Patent Application
20060068813
The present invention relates generally to mobile communication systems, and more particularly to a method of paging a mobile unit in a mobile communication system.
In mobile or wireless communication systems, mobile units have the capability to roam within the wireless communication system. When the wireless communication system receives a call request for a mobile unit, the communication system has to locate the mobile unit in order to complete the call.
A mobile unit registers with the wireless communication system when powering up, and also when placing a call. It would be desirable to have the communication system know the location of all mobile units at all times, but this would require each mobile unit to send a location update each time it moved into a new cell. This would be extremely inefficient, since the communication system only need to know the location of the mobile unit when an incoming call request is received, and also because the number of update messages would consume bandwidth that is needed by the communication system to handle calls within the communication system.
The process of paging a mobile consumes many resources within the MSC. These resources include bandwidth over the paging channel, processor occupancy of internal network elements, and the bandwidth for signaling messages between those elements. Paging is costly because the page is generally broadcast to many cells which might contain the mobile unit. If too few cells receive the page, one risks missing the mobile unit. If too many cells receive each page, resource shortages may ensue.
In current wireless communication systems, an MSC (Mobile Switching Center) pages a mobile unit in suspected cells upon receiving a call request for the mobile unit. Paging algorithms typically start by broadcasting to a narrow group of cells, and then broadcast to progressively larger groups in hope of eventually finding the mobile unit. Ideally, one would start with a very small number of cells, perhaps just a single cell, and then progressively widen the scope. Unfortunately, delivering a call to a mobile unit is time-critical. The calling party is likely to abandon the call before the mobile unit is found if too many paging attempts are necessary.
The present invention provides a method for determining a mobility index for a mobile unit in a wireless communication system. The mobility index is used to determine an efficient paging strategy for locating a mobile unit.
Often times a mobile unit is not mobile, but stationary. A typical subscriber of a mobile unit is mobile during certain times of the day, such as when driving to and from work, but is stationary during other times of the day, such as when at work or at home. The present invention utilizes a method for determining the best method of paging a mobile unit, based on the likelihood that the mobile unit is in a particular cell or other interaction area.
The present invention provides a method for paging a mobile unit by an MSC. The MSC calculates the probability that a mobile unit is in the same cell or group of cells as when it last interacted with the MSC based on a mobility index associated with that interaction and the elapsed time since that interaction. The MSC can therefore avoid wasting time on paging strategies with a low probability of success.
The mobility index is preferably calculated based on the last interaction of the mobile unit with the MSC. The mobility indices are preferably calculated for every sector of every cell, and also for the MSC as a whole. The mobility indices can also be calculated for any well-defined interaction area. The MSC can calculate the mobility index for being within the same cell and its immediate neighbors, for being within the same sub-MSC paging area, or for being within the same MSC. In further exemplary embodiments, the MSC can calculate the mobility index for different times of day and for different subscribers.
To calculate the mobility index, the MSC determines the last interaction of the mobile unit with the MSC. The MSC determines the cell and sector of the mobile unit for the last interaction. In an exemplary embodiment, the mobility index is a function of the cell sector. A cell sector serving a busy highway implies high mobility, while a cell sector serving a shopping mall, for example, implies low mobility. The MSC determines the time of the last interaction.
A location-based registration which occurs when a mobile unit moves into a new paging area implies high mobility. A time-based registration, periodically generated by the mobile unit, implies low mobility. All aspects of the interaction: cell, sector, interaction type, time of day, and the class of mobile unit can be used to calculate a mobility index. The exemplary embodiment of the present invention uses cell, sector, and interaction type for calculation.
After determining the mobility index, the MSC then determines the paging strategy. The paging strategy is preferably determined based upon the probability of paging success, which is a function of the mobility index and the elapsed time since the last interaction.
The present invention allows an MSC to determine its first paging strategy based on the mobility implied by the last interaction the mobile unit had with the MSC and how long ago the interaction occurred. The present invention thereby allows an MSC to use a single cell page and/or neighbor page on a mobile unit expected to be stationary. Considerable paging load is thereby reduced in the wireless communication system. The present invention provides a methodology for wireless services to locate mobile units through an efficient paging strategy based on each mobile unit's mobility index, thereby providing for the delivery of calls with minimal yet effective paging resource consumption.
The present invention can be better understood with reference to
Base stations 111 through 117 communication with mobile units located within their cells. Mobile units can move from cell to cell, in which case communication with the mobile unit is handed off to the new base station that is associated with the new cell.
It has been found that much of the time, mobile units are not mobile. A typical mobile unit spends a minority of its time in motion, such as when commuting to and from work, and a majority of its time stationary, such as at home or when in the office.
By using only a cell page or neighbor page on a mobile unit known to be stationary, considerable paging load can be reduced. A cell page to cell 117 pages all sectors of cell 117 and nothing else. A neighbor page would likewise page cell 117 plus all cells with a handoff relationship to cell 117, cells 101 through 106. As used herein, the term “focused pages” refers to cell and neighbor pages.
The confidence in mobility index 201 is affected by time. For example, at the time of any interaction with the communication system, the probability that a mobile unit is in the domain of the cell serving that interaction is close to 100%. As time progresses, the probability that the mobile unit remains in the domain of the cell that served the last interaction drops. In a preferred embodiment, the probability over time can be modeled as an exponential distribution. The probability can be depicted by Eq. 1.
Ppageable=e−t/M (Eq. 1)
Where:
Ppageable is the probably that the mobile unit is pageable in a cell associated with the interaction.
In accordance with an exemplary embodiment of the present invention, a mobility index M 201, which represents the mean decay time, is used to predict the probability of paging success at any time t after the last interaction. Mobility index 201 is preferably calculated by recording page response activity in the MSC. The communication system then determines the number of paging messages to initially transmit based upon mobility index 201. The paging type that is selected is the paging type that has a predetermined probability of success.
As relevant data, the table includes a mobility index 714 and cutoff time 715. Cutoff time 715 includes the elapsed time since the last interaction representing a cutoff for using a focused page. This data allows the assignment of a unique mobility index and cutoff time for each combination of cell, sector, and interaction type. The intention of this data is that when a mobile unit completes an interaction matching a record in this data base, the mobile unit will be assigned the corresponding mobility index and cutoff time. Rows 701 through 703 represent sample records in this database.
An exemplary embodiment of the present invention utilizes modeling techniques to generate the Cellular Mobility Database in
For a bucket containing page responses for times t through 2t, it can be shown that, given that {overscore (t)}=the average time for samples in the bucket:
Then mobility index 201 can be calculated as:
PBucket=e−{overscore (t)}/M
Mobility Index M 201 can then be calculated using the following equations:
ln(PBucket)=−{overscore (t)}/M
M=−{overscore (t)}/ln(PBucket)
Mobility index M 201 is used to calculate a cutoff time for any probability Pcutoff as:
tCutoff=−M ln(Pcutoff)
Suppose mobility index 201 is 20 minutes and we have selected PCutoff of 95%. That is, we want to know when there is a 95% chance that the mobile unit is within its target area. We then calculate tCutoff as:
tCutoff=−M ln(Pcutoff)=−20 ln(0.95)=1.03 minutes
Call processing can use the value tCutoff to decide whether or not to use a focused page. A focused page is a page that is sent on a predetermined number of base stations. Periodically, as a low priority process, tCutoff should be recalculated.
Although the mobility index can be calculated using data from only one page response, it is preferable to utilize larger samples. In an exemplary embodiment of the present invention, the mobility index is calculated using at least 5/(1−PBucket) samples.
The MSC looks up (301) data in the Mobile Unit Paging Database including the time of the last interaction, the interaction type, and the cutoff time associated with that interaction. The cutoff time is preferably calculated based on the last interaction of the mobile unit with the MSC, and is depicted in greater detail in
In an exemplary embodiment, the cutoff times are calculated for every cell, and also for the MSC as a whole. In a further exemplary embodiment, the cutoff times are calculated for larger interaction areas. The algorithm calculates the cutoff time for being within the for a given probability same cell as the last interaction. The MSC can calculate the cutoff time for being within the same cell and its immediate neighbors, for being within the same sub-MSC paging area, or for being within the same MSC. In further exemplary embodiments, the MSC can calculate the cutoff times for different times of day and for different subscribers.
If data does not exist (302), the MSC uses (307) standard paging procedures.
If the elapsed time since the last interaction is greater than the cutoff time as determined in step 304, the MSC uses (307) standard paging procedures.
If data exists and the elapsed time since the last interaction is less than the cutoff time, the MSC issues (305) a focused page.
A mobile unit can have different mobility indices for different levels of granularity. The mobility index for a single cell page might imply a low probability of success while the mobility index for paging a cell and all its neighbors might imply a high probability of success.
At the conclusion of an interaction between the mobile unit and the MSC, the MSC determines (401) the interaction type, location, and time. In accordance with an exemplary embodiment of the present invention, interactions between a mobile unit and an MSC are classified as either high mobility interactions, low mobility interaction, or ambiguous interaction. High mobility interactions include a call with more than one handoff and a location-based AR. Low mobility interactions include a call with no handoffs and a time-based AR. Ambiguous interactions include a call with one handoff, power-on AR, and delivery of an SMS message or a MWI message.
In an exemplary embodiment, the mobility index is a function of the cell sector. A cell sector serving a busy highway implies high mobility. A cell sector serving a shopping mall implies low mobility. The cutoff time is a function of the mobility index and the desired probability for a successful focused page.
The MSC then looks up (403) the cutoff time associated with the data in the cellular mobility database. In an exemplary embodiment, each cell/sector/interaction will have its own cutoff time.
The MSC then updates (405) the mobile unit Paging Database entry for the mobile unit. The database is updated with the cutoff time, cell, sector, and interaction time.
After a page response, the MSC looks up (502) the mobile unit in the Mobile Unit Paging Database in order to determine the interaction time, interaction type, and interaction sector cell. If a record exists (503), the data is used to look up (504) a record in the Cellular statistics data base.
If the records are found (505), the MSC uses the Mobile unit paging data base entry and the current time to determine (506) the elapsed time and thus the bucket associated with the interaction. In accordance with an exemplary embodiment of the present invention, there are 15 buckets with each bucket containing 120 seconds of elapsed time. That is, the first bucket contains data for times between 0 and 119 seconds, the second bucket contains data for times between 120 and 239 seconds, and so on.
If the bucket exists (507), the MSC increments (508) the bucket response counter by 1 and the bucket time by the elapsed time between the interaction and the page response. If the page response is within the focused paging area of the interaction (509), the success counter is incremented (510) by one.
{overscore (t)}=bucket time/bucket responses
PBucket=bucket successes/bucket responses
Buckets in the Cellular Statistics Database are useable for calculating the mobility index if there are enough samples. In accordance with an exemplary embodiment of the present invention, the minimum number of samples calculated by the MSC is:
min_samples=5/(1−Pcutoff)
These samples should come from buckets with bucket probability close to the probability of interest, Pcutoff.
The Cellular Mobility Database generation algorithm is preferably run periodically. The algorithm iterates through and looks up (602) each record in the Cellular Statistics Database. For each record that exists (603), the algorithm selects (604) a statistically valid set of buckets. If the buckets are statistically valid (605), the algorithm sums (606) those buckets together. That is the fields, bucket_responses, bucket successes, and bucket time are summed together to produce total responses, total successes, and total times.
The algorithm calculates (607) data as:
{overscore (t)}=total times/total responses
P=total successes/total responses Mobility Index M=−{overscore (t)}/ln(P) tCutoff=−M ln(PCutoff)
The algorithm updates (608) the Cellular Mobility Database with this data.
While this invention has been described in terms of certain examples thereof, it is not intended that it be limited to the above description, but rather only to the extent set forth in the claims that follow.
