HK1108596B

HK1108596B - Initial cell search method and equipment in mobile communications systems

Info

Publication number: HK1108596B
Application number: HK07113816.8A
Authority: HK
Inventors: 托里尼．帕莱纽斯
Original assignee: Lm爱立信电话有限公司
Priority date: 2004-12-06
Filing date: 2005-12-05
Publication date: 2010-02-26

Description

Initial cell search method and apparatus in mobile communication system

Technical Field

The present invention relates to a mobile communication system, and more particularly, to an initial cell search method and apparatus in a mobile communication system.

Background

Mobile communication systems, such as cellular communication systems, allow mobile User Equipment (UE) to communicate wirelessly by establishing a wireless (e.g., radio) link between the UE and one of several Base Stations (BSs) geographically distributed throughout a service area. Mobility is provided by means of a protocol that allows a UE to be handed off from a first BS to another BS when the UE moves from the coverage area of the first BS to the coverage area of the other BS.

Various base stations are connected (e.g., by wireless and/or wired links) to a Public Land Mobile Network (PLMN) that provides the infrastructure necessary to service calls. PLMNs also typically have a connection to a Public Switched Telephone Network (PSTN) so that calls may be routed to wireline communication devices not associated with the PLMN.

A UE that is powered on typically temporarily "campon" the control channels of the appropriate base station even when the UE is not actively answering the call. This allows the UE to be notified and respond when the UE is the recipient of the call, and also allows the user to quickly initiate his or her own call.

However, when the UE is first powered on, or when the network has been interrupted for a long time (e.g., when the UE has left the coverage area for a long time), the terminal must perform an initial cell search procedure to identify which cells are available (each cell being associated with a base station). The UE will select the best cell among the available cells it finds during the search.

Since the UE may "wake up" essentially anywhere (e.g., in a different region than the one it was last powered on), the initial cell search typically involves searching for the presence of control channels throughout the entire available radio frequency band. One obstacle in this respect is that the accuracy of the oscillator of the UE may vary, mainly due to temperature fluctuations of the frequency generating components. As long as the internal temperature is stable and no other factors affecting the frequency occur, the accuracy of the generated frequency will be stable (and hence the error will also be stable). The change in the UE internal temperature may be due to a change in activity performed by the UE (e.g., the UE starts to receive or transmit data) and/or due to a change in the environment surrounding the UE.

Due to the possibility of varying frequency errors, in the event that a frequency error results in a wide difference between the frequency generated by the UE and the exact frequency being used by the transmitting base station, the conventional initial cell search procedure must monitor not only the center frequency of the potential control channels within the available radio frequency band, but also a certain number of frequencies on both sides of the "desired" center frequency.

An example related to the Wideband Code Division Multiple Access (WCDMA) standard for mobile communications is now given to illustrate a conventional initial cell search process. However, the invention presented herein should not be seen as limited to application only to WCDMA systems, as it is equally applicable to other mobile communication systems as well.

Conventional initial cell search techniques typically assume that the frequency error is large, on the order of about 10 parts per million (10ppm) when searching for a carrier, which means 20kHz over the 2GHz band. A good level of accuracy in the frequency generation will allow coherent integration of the received signal to obtain good performance. However, when the frequency error is large, the coherence in the receiver deteriorates, and thus the performance of the receiver also deteriorates. This results in a long search time.

To improve search time, one approach involves using several searches at different center frequencies, each of which exhibits a better level of accuracy. For example, errors of 20kHz can be compensated by performing 4 searches (), each search bearing an error of 5 kHz. For carrier wave f_c. + -. 5kHz and f_cThese searches were performed at ± 15 kHz. This method has a disadvantage in that it consumes about 4 times as long as the time consumed by a single search in the case where the frequency error is less than 5 kHz.

Fig. 1a to 1c are flowcharts illustrating a conventional initial cell search algorithm using the method of searching for a plurality of center frequencies on both sides of an actual desired center frequency just described. For example, the initial cell search procedure may be applied to a Universal Mobile Telecommunications System (UMTS). Fig. 1a illustrates an overview of the overall process. The purpose of the search is to identify the carrier frequency being used by the cell associated with the target PLMN. To initiate the search, an initial search list including all legitimate UMTS Absolute Radio Frequency Channel Numbers (UARFCNs) is merged (block 101).

One aspect of this approach is that finding a cell located on one center frequency can allow other neighboring frequencies to be excluded in subsequent searches, which has the effect of speeding up the overall search time. Thus, to increase the likelihood of finding a cell, the initial search first performs a history list search (block 103). The history list may for example comprise a number of the latest frequencies on which suitable cells were found (e.g. 5, although this number is not an issue).

FIG. 1b is a flow chart of an exemplary history list search algorithm 103. In this example, the history list includes a number of the latest frequencies on which suitable cells were found. This list is continually updated each time a new PLMN/frequency is found, as shown in block 121. When the UE is deactivated/powered off, the history list is stored in a non-volatile memory for later use when the UE is powered on again.

To begin the actual search, the first UARFCN in the history list is selected (block 123). A search loop is then entered which runs a cell search on the selected UARFCN and then removes the UARFCN from the initial search list (block 125). If a new cell is found ("yes" path out of decision block 127), the information received from the cell is used to determine if it is from the target PLMN (decision block 129). If the cell is from the target PLMN ("yes" path out of decision block 129), the search algorithm need not be performed.

However, if the found cell is not from the target PLMN ("no" path out of decision block 129), then all UARFCNs that are ± 3MHz away from the UARFCN associated with the found cell are removed from the initial search list (block 130). This has the effect of speeding up the overall search time, since removing these UARFCNs from the initial search list will prevent these carriers from being searched again in the subsequent number of cycles of the initial cell search.

After block 130, or if no cell has been found on the selected UARFCN ("no" path out of decision block 127), a determination is made as to whether the last UARFCN in the history list has been selected (decision block 131). If not ("no" path out of decision block 131), the next UARFCN in the history list is selected (block 133) and the loop is repeated by returning the process to block 125. It is determined that the last UARFCN in the history list has been selected ("yes" path out of decision block 131) constitutes the end of the history list search 103.

Returning to fig. 1a, when the history list search is completed, the next activity involves the following processing for the Downlink (DL) band.

First, the initial search list is narrowed by filtering out some frequencies based on the Received Signal Strength Indicator (RSSI) (block 105). The filtration process involves:

performing an RSSI scan on each UARFCN in the initial search list;

removing all UARFCNs satisfying RSSI ≦ -100dBm from the initial search list for any frequency within a distance of + -100kHz, + -300 kHz, + -500 kHz from the center frequency of the DL band;

remove all UARFCNs that satisfy RSSI ≦ -95dBm from the initial search list for any frequencies that are not + -100kHz, + -300 kHz, + -500 kHz from the center frequency of the DL band.

By removing frequencies that are unlikely to result in finding a cell, the search effort is further reduced to testing only the most likely carriers in the band.

Finally, the resulting (filtered) list is searched (107). FIG. 1c is a flow chart illustrating an exemplary searcher 107. The approach used is to search first for the most likely frequency and then for all other frequencies in the search list. Referring now to FIG. 1c, the hub to be usedFrequency f_cSet equal to the center frequency (e.g., f) of the DL band_c2112.5MHz) and is selected to have a frequency f_cUARFCN at 100kHz (block 141).

If the selected UARFCN is in the initial search list ("yes" path out of decision block 143), a cell search is performed on the selected UARFCN and the selected UARFCN is removed from the initial search list (block 145) to prevent a second search thereof. If the cell search finds a new cell ("yes" path out of decision block 147), information received from the cell is used to determine whether it is from the target PLMN (decision block 148). If it is ("yes" path out of decision block 148), no further search need be performed.

However, if the found cell is not from the target PLMN ("no" path out of decision block 148), then all UARFCNs that are ± 3MHz away from the selected UARFCN associated with the found cell are removed from the initial search list (block 149).

Thereafter, either if no new cell is found ("no" path out of decision block 147) or if the selected UARFCN is not found in the initial search list ("no" path out of decision block 143), the following algorithm is performed: either the next UARFCN to be used in the subsequent round is selected or the initial search is ended (block 151). To perform the next cycle, processing returns to decision block 143.

The processing associated with block 151 may be performed in any of several ways (i.e., either selecting the next UARFCN to be used in a subsequent round of rotation or ending the initial search). For example, the carriers may be sorted in RSSI order (the strongest carrier appears first) and searched in that order until all carriers have been selected for searching (at which point the initial search is ended).

In one embodiment, the entire frequency band is divided into several smaller frequency bands. For each of these smaller bands, a known center frequency is selected, and then block 151 ensures that at some future time, f will be selected_c±100kHz、f_c. + -. 300kHz and f_cEach of the carriers defined by ± 500 kHz.

For more information on the well-known initial Cell search technique, the interested reader is referred to U.S. patent publication No. US 2004/0203839A1 (Ostberg et al, "Mobile Terminals and Methods for Forming Fast Initial frequency Scans and Cell searchs") published at 10.14.2004.

One problem with the conventional initial cell search algorithm is that the search for all carriers takes a long time. In some cases it may take several minutes before finding an allowed PLMN. One effect of this on the UE is that the time to register with the network is long, which in turn means that the time from first powering the UE on to the time a call can be made is long. This adversely affects the user of the UE.

Another impact on the UE is that the current consumption when performing the initial cell search algorithm is high.

It is therefore desirable to propose an initial cell search apparatus and method that can more quickly identify cells associated with allowed PLMNs.

Disclosure of Invention

It should be emphasized that the term "comprises/comprising" when used in this specification is taken to specify the presence of stated features, integers, steps or components; but the use of this term does not preclude the presence or addition of one or more other features, integers, steps or components or groups thereof.

In accordance with an aspect of the present invention, the above and other objects are achieved in methods, apparatuses, and machine-readable storage media for performing an initial search for a cell search in a telecommunication system. The initial cell search comprises the steps of: defining a set of carrier frequencies to be searched; and selecting either a wide cell search mode or a narrow cell search mode as a cell search mode, wherein the selection of the cell search mode is based on a frequency generation accuracy level. The wide cell search mode searches for a wider frequency range than the narrow cell search mode. For each carrier frequency in the set of carrier frequencies to be searched, the most recently selected cell search pattern is used when searching for cells transmitting on that carrier frequency. The narrow cell search mode is used only when the frequency generation accuracy level is better than an expected worst frequency generation accuracy level.

In another aspect, the initial cell search comprises the steps of: the wide cell search mode is initially selected for use whenever a cell is searched that transmits on one of the set of carrier frequencies to be searched. When a cell search is found as a result of performing a cell search on one carrier frequency of the set of carrier frequencies to be searched, a frequency generation accuracy level is improved using a signal received from the found cell. In response to an improvement in the level of frequency generation accuracy, the narrow cell search mode is then selected for use when a next search is performed for a cell transmitting on another carrier frequency of the set of carrier frequencies to be searched.

In an alternative embodiment, the initial cell search comprises the steps of: the method further includes initially selecting one of the wide cell search mode and the narrow cell search mode to use each time a cell is searched that is transmitted on one of the set of carrier frequencies to be searched, wherein the initial selection is based on one or more parameters indicative of a current frequency generation accuracy level. When a cell search is found as a result of performing a cell search on one carrier frequency of the set of carrier frequencies to be searched, a frequency generation accuracy level is improved using a signal received from the found cell. In response to an improvement in the level of frequency generation accuracy, the narrow cell search mode is then selected for use when a next search is performed for a cell transmitting on another carrier frequency of the set of carrier frequencies to be searched.

In another aspect, the one or more parameters indicative of the current frequency generation accuracy level include one or more automatic frequency control parameters.

According to another aspect of the present invention, there is provided an apparatus for searching for a cell in a telecommunication system, the apparatus comprising logic for: logic that defines a set of carrier frequencies to be searched; logic that provides a wide cell search mode and a narrow cell search mode for use by logic that selects a cell search mode; logic that selects either the wide cell search mode or the narrow cell search mode as a cell search mode, wherein the cell search mode is selected based on a frequency generation accuracy level of a local frequency generator in a mobile terminal; logic for using, for each carrier frequency in the set of carrier frequencies to be searched, a most recently selected cell search pattern when searching for a cell transmitting on that carrier frequency, wherein: the wide cell search mode searches for a wider frequency range than the narrow cell search mode; and the narrow cell search mode is used only when the frequency generation accuracy level is better than an expected worst frequency generation accuracy level.

Drawings

The objects and advantages of the invention will be understood by reading the following description in conjunction with the drawings, in which:

fig. 1a to 1c are flowcharts illustrating a conventional initial cell search algorithm.

Fig. 2a to 2c are flowcharts illustrating an improved initial cell search algorithm according to the present invention.

Detailed Description

Various features of the present invention will now be described with reference to the drawings, wherein like reference numerals are used to refer to like parts throughout.

Various aspects of the invention will now be described in more detail in connection with several exemplary embodiments. To facilitate an understanding of the invention, aspects of the invention are described in terms of sequences of actions to be performed by components of a computer system. It will be recognized that in various embodiments, various actions could be performed by specialized circuits (e.g., discrete logic gates interconnected to perform a specialized function), by program instructions being executed by one or more processors, or by a combination of both. Moreover, the invention can additionally be considered to be embodied entirely within any form of computer readable carrier, such as solid-state memory, magnetic disk, optical disk or carrier wave (e.g., radio frequency, audio frequency or optical frequency carrier waves) containing a suitable set of computer instructions that would cause a processor to carry out the techniques described herein. Thus, the various aspects of the invention may be embodied in many different forms, all of which are considered to be within the scope of the invention. For each of the various aspects of the invention, any such form of embodiment may be referred to herein as logic designed to perform the recited action, or alternatively as "logic that" performs the recited action.

As described in the background section, the length of time of the initial cell search algorithm is directly related to the number of frequencies that must be searched. In the conventional art, the number of frequencies to be searched is set higher than the number actually required by an ideal receiver because the receiver is not ideal in practice-they generate frequency errors. The accuracy of the oscillator of the UE may vary, mainly due to temperature fluctuations of the frequency generating components. The accuracy of the generated frequency is stable (and conversely the error is stable) as long as the internal temperature is stable and no other things occur that affect the frequency. The change in the internal temperature of the UE may be due to a change in activity performed by the UE (e.g., the UE begins to receive or transmit data) and/or due to a change in the environment surrounding the UE.

According to an aspect of the present invention, information on a frequency error of a UE is considered in a cell search procedure. When the error is known to be low, less frequent searches are performed, thereby speeding up the search without adversely affecting the search results.

In another aspect of the invention, knowledge about frequency errors is inferred from the recent history of UE operation. For example, if a UE initially has a connection to the network but then breaks the connection to the network, for example, due to leaving coverage, the UE initially has accurate frequency generation. Accordingly, it can be considered that a UE whose coverage is known to have recently left generates a relatively accurate frequency, and thus the number of frequencies searched in the cell search operation can be reduced, thereby reducing the time taken to search for a cell on each carrier and reducing power consumption.

In yet another aspect of the present invention, multiple steps are taken in cell search to improve frequency accuracy. The knowledge of the improved frequency accuracy is then used to shorten the cell search time. For example, consider the example in the context of a WCDMA system (the invention is not limited to WCDMA systems). When the UE performs an initial cell search and finds a cell, the UE reads a Broadcast Channel (BCH) to determine whether the cell is associated with the target PLMN and also determines other cell information. If the cell belongs to a different PLMN than the PLMN to be searched, or alternatively if the UE is not allowed to connect to the cell, the UE continues to search for other carriers for the cell associated with the target PLMN. In one aspect of the invention, when the UE reads the BCH, it does its frequency generation at the same time.

When the UE has connected to the network in this way, it has a very good estimate of the frequency, which in many cases will be valid for a rather long time, depending on the temperature variation and long-term stability of the crystal oscillator. During this time, the UE may continue to search for the rest of the spectrum with smaller frequency errors.

For example, by basing the cell search algorithm on a frequency error equal to only a few kHz instead of, say, 10ppm (e.g., 20kHz in the 2GHz band (as in conventional search algorithms)), the cell search algorithm may be able to perform a cell search on each carrier instead of four times, thereby resulting in a 4-fold improvement in cell search speed. This still represents a 2-fold speed improvement even though the cell search algorithm has to perform two cell searches per carrier.

Thus, due to better coherence of the receiver with the received signal, the UE may have the following capabilities:

1. the cell search time is shortened, thus shortening the time for the UE to connect to the network.

2. Sensitivity in searching for new cells is improved so that weaker cells can be found in a time equal to that required by conventional techniques to find only stronger cells. This application of the inventive technique will reduce the risk of not finding a suitable network.

3. Since the activity of the UE is reduced, the power consumption of the UE is improved when the UE is out of coverage.

An exemplary embodiment will now be described in connection with fig. 2a to 2d, which fig. 2a to 2d are flow charts illustrating a new initial cell search algorithm. The initial cell search may be applied, for example, to a Universal Mobile Telecommunications System (UMTS).

The new initial cell search algorithm employs two different types of cell searches: a "wide" cell search and a "narrow" cell search. The wide cell search is employed when frequency accuracy is not certain to be good, and in this example involves: for each searched carrier, 100ms per bin (frequency bin) and 4 bins per carrier are used. This is the type of search performed in conventional systems. The narrow cell search in this example involves: for each searched carrier, 100ms per band and only 1 band (i.e., one search) per carrier is used. It will be appreciated that in other (alternative) embodiments, the specific duration and number of bands per band defining the wide cell search and narrow cell search may be different. However, in embodiments, a wide cell search will involve more frequency bands (i.e., more searches) than a narrow cell search, or more generally expressed as: a wide cell search will involve searching a wider frequency range than a narrow cell search, and thus a wide cell search consumes more time and/or energy than a narrow cell search.

In the logic flow to be introduced, whether to perform a wide cell search or a narrow cell search is controlled by a parameter "CS", which may have a value representing "wide" or "narrow". This parameter CS is initially set to "wide", but is changed to "narrow" once a cell from any PLMN is found. As described earlier, the UE utilizes discovery of any cell as an opportunity to improve its frequency accuracy. The improved frequency accuracy is then considered to remain unchanged for the remaining duration of the cell search algorithm.

Turning now to exemplary embodiments, FIG. 2a illustrates an overview of the overall process. The purpose of the search is to identify the carrier frequency being used by the cell associated with the target PLMN. To begin the search, an initial search list including all legitimate UMTS Absolute Radio Frequency Channel Numbers (UARFCNs) is merged and the parameter CS is initialized to "wide" (block 201).

One aspect of this approach is that the discovery of a cell on one center frequency may allow other neighboring frequencies to be excluded from subsequent searches, which has the effect of speeding up the overall search time. Thus, to increase the likelihood of finding a cell, the initial search first performs a frequency accuracy based history list search (block 203). The history list may for example consist of the 5 most recent frequencies on which a suitable cell was found.

Fig. 2b is a flow diagram of an exemplary embodiment of the improved history list search algorithm 203. In this example, the history list consists of a number of most recent frequencies (e.g., 5 most recent frequencies) on which suitable cells were found. This list is continuously updated each time a new PLMN/frequency is found, as shown in block 221. When the UE is deactivated/powered off, the history list is stored in a non-volatile memory for later use when the UE is powered on again.

To begin the actual search, the first UARFCN in the history list is selected (block 223). A search loop is then entered which runs a cell search on the selected UARFCN (either a wide search or narrow search algorithm is selected under control of the current state of the CS parameters) and then removes the UARFCN from the initial search list (block 225). If a new cell is found ("yes" path out of decision block 227), the information received from the cell is used to determine if it is from the target PLMN (decision block 229). If the cell is from the target PLMN ("yes" path out of decision block 229), the search algorithm need not be performed.

However, if the found cell is not from the target PLMN ("no" path out of decision block 229), then all UARFCNs that are ± 3MHz away from the UARFCN associated with the found cell are removed from the initial search list (block 230). This has the effect of speeding up the overall search time, since removing these UARFCNs from the initial search list prevents searching for these carriers in the subsequent round of the initial cell search. Furthermore (also shown in block 230), the UE takes advantage of its (temporary) connection to this cell by increasing its frequency accuracy and setting the CS parameter equal to "narrow", thereby ensuring that all further cell searches will be narrow cell searches.

After block 230, or if no cell has been found on the selected UARFCN ("no" path out of decision block 227), a determination is made as to whether the last UARFCN in the history list has been selected (decision block 231). If not ("no" path out of decision block 231), the next UARFCN in the history list is selected (block 233) and the loop is repeated by returning the process to block 225. It is determined that the last UARFCN in the history list has been selected ("yes" path out of decision block 231) constitutes the end of the history list search 203.

Returning to fig. 2a, when the history list search is completed, the next activity involves the following processing for the Downlink (DL) band.

First, the initial search list is narrowed by filtering out some frequencies based on the Received Signal Strength Indicator (RSSI) (block 205). In an exemplary embodiment, the filtering process involves:

performing an RSSI scan on each UARFCN in the initial search list;

Finally, the resulting (filtered) list is searched (207). FIG. 2c is a flow chart illustrating an exemplary refined searcher 207. The approach used is to search first for the most likely frequency and then for all other frequencies in the search list. Referring now to fig. 2c, the center frequency f to be used_cSet equal to the center frequency (e.g., f) of the DL band_c2112.5MHz) and is selected to have a frequency f_cUARFCN at 100kHz (block 241).

If the selected UARFCN is in the initial search list ("yes" path out of decision block 243), a cell search is performed on the selected UARFCN (either the wide search or narrow search algorithm is selected under control of the current state of the CS parameters) and the selected UARFCN is removed from the initial search list (block 245) to prevent a second search thereof. If the cell search finds a new cell ("yes" path out of decision block 247), the information received from the cell is used to determine whether it is from the target PLMN (decision block 248). If it is ("yes" path out of decision block 248), no further search need be performed.

However, if the found cell is not from the target PLMN ("no" path out of decision block 248), then all UARFCNs that are ± 3MHz away from the selected UARFCN associated with the found cell are removed from the initial search list (block 249). Furthermore (also shown in block 249), the UE takes advantage of its (temporary) connection to the cell in such a way as to increase its frequency accuracy and ensure that the CS parameter is set equal to "narrow" based on now having a good level of frequency accuracy, thereby ensuring that all further cell searches will be narrow cell searches.

Thereafter, either if no new cell is found ("no" path out of decision block 247) or if the selected UARFCN is not found in the initial search list ("no" path out of decision block 243), the following algorithm is performed: the next UARFCN to be used in the next cycle of the loop is selected or the initial search is ended (block 251). To perform the next round of the loop, processing returns to decision block 243.

The processing associated with block 251 may be performed in any of several ways (i.e., selecting the next UARFCN to be used in a subsequent round, or ending the initial search), none of which is critical to the present invention. For example, the carriers may be sorted in RSSI order (the strongest carrier appears first) and searched in that order until all carriers have been selected for searching (at which point the initial search is ended).

In one embodiment, the entire frequency band is divided into several smaller frequency bands. For each of these smaller bands, a known center frequency is selected, and then block 151 ensures that f will be selected in the near future_c±100kHz、f_c. + -. 300kHz and f_cEach of the carriers defined by ± 500 kHz.

Next, it can be seen in this exemplary embodiment, as well as other embodiments, that the initial cell search algorithm dynamically selects either a wide cell search or a narrow cell search for a selected frequency as a function of UE frequency generation accuracy.

The above embodiments also illustrate the following aspects: as long as the UE is connected to a cell during the initial cell search algorithm, the UE may improve its frequency generation accuracy even if the cell is not associated with a target PLMN, or is otherwise a cell that the UE is not allowed to access. Once the frequency generation accuracy is improved, the UE then performs more efficient subsequent cell searches (each of which is performed on the selected carrier frequency) because it does not have to take into account the possibility of low frequency generation accuracy.

In an alternative embodiment, the sensitivity of the search for weak cells may be increased while maintaining comparable speed performance (or better than the latter) to conventional initial cell search routines by modifying the initial cell list filtering process performed by block 205 to not remove weaker cells (as opposed to filtering out weaker cells in the exemplary embodiment described above). For example, the RSSI threshold level that determines whether to remove a particular UARFCN may be set to a lower value to allow a certain number of weaker cells to remain in the initial cell list. Perhaps the most straightforward way to do a deep search for a cell is to have essentially no threshold at all. For practical purposes, this is equivalent to setting the threshold to-100 dBm, since this level can typically be triggered by the internal noise of the receiver. To just find strong cells, the threshold may be from-95 dBm up to-80 dBm, or even higher. The particular threshold level that should be selected depends on the application and therefore cannot be explicitly set forth herein. Those skilled in the art can readily determine appropriate values for achieving their desired performance levels.

It should be observed that in the exemplary embodiments illustrated in fig. 2a to 2c, the initial cell search algorithm is initialized such that the default search mode is to perform a wide search (i.e., a more time consuming search) and to switch to performing a more efficient narrow cell search only after the frequency generation accuracy level of the UE has been improved (i.e., by improving frequency generation with signals received from cells that have been found). However, in yet another alternative set of embodiments, the initial cell search mode is set based on some other parameter that represents the current frequency generation accuracy level of the UE. For example, in some UEs, a previously determined Automatic Frequency Control (AFC) setting may indicate a current performance level of the frequency generator of the UE. Thus, these parameters can be utilized to determine whether the initial cell search algorithm will begin performing a wide cell search or a narrow cell search. Whether such an approach is applicable to a particular UE depends on the particular design and structure of the UE, and in some cases the frequency accuracy of the UE may change very quickly when the UE is first turned on due to, for example, heat dissipation by one or more chips and other effects affecting frequency generation. In this case, the CS parameter is preferably initialized to represent a wide cell search.

An analysis of the performance improvement that can be achieved with the new techniques described above is now described. The performance of the conventional cell search algorithm will be compared with the performance of one embodiment of the new cell search algorithm, combining 3 different cases. In each case, it is assumed that both conventional and new techniques divide the entire frequency band into several smaller frequency bands. For each of these smaller bands, a known center frequency is selected, and then block 151 ensures future selection by f_c±100kHz、f_c. + -. 300kHz and f_cEach of the carriers defined by ± 500kHz is searched.

In the first case, only 3 carriers are available for the UE to detect. The carriers are transmitted in some of the most likely carriers. It is assumed that the RSSI of all remaining frequencies measured is below-100 dBm. Assume that the UE is to find an unreachable PLMN (i.e., the initial cell search algorithm will not end early due to finding an acceptable cell). It is also assumed that the history list is empty.

Looking first at the conventional initial cell search algorithm, a search is started for the most likely carrier of ± 100kHz from the middle of the 125 MHz channels at an RSSI greater than-95 dBm. Only 3 of the 5MHz channels meet this. A cell is found in each of the 3 5MHz channels and, on average, in 1.5 of the 3 channels, on the first carrier searched. The remaining 95 MHz channels are not searched.

The algorithm also removes from the search those carriers that are located ± 3MHz from the carrier with the found cell. Since the RSSI measurements show that the measured values deviate by about ± 4.5MHz from the received carrier, in the worst case there is about 1.5MHz on both sides of the carrier where the search is also performed.

This means that a total of 4.5 carriers are searched among the most likely carriers. Then another 7 carriers are searched on both sides of each detected cell. Assuming (due to the wide search) that 4 searches are performed per carrier, each 400ms, plus about 1.5s for reading BCH on each found carrier, the search would cost in the worst case:

0.4*4*4.5+0.4*4*7*2*3+3*1.5s＝7.2+67.2+4.5s＝78.9s。

now, considering the performance of the new initial cell search algorithm in this first case, even if the UE does not decide to camp on a found cell, the following estimates for identifying the frequency carrier associated with the cell can be derived: the frequency will be locked after the maximum carrier. Thereafter, the search would take 400ms instead of 1600ms, since a narrow cell search would be performed instead of a wide cell search. The search order is the same as the above.

Thus, with the new algorithm, a total of 4.5 carriers out of the most likely carriers are searched, and a wide search can be performed on 2 carriers among these carriers. Then, another 7 carriers are searched on both sides of each detected cell. Therefore, in the worst case the search will cost:

0.4*4*2+0.4*2.5+0.4*7*2*3+3*1.5s＝4.2+16.8+4.5s＝25.5s。

the new initial cell search algorithm shows a significant improvement over the conventional approach.

Now consider a second case where only 3 carriers are available for the UE to detect. It is also assumed that these 3 carriers are transmitted on some of the most likely carriers and the history list is empty. The RSSI measured for all other frequencies was above-95 dBm. Suppose the UE is to find an unreachable PLMN (i.e., the initial cell search algorithm will not end early due to finding an acceptable cell). For example, these conditions are for the 1900MHz band and since the non-linear characteristic of the receiver is effective, the non-linear characteristic results in a very high RSSI level for all frequencies in case the input signal from one carrier is very high. The test point at-52 dBm should be 33dB for the adjacent channels, as needed. Even this would result in an RSSI level on the adjacent channel equal to-75 dBm. At higher input levels (in which case the linearity of the receiver will degrade, which means levels above or approximately above-40 dBm), the RSSI level rises across the band. This tends to be quite common in certain environments where any operator has micro cells, but it is reasonable that degradation occurs where RSSI levels in most of the frequency band are greater than-95 dBm.

First consider a conventional initial cell search algorithm that starts searching for the most likely carrier of + -100kHz from the middle of the 125 MHz channels at an RSSI greater than-95 dBm. All 5MHz channels meet this. Of these 125 MHz channels, cells were found on 3 of them, and on average, in 1.5 of these channels, cells were found on the first carrier searched.

Those carriers located within ± 3MHz from the carrier associated with the found cell are removed from the search. This means that 18MHz (90 carriers) are excluded from the search after these 3 cells are found. Thus leaving 277-90-187 carriers to be searched, on which there are no cells but the RSSI level is above the threshold.

This means that a total of 4.5 carriers closest to the carrier used will be searched. In addition, a search will also be made for 187 other carriers. Assuming each search for 400ms, plus a BCH read time of 1.5s for each found carrier, this conventional search would take:

0.4*4*4.5+0.4*4*187+3*1.5s＝18.6+4.5s＝306.4s。

considering now the new initial cell search algorithm performed in this second case, the search starts on the most likely carriers (if they are defined). There are 3 carriers with detectable cells, so the UE searches between 1 to 105 MHz channels, which means that it searches between 1 to 20 carriers before it finds the first cell; this takes between 1.6s and 20 x 1.6 s-32 s, on average about 16.8 s. Then, the remaining 5MHz channels (between 2 and 11 5MHz channels) are searched; this takes between 2 x 0.4s and 11 x 0.4s, on average about 2.6 s. After these searches are performed, the algorithm continues to search for (not the most likely) 187-9 x 2 carriers-169 carriers.

This means that the entire search will be costly

16.8+2.6+0.4*169+3*1.5s＝91.5s。

Also, the new initial cell search algorithm shows a significant improvement over conventional techniques.

Consider now a third case where there are 12 carriers available for the UE to detect. It is assumed that the carriers are transmitted on some of the most likely carriers and the history list is empty. The measured RSSI was above-95 dBm for all frequencies. Assume that the UE is to look for an unreachable home PLMN.

First consider a conventional initial cell search algorithm that starts searching for the most likely carrier of + -100kHz from the middle of the 125 MHz channels at an RSSI greater than-95 dBm. All 5MHz channels meet this. Cells were found in all of these 125 MHz channels and on average in 6 of these channels, on the first searched carrier; the cell is found in the second search on the other channel.

Those carriers located within ± 3MHz from the carrier associated with the found cell are removed from the search list. This means that all other carriers will be excluded from the search.

Therefore, a total of 18 carriers close to the used carrier will be searched. Assuming that 4 searches are performed for each carrier found, each search being 400ms plus 1.5s for reading the BCH, it can be estimated that the search will cost:

0.4*4*18+12*1.5s＝28.8+18s＝46.8s。

considering now the expected performance of the new initial cell search algorithm operating under the conditions defined by the third case, it can be seen that a total of 18 carriers close to the carrier used will be searched. The frequency will be locked after a maximum of two carriers. Therefore, it can be estimated that the search costs:

0.4*4*2+0.4*16+12*1.5s＝27.6s。

these estimates are summarized in the following table:

	routine algorithm	New cell search algorithm
	routine algorithm	New cell search algorithm	Case 1	78.9s	25.5s
Case 2	306.4s	91.5s	Case 1	78.9s	25.5s
Case 2	306.4s	91.5s	Case 3	46.8s	27.6s

Table 1: performance comparison between conventional and new initial cell search techniques

The invention has been described with reference to specific embodiments. However, it will be readily understood by those skilled in the art that the present invention may be embodied in specific forms other than those of the embodiments described above. The described embodiments are merely illustrative and should not be considered restrictive in any way. The scope of the invention is indicated by the appended claims rather than the foregoing description, and all changes and equivalents that fall within the range of the appended claims are intended to be embraced therein.

Claims

1. A method of searching for a cell in a telecommunications system, the method comprising the steps of:

defining a set of carrier frequencies to be searched;

providing a wide cell search mode and a narrow cell search mode for use in the selecting step;

selecting the wide cell search mode or the narrow cell search mode as a cell search mode, wherein the cell search mode is selected based on a frequency generation accuracy level of a local frequency generator in a mobile terminal;

for each carrier frequency in the set of carrier frequencies to be searched, using a most recently selected cell search pattern when searching for cells transmitting on that carrier frequency,

wherein:

the wide cell search mode searches for a wider frequency range than the narrow cell search mode; and is

The narrow cell search mode is used only when the frequency generation accuracy level is better than an expected worst frequency generation accuracy level.

2. The method of claim 1, further comprising the steps of:

initially selecting the wide cell search mode to use whenever a cell is searched that transmits on one of the set of carrier frequencies to be searched;

finding a cell as a result of performing a cell search on one carrier frequency of the set of carrier frequencies to be searched;

improving a frequency generation accuracy level using a signal received from the searched cell; and

in response to an improvement in the level of frequency generation accuracy, selecting to use the narrow cell search mode when a next search is performed for a cell transmitting on another carrier frequency of the set of carrier frequencies to be searched.

3. The method of claim 1, comprising the steps of:

the order in which the carrier frequency searches are arranged is based on the expected likelihood of finding the cell associated with each carrier frequency.

4. The method of claim 3, wherein the search for carrier frequencies known to be most recently associated with a suitable cell precedes the search for other carrier frequencies in the set of carrier frequencies to be searched.

5. The method of claim 1, further comprising the steps of:

initially selecting one of the wide cell search mode and the narrow cell search mode to use whenever searching for a cell transmitting on one of the set of carrier frequencies to be searched, wherein the initial selection is based on one or more parameters representing a current frequency generation accuracy level;

in response to an improvement in the level of frequency generation accuracy, it is ensured that the narrow cell search mode is selected for use when a next search is performed for a cell transmitting on another carrier frequency of the set of carrier frequencies to be searched.

6. The method of claim 5, wherein the one or more parameters representative of a current frequency generation accuracy level comprise one or more automatic frequency control parameters.

7. An apparatus for finding a cell in a telecommunication system, the apparatus comprising logic to:

logic that defines a set of carrier frequencies to be searched;

logic that provides a wide cell search mode and a narrow cell search mode for use by logic that selects a cell search mode;

logic that selects either the wide cell search mode or the narrow cell search mode as a cell search mode, wherein the cell search mode is selected based on a frequency generation accuracy level of a local frequency generator in a mobile terminal;

logic for using, for each carrier frequency in the set of carrier frequencies to be searched, a most recently selected cell search pattern when searching for cells transmitting on that carrier frequency,

wherein:

8. The apparatus of claim 7, further comprising:

a logic section for initially selecting the wide cell search mode to be used whenever a cell to be transmitted on one carrier frequency among the set of carrier frequencies to be searched is searched;

logic that finds a cell as a result of performing a cell search on one carrier frequency of the set of carrier frequencies to be searched;

logic for improving a frequency generation accuracy level using a signal received from the searched cell; and

logic that selects to use the narrow cell search mode when performing a next search for a cell transmitting on another carrier frequency of the set of carrier frequencies to be searched in response to an improvement in a frequency generation accuracy level.

9. The apparatus of claim 7, further comprising logic to:

logic that arranges the order in which carrier frequencies are searched based on the expected likelihood of finding a cell associated with each carrier frequency.

10. The device of claim 9, wherein the search for a carrier frequency known to be most recently associated with a suitable cell precedes the search for other carrier frequencies in the set of carrier frequencies to be searched.

11. The apparatus of claim 7, further comprising logic to:

logic that initially selects one of the wide cell search mode and the narrow cell search mode to use whenever searching for a cell that is transmitting on one of the set of carrier frequencies to be searched, wherein the initial selection is based on one or more parameters that indicate a current frequency generation accuracy level;

logic that ensures that the narrow cell search mode is used when a next search is performed for a cell transmitting on another carrier frequency of the set of carrier frequencies to be searched in response to an improvement in the level of frequency generation accuracy.

12. The apparatus of claim 11, wherein the one or more parameters indicative of a current frequency generation accuracy level comprise one or more automatic frequency control parameters.