Location Update Schemes

Location update schemes can be classified into static and dynamic. Presently, most of the schemes are static because of their simplicity. Static schemes define specific conditions, under which a UE should perform location update procedure; the conditions are dependent on the network configuration and a UE’s current movement, but independent on a UE’s movement history. Such static mechanisms require low computation and are easy to implement because of the lack of independent UE tracking.

A variety of static location update schemes are discussed below.

The always-update mechanism is the simplest location update scheme, performing a location update whenever the UE move into a new cell. The network always knows the location of every UE and requires no paging to locate the user when an incoming

3

call arrives.

The always-update scheme performs well for those UEs who have low mobility pattern and high call arrival rate. On the other hand, it performs quite poorly for those UEs who have high mobility pattern, which means they require many location updates.

Although the always-update scheme has not been used in practice, it forms the basis of many complex mobility management mechanisms, such as location area update and profile-based update.

The never-update scheme is the counterpart to always-update, which means never require location update, but may result in excessive paging cost if the network is large or the call arrival rate is high. Although these two strategies are the extremes case of mobility management, they are often combined to form a more efficient mobility management strategy, which is more suitable for the differences in UE and network characteristics.

In reporting center scheme, some cells are selected as reporting-centers; a UE updates its location only when it enters into a new reporting-center. When an incoming call comes, the vicinity of the last reported reporting-center is paged to locate the target UE. Selecting the reporting centers that provide good trade-off between paging and update operations is a complex task [2].

The topology of reporting cells may be bounded or unbounded, as shown in Figure 1-1. The unbounded scheme required a smaller number of reporting cells, and thus reduce the number of location updates. The strategy, however, needs to be accompanied with an intelligent paging scheme to determine the cells to be paged.

Figure 1-1 (a) Bounded reporting cell configuration (b) Unbounded reporting cell

4

configuration

The performance improvements with a reporting cell topology are somehow limited, not only because the mobility patterns of UEs are not considered, but also it is difficult to obtain an optimal assignment of the reporting cells. The selection of an optimal set of reporting cells is an NP-complete problem.

The location area topology is widely used to control the occurrence of location updates. Here, the network is partitioned into groups of cells. The scheme is similar to the always-update mechanism; UE only updates their location as they leave the current location area. Figure 1-2 shows a partition with four separate location areas.

Figure 1-2 Network partitioned into four location areas

If the network knows the UE’s current location area, the paging process pages the cells inside the location area. The location update process may occur periodically or on the crossings of location area boundary. The periodical location update process requires the UE inform the network its location at a regular time interval. The boundary crossing method updates the UE’s location whenever the UE move to another location area. However, this method has its weakness, particularly for those UE who are tend to move along the boundaries between two or more location areas.

5

Figure 1-3 Ping-pong effect

Figure 1-3 shows the ping-pong effect; this is the main weakness of location area scheme. We can see that the UE moves back and forth between the boundaries of two or more location areas. This may result in a high location update rate with low physical mobility.

We summarize the static location update scheme in table 1-1:

Table 1-1 Comparison between major static LU schemes LU Scheme Update cost Paging

cost

low high several cells high computational

overhead

Location area low high several cells ping-pong effect Dynamic location update schemes work on the principle that the UEs may have different movement patterns, and thus the location update schemes should treat them differently. Under threshold (time, movement or distance)-based schemes, location update occurs when the UE’s behavior is beyond the threshold value. A variety of dynamic location update schemes are discussed below.

The time-based strategy [3, 4] requires that the UEs to update their locations at constant time intervals. This time interval can be optimized per UE to reduce the redundant location update. The mechanism only requires the UE to maintain a timer, which is low computational overhead and easy to implement.

6

Figure 1-4 Time-based location update

Figure 1-4 illustrates the updates (U1, U2, and U3) which are performed at each time interval ∆t, regardless of individual movements. The time-based scheme however has a high degree of overhead when the UE has only moved a very small distance, or the UE has not moved at all.

The movement-based update scheme [5, 6] requires the UEs to update their locations after a given number of boundary-crossings to other cells in the network. The boundaries-crossing threshold can be optimized per UE, according to individual movement pattern and call arrival rate.

Figure 1-5 Movement-based location Update

Figure 1-5 illustrates a movement-based scheme, with a movement threshold of two. Thus, the UEs update its location whenever the UE crosses two boundaries

The distance-based scheme [7] requires the UE to perform update process whenever the UE has moved for a pre-defined distance from the cell where the UE last performed location update. The distance threshold can also be optimized per UE according to movement pattern and call arrival rate.

7

Figure 1-6 Distance-based location update

A distance-based scheme is showed in Figure 1-6. The update process occurs when the UE travels beyond a certain radius from the previous updated location. This scheme has the benefit of not requiring an update when the UE move back and forth within a small subset of cells, provided that these cells inside the distance threshold radius.

For profile-based scheme [8], the network maintains a profile for each UE, according to the previous movement pattern. The profile contains a list of most probable cells for the UE to visit. The UE updates its location only when moving to a cell which is not contained in the list.

We summarize the dynamic location update schemes in table 1-2:

Table 1-2 Comparison between major dynamic LU schemes LU Scheme Update cost Paging

cost

Location accuracy

Major drawback

Time based low high several cells unnecessary

updates by

Distance based low high several cells high

computational overhead on UE side

Profile based low high several cells May suffer from low predictability

8