,000,000 people today assuming that 15% of the population will subscribe to their cellular service, or 150,000 people.

• However, to accommodate possible increase in the percentage of subscribers or the same percentage of subscribers but an increase in population, the network designer may build the network to provide acceptable GOS for 200,000. Such move guarantees that the network will need any expansion for possibly 5 years.
• In some cases, it may be difficult to predict the need for network expansion or even when network expansion is predictable, the time for network expansion arrives. There are several techniques to expand an already existing network or to add more capacity to a network being built. In the following we discuss two techniques.

Cell Splitting

• We have seen that reducing the size of cells of a cellular system keeps the SIR constant but results in an expansion of the network capacity because the smaller cells cover less area and therefore more cells would be required to cover the whole region which directly reflects on the network capacity.
• If the network is already functioning, it may be found that the network needs expansion only in specific regions and not network‐wide expansion. In this case, a cell (or multiple cells) can be split into smaller cells and frequencies are redistributed in a way that does not cause additional interference. This is shown in the following figures.
• The first figure shows a cell that has reached it capacity and needs to be split. This cell is split into several cells. Since the area of a cell is proportional to R2. So, reducing the cell readius to one half of its original value, for example, the area of the cell drops to one quarter of its original value.
• Therefore, theoretically, 4 of the smaller cells can fit into 1 of the large cells. However, since it is not possible to fit 4 quarter‐size hexagonal cells completely into 1 full‐size hexagonal cell, some regions will have to be covered by adjacent cells.

IMAGE 1

Original Cell Distribution

IMAGE 2

Cell Distribution following the splitting of cell

Sectoring

• The sectoring technique increases the capacity via a different strategy. In this method, a cell has the same coverage space but instead of using a single omni‐directional antenna that transmits in all directions, either 3 or 6 directional antennas are used such that each of these antennas provides coverage to a sector of the hexagon.
• When 3 directional antennas are used, 120° sectoring is achieved (each antenna covers 120°), and when 6 directional antennas are used, 60° sectoring is achieved (each antenna covers 60°).

IMAGE 3

• Dividing the cells into sectors actually reduces the network capacity because the channels allocated to a cell are now divided among the different sectors. In fact, handoff takes place when a cell phone moves from one sector to another in the same cell.
• The gain in network capacity is achieved by reducing the number of interfering co‐channel cells. If sectoring is done in a way that channels assigned to a particular sector are always at the same direction in the different cells (i.e., group A of channels is assigned to the sector to the left of the tower in all cells, and group B of channels is assigned to the sector at the top of all cells, and so on), each sector causes interference to the cells that are in its transmission angle only.
• Unlike the case of no sectoring where 6 interfering co‐channel cells from the first‐tier co‐channels cells cause interference, with 120° sectoring, 2 or 3 co‐channel cells cause interference and with 60° sectoring, 1 or 2 co‐channel cells cause interference. The number of cannel interfering cells depends on the cluster shape and size.
• By having less than 6 interfering first‐tier co‐channel cells causing interference, the SIR is increased for the same cluster size. This allows us to reduce the cluster size and achieve the same original SIR, which directly increases the network capacity.