WO2013160708A1 - Automatic identification of neighbors in wcdma network - Google Patents

Description

AUTOMATIC IDENTIFICATION OF NEIGHBORS IN WCDMA NETWORK

DESCRIPTION

Technical Field

Present invention belongs to the field of communication technique, specifically to selecting arrangements to which subscribers are connected via radio links that enable completing call to or from mobile subscribers, and more specifically to handover arrangements and data processing thereof.

Technical Problem

WCDMA (Wideband Code Division Multiple Access) is well known air interface standard found in 3G mobile telecommunications network established in late 1990s. The essential characteristic of the WCDMA network is that a data signal in WCDMA is scrambled by utilizing a user-specific pseudo noise (PN) code at the transmitter side to spread the signal over the entire band. The WCDMA system uses several codes. In theory, one type of code should be enough, but in practice, radio path physical characteristics require that the WCDMA system uses different codes for different purposes. Among others, scrambling code (SC) is used in the downlink direction for cell/sector separation; and also in the uplink direction. In this case, the users, i.e. their user equipments (UE) , are separated from each other using this code.

Each cell in the WCDMA network transmits on the same frequency, but using a different SC, and different signals are distinguished by different codes. WCDMA network enables 512 different scrambling codes in the downlink that can be used, and evidently there is a need for the repetition of the same code across the network. The necessity -of scrambling code repetitions across the network has significant impact in network planning, i.e. physical reuse of the same SC should be planned for the places with maximum possible mutual distance. This is extremely challenging. requirement, especially in the dense urban areas, where the cells are. very close to each other.

In some cases the SC reuse pattern is too tight and the signal originating cell cannot be easily identified solely by SC. The observed problem is prominent during a handover procedure. Any UE constantly measures signals strength received not only from the currently serving cell but also from the neighboring cells, and said UE reports back the signals in a form of Measurement Report (MR) . Each MR contains the measured quantities of all present signals and their identities expressed in form of their SC.

It is possible that the same SC is used more than once in the vicinity of UE; vicinity herby defined as being able to be received by UE; which prevents the exact transmitting cell identification based only on measured signal strength and its SC.

Mobility is the essential feature of any wireless network. In order to ensure the mobility in WCDMA network, neighbor relations (NCELLs) have to be defined as network parameters; each neighbor relation consists of home cell (HCell) and list of neighbor cells (NCell) . When the UE reports the signal with the SC, the Radio Network Controller (RNC) checks the list of all defined relations where the HCell is the currently serving cell of the UE. If one of those relations have the NCell with the SC equal to the one measured, RNC concludes that the handover has to be initiated to the NCell with the same SC.

UE can also measure and report the signals with the SC that do not belong to any of the neighbors listed for the serving cell. Recorded event is classified as a "detected neighbor" and does not result with the handover. Namely, if RNC using the SC cannot identify to which cell the call has to be handed over, the handover is not executed, and that might result in eventual loss of the signal and the dropped calls.

The present invention solves mentioned handover problem, i.e. "detected neighbor" and improves mobility despite tight SC reuse pattern via enlargement of identification space from WCD A standard 512 SCs to maximum possible 131072, via further cells discrimination. Discrimination is performed by further characterization of each SC with possible 256 observed different time constants expressed in terms of frames and referenced to one arbitrary cell within network. Such improved cell characterization is used for forming unique synchronization topological object using neighbor cells relations and measurement reports. In addition, said relations being adequately processed by the data-processing system are used for merging, learning and optimization of synchronization topological object handled by the data-processing system.

Previous State of Art

The general information regarding state of the art can be found in: o H. Kaaranen, A. Ahtiainen, L. Laitinen, S. Naghian and V. Niemi;

"UMTS Networks", 2^nd Edition; 2005, John Wiley & Sons, Ltd; and o http: //www.3gpp.org/

Regarding the above cited technical problem and in order to improve mobility and to reduce the chance of call drops, operators usually perform one of the following general methods well known in the art:

(i) to optimize the neighbor list in the networks and by doing so, to reduce the chance of "detected neighbor"; or

(ii) to use randomly one or all cells with the same SC on the same RNC and to perform the handover to all of them; or (iii) to rely on geographical data and to guess/estimate the most likely of all candidates with the same SC, based on the distance and the observed direction of the candidates.

Method (i) ; optimizing the neighbor list is a challenging task since the length of the neighbor list is limited to 31 neighbors and usually, operator puts extra limitations like mutuality of neighbors, inter-frequency rules etc. Moreover, the WCDMA networks are dynamic networks as they are constantly built, expanded and the users' behavior change over the course of the year, weeks and even days. In order to achieve good optimal neighbor list and to minimize the detected neighbor events, constant and real time optimization of the neighbor list is required and that is not fully achievable.

Method (ii) ; taking the handover candidates with the same SC within the same RNC has few problems - it may happen that in said RNC is more than one cell with the same SC beside one already being reported. It that case, it is not possible to identify the actual serving cell unambiguously. Additionally, in 20-30% of the cases, the actual serving cell is controlled by another RNC and Iur handover should occur, as known in the art. If method (ii) is used, 20%-30% of handovers will still be missing due to the fact that Iur candidates are not considered. Moreover, if all of the possible handover candidates are used, the downlink resources in the system are wasted as the actual downlink capacity is needed from one or from none of those possible handover candidates.

Method (iii) ; using a geographical data and guessing the candidate based on the location and directiveness of the candidates can be effective, but the source of the data is troublesome. RNC does not contain the accurate geographical locations of transmitters and even rough information about location cannot be transferred over the Iur interface. So, even if the whereabouts of the own controlled cells are known, the lack of the information about the neighboring cells, controlled by neighboring RNCs will introduce the same level of uncertainty as described previously in Method (ii) . Beside the general approaches described above, a lot of efforts have been done to solve observed problems in handover . procedures . These efforts are documented in form of patent and non-patent literature. The international patent application PCT/KR2006/003389 (KIM, Shin- Jae), filed on June 28, 2006 and published as WO 2007/027034 Al teaches about method and apparatus for optimizing neighbor list automatically in asynchronous WCDMA network. The described method includes the steps of: collecting neighbor list data, call fault data, handover statistical data, base station location data and PSC (primary scrambling code) information data of each base station sub- cell (or, sector) in a nationwide network; extracting all target sub-cells (or, sectors) available for handover by analyzing the collected data; endowing a weighting factor to the extracted target sub-cells (or, sectors) according to importance and then sorting calculated results so as to determine priorities; and subsequently inputting the target sub-cell (or, sector) information to the neighbor list according to the priorities; as described in paragraph 18. The cited document describes general state of the art. This method is similar to above mentioned method (i) .

Document PCT/US2009/041297 (CATOVIC, A. et all) filed on April 21, 2009 and published as WO 2009/132034 Al teaches about systems and methods that facilitate management of automatic neighbor relation functions in wireless networks. The system can include components and/or devices that ascertain whether or not to add or remove a neighbor relation based on information associated with an operations and management system, wherein the operations and management system dispatches add or remove requests to a base station that establishes, updates, and/or maintains a neighbor relations table and/or set of neighbor relations that includes neighbor relations between cells. The cited document describes general state of the art .

As the closest prior art, we identify European patent application published as EP-A-1722588 (KIM, Young-Hun) , assigned to Samsung Electronics Co., Ltd., filed on May 11, 2006 that teaches about hard handover method and Radio Network Controller (RNC) therefore in a mobile telecommunication system. The document discloses information regarding improved performances of an interfrequency hard handover. In cited document the RNC stores information about a timing difference between first and second Node Bs for supporting a soft handover. When receiving a request for a hard handover from the first Node B to the second Node B, the RNC computes information about a timing difference between the Node Bs for the hard handover using the information about the timing difference between the Node Bs stored at a soft handover time. The essential calculation process performed by the "calculator for computing timing difference" is disclosed in paragraphs 0093 and 0094 of EP-A-1722588.

The present invention uses similar calculation process for establishing sub-tree relations to referent cell. Besides the fact that the mentioned prior art in EP-A-1722588 applies on interfrequency, while the present invention applies on intrafrequency handovers, it does not teach about non-trivial features of the present invention that can be summarized below:

(i) synchronization tree formation, i.e. formation of synchronization topological object or objects from various sub-trees by merging;

(ii) simplification procedure of the synchronization topological object;

(iii) mechanism for learning/guessing missing neighbors; and

(iv) fine tuning of the synchronization topological object;

which is performed by the present invention.

Summary of invention

Present invention solves before listed technical problems via method of operating data-processing system for automatic identification of neighbors in WCDMA network, said method comprising steps of: (I) receiving measurement report by a RNC obtained from user equipment serviced by the cell A being characterized by its scrambling code containing reports of some set of cells B_j. characterized by their individual scrambling codes;

(II) detecting the actual transmitting cells of received signals, based on the reported scrambling codes and querying the available monitored set data for cells with the same scrambling code;

(III) forming a sub-tree ST_j where said RNC extracts a corresponding tree relations A (Ci-^U) among all sub-tree ST_j constituents, defined as:

A (Ci->U) = (A (Ci) -A (U) ) for each Ci from the set {A, Bi, B₂, ...} , where U is an arbitrary reference cell from the set {A, Βχ, B₂, ...}, and constructed sub-tree ST_j is being defined by the cell set {A, Bi,..., B_n}j and set of relations A (Ci->U) ;

(IV) merging of any previously formed synchronization topological objects or newly recorded sub-tree STi from step (III) to existed synchronization topological objects; sub-trees, branches, islands or trees; handled by the serving RNC, using data relations obtained from step (III) to form larger synchronization topological objects; wherein said larger synchronization topological objects have each cell listed only once and characterized by used scrambling code and Δ relations with the neighboring cells forming said topological object; by maintaining the values of relations Δ between the said topological cell neighbors constant; and

(V) providing simplification of each synchronization topological object by selecting referent cell U to which each other cells in the synchronization topological object are referenced; and performing summation of all interconnected cell relations A(Ci->C_j) along the connecting path between any desired cell M and the selected referent cell U within the synchronization topological object, using shortest path possible and counting the directions of the inter-relations that influence a +/- sign in the sum; and where A(M->U) is being defined via relation :

Δ(Μ->ϋ) = ((A(M-»Ci) + A(C_!^C₂) + ... + Δ (C_n-»U) ) mod 256 where said steps (I)-(V) are executed independently or as a part of the decision algorithm stored also in said data-processing system applied to data extracted from measurements report or part of the data from the measurement report taken in step (I) .

WCDMA network should be able to perform process of learning, merging and synchronization of topological object hand-in-hand with the identification of missing neighbors. This task is possible to perform via algorithm depicted on Figure 10 and explained in detail in the following text.

Brief Description of the Drawings

Figure 1 represents home cell A to which an UE is connected and where said UE is measuring signal on SC = Y. Figure 2 depicts typical handover position of the UE moving from the serving cell 1 to another cell 2.

Figure 3 shows establishing of the cell relation Δ(Β->Α) as the difference between beginning of next frame in uplink transmission occurred in time T(UE_Tx) and the downlink signal frames began at times (Rx_SFN_Cell_A) and T (Rx_SFN_Cell_B) . Figure 4 shows the subtree formation by referencing cells B and C to the referent cell A.

Figure 5 describes picking the information from various MR reports in time, while Figure 6 describes the way RNC being part of data- processing means can merge various sub-trees into larger topological object. Figure 7 describes a possible way of referencing each cell within synchronization topological object to only one cell of the same object, i.e. cell A.

Figure 8 shows a complex structure obtained from all MR reports received and added to the single synchronization tree in service RNC. Figure 9 shows a closed loop used for fine tuning of some synchronization topological object.

Figure 10 represents the algorithm for learning/guessing missing neighbors, merging and further optimization of the synchronization topological objects by each recorded measurement report.

Detailed Description

It is instructive to review briefly the technical problem once again. Figure 1 represents home cell A to which an UE is connected and where said UE is measuring signal on SC = Y. UE may receive the signal with SC = Y from different directions (from cells B, C, D, ...) . The question under consideration is what is the transmitter of SC = Y?

In a standard solution it is necessary to check the monitored set of cells in the active set on that particular connection, maintained by

RNC and verify if there is any monitored neighbor with mentioned SC

= Y. In absence of such cell(s) the measurement is treated as "missing neighbor". Therefore, neighbor list has to be maintained and optimized constantly in order to keep all relevant neighbors on the list.

Observed problem arises from the mobility request as one of the essential features of any mobile network. Mobility presumes various types of handovers between cells, described and defined in the art (cf. H. Kaaranen, A. Ahtiainen, L. Laitinen, S. Naghian and V. Niemi; "UMTS Networks", 2^nd Edition; 2005, John Wiley & Sons, Ltd.) Standard handover situation is depicted in Figure 2. Mobile UE (3) is moving in direction (4) while connected to RBS A (1) and being within the signal range of RBS B (2) . RBS A (1) is served by the RNC (8) connected to the said RBS A (1) via connection (9) . The UE (3) is permanently measuring downlink signal (6) from the RBS A (1), downlink signal (7) from the RBS B (2), and any other downlink signal that can be detected. The UE (3) reports the signal quality, SC and the frame offset in the form of the MR sent via uplink connection (5) to RBS A (1) ; and RBS A (1) transmits the information via connection (9) to the RNC (8) .

WCDMA network has standardized frame structure that is divided into 15 slots, each of length 2/3 ms and, thus, the total frame length is 10 ms . Based on this, one WCDMA frame is able to handle 38400 chips. System chip rate (SCR) is expressed as 3,84 Mchip/s, and chip is a pulse of a direct-sequence spread spectrum (DSSS) code that is well defined in the art.

So, WCDMA network is based on precise timing and the present invention exploits information which is included in the MR and sent by the UE. Beginning of next frame in uplink transmission occurred in time T(UE_Tx) and the downlink signal frames began at times T (Rx_SFN_Cell_A) and T (Rx_SFN_Cell_B) as shown on Figure 3. Abbreviations Tx and Rx are used in common sense. The MR of signals consist inter alia of particular SCs, signal strength and the time differences observed between SFN (system frame number) and CFN (connection frame number), according to the 3GPP 25.215 specification details.

The SFN-CFN observed time difference Δ to cell Z, Δ(Ζ) expressed in frames, is defined as:

Δ(Ζ) = OFF + Tm(Z)/38400 where Tm(X) is defined as: Τπι(Ζ) = (T(UE_Tx) - TO) - T (Rx_SFN_Cell_Z) with the possible values within the range [0,38399] expressed in chips, and where TO denotes a constant of 1024 chips described in the art. Variable OFF is defined as:

OFF = (SFN_Cell_Z - CFN_Tx) mod 256 representing the offset between the two signals expressed in the number of frames ranging from 0 and 255. CFN_Tx is the connection frame number for the UE transmission of an uplink DPCCH frame at the time T(UE_Tx), while SFN_Cell_Z is the system frame number for the neighbouring P-CCPCH frame received in the UE at the time T (Rx_SFN_ Rx_SFN_Cell_Z) . DPCCH stands for Dedicated Physical Control Channel, and P-CCPCH for Primary Physical Common Control Channel, both being defined in the art.

Referring back to the Figure 3, we are now in position to calculate Δ(Ζ) for each and every signal given by the MR. Furthermore; it is possible to reference other downlink signals to any particular downlink signal used as the reference. We define a value that expresses, in number of frames, earlier signal occurrences of the recorded signals from the cell B to the signal from A:

Δ (B- A) = Δ (B) -Δ (A)

Information Δ(Β->Α) has a value within the range [0,255]. Its value depends on the location of the UE in relation to the transmitters locations, but its integer value should be stabile irrespectively where UE is located. RBSes are synchronized to the network and are also having accurate synchronizing clock. Because of that, the information Δ(Β->Α) is usually changed only during the RBS restart, but the present invention provides the methodology how to address the issue of RBS restart. How this information can be helpful in determination of missing neighbor? Let us suppose at the moment that we have NC number of cells in some WCD A network on the same frequency, and N number of available SCs, where N is evidently less than or equal to 512. The number of candidates for missing neighbor is still NC/N. Therefore all transmitters are classified into 512 classes by their SC code, and theoretically number of candidates under assumption of uniform SC distribution is NC/512.

Introducing reference relations to any potential cell, such as above described relation Δ(Β->Α) where cell B is measured to cell A, gives further characterization of each cell within the network. If only the integer part of Δ(Β->Α) is taken into account, said 256 possible values enlarge further characterization space of cells containing the same SC. A set of cells having same SC is possible to additionally discriminate via reference relations Δ() that leads to enlarged space of 512 x 256 = 131 072 maximum transmitter's characterizations within a WCDMA network. Now, searching for the missing neighbors is narrowed 256 times to NC/ (512*256) .

The whole procedure is possible to carry out in several phases. It is necessary to emphasize that the calculations are to be performed by serving RNC. The mentioned phases are learning, merging and synchronization of formed topological object.

Learning phase is depicted on the Figure 4 by forming sub-tree. Referring to the Figure 3, we may imagine scenario where MR measured by the UE consist of two, three, or even more signals, denoted hereby as signal A, B, C, ...

In the present example (Figure 4-7) we are referencing other signals to the signal received from Cell A. Referencing can be established to any arbitrary cell.

It is possible to establish following relations Δ(Β->Α) and Δ (C- A) . Reference calculations and handling will be conducted by the serving RNC (8) on Figure 2. Evidently each MR recorded by the RNC will contribute in establishing various sub-tree relations, not necessary interconnected, as shown on the Figure 5, i.e. (Δ(Β- Α), Δ (C->A) }_MR1;

(E^D)}_MR2; {A(B- D), A(E^D)}_MR3 etc.

For sake of simplicity we may generalize above approach in three following steps:

STEP I : receiving measurement report obtained from user equipment serviced by the cell A being characterized by its scrambling code containing reports of some set of cells Bi characterized by their individual scrambling codes;

STEP I I : detecting the actual transmitting cells of received signals, based on the reported scrambling codes and querying the available monitored set data for cells with the same scrambling code; and

STEP III : forming a sub-tree ST_j where said RNC extracts a corresponding tree relations Δ( (^-> υ ) among all sub-tree STj constituents, defined as:

A(Ci^U) = (A(Ci)-A(U)) for each Cj from the set {A, Βχ, B₂, ...}, where U is an arbitrary reference cell from the set {A, Βχ, B₂, ...} , and constructed sub-tree ST_j is being defined by the cell set {A, Bi,..., B_n}j and set of relations Δ (Ci- U) .

STEPS I, II and III defined learning procedure about the WCDMA network. Evidently a number of sub-trees recorded in time increases. Amount of data that can be used to merge different synchronization topological objects (sub-trees, branches, islands or trees) also increase in time with the same rate. To understand merging process we revert to special case depicted on Figures 5 and 6. Figure 5 shows the information received by three measurement reports MRl, MR2 and MR3. Evidently it is not possible to find mutual relations between the cells via information picked in the MRl and MR2. However, information picked in MR3 leads to the possibility of merging different synchronization topological objects formed by MRl and MR2 by using the information extracted from MR3 as shown by the Figure 6, all cells being referenced in the example to the cell A.

In time, formed synchronization topological object or more objects grow and become more and more complex, as shown on the Figure 8; where we again use reference cell U as the origin of the topological object.

Such a complex relationships within one synchronization topological object become less useful and less reliable and require optimization .

Optimization process of synchronization topological object presented on Figure 6 is shown in Figure 7. All members of the same topological object are referenced to some particular cell, i.e. for simplicity cell A in the example. Therefore, Δ(ϋ- Α) and Δ (E- A) can be expressed as follows:

A(D- A) = (-Δ (B- D) +Δ (B" A) ) mod 256 and

Δ(Ε- Α) = (Δ (E->D) -Δ (B->D) +Δ (B-»A) ) mod 256 Hereby, we used identity defining the relation direction:

Δ (B- A) = -Δ (A- B)

Mod 256 has been introduced to keep the relation information Δ(Χ) within the prescribed range [0,255], otherwise data are being useless. Mod 256 as used in this specification keeps decimal fractions intact.

The main idea of merging and optimizing of some synchronization topological object is that each cell is characterized by its own scrambling code and reference relation to some referent cell; each cell is being mentioned only once within optimized synchronization topological object.

For sake of simplicity we may generalize above approach of merging and optimizing via two further steps, taking into account before mentioned steps I and II and III;

STEP IV : merging of any previously formed synchronization topological objects or newly recorded sub-tree STi from STE P III to existing synchronization topological objects; sub-trees, branches, islands or trees; handled by the serving RNC, using data relations obtained from step III to form larger synchronization topological objects; wherein said larger synchronization topological objects have each cell listed only once and characterized by used scrambling code and Δ relations with the neighboring cells forming said topological object; by maintaining the values of relations Δ between the said topological cell neighbors constant (please compare cell D relations on Figures 5 and 6 ) ; and

STEP V: providing simplification of each synchronization topological object by selecting referent cell U to which each other cells in the synchronization topological object are referenced; and performing summation of all interconnected cell relations A ( Ci->C_j ) along the connecting path between any desired cell M and the selected referent cell U within the synchronization topological object, using shortest path possible and counting the directions of the inter-relations that influence a +/- sign in the sum; and where Δ(Μ->ϋ) is being defined via relation:

Δ(Μ->ϋ) = ((A(M-^d) + A ( Ci- C₂ ) + ... + Δ ( C_n^U ) ) mod 256 Figure 8 shows a complex structure obtained from various MR reports received and added to the single synchronization tree in service RNC. All cells should be listed, not only from own RNC but also from neighboring RNCs allowing the relations Δ to be uniquely established across the entire WCDMA network. In practice, each RNC will collect information about its own cells and all cells where calls initiated from that RNC may end up.

Being dynamical system, WCDMA network should be able to perform process of learning, merging and synchronization of topological object already described hand-in-hand with the identification of missing neighbors. This task is possible to perform via algorithm depicted on Figure 10 divided hereby in four steps, branching at step 2:

STEP 1 : Let the MR from particular UE to be received by the RNC. Let said MR reports SC Y from some, at the moment unidentified, cell W.

In this step RNC is aware of the active communication performed via service Cell A with the SC X, and about the position of cell A within synchronization topological object.

In the present consideration, let the Cell U is the cell to which all other cells are referenced in the topological object, therefore, the position of cell A in the object is defined as: Δ (A->U) .

As explained in details before, RNC is able to extract frame difference relation Δ (W- A) , and the unknown cell W has been characterized now by the SC Y and with the relation Δ (W- A) . Using Δ(Α- υ) , cell W can be also referenced to the root of the synchronization object U by:

A(W- U) = Round (A(W-»A) + Δ (A- U) ) mod 256 STEP 2 : RNC will consult the list of all available cells in the monitored set and check if there is any cell in monitored set with the scrambling code Y, let us say cell B.

STEP 3 : If cell in STEP 2 exists, the information about the cell and the offset from measurement report will be used in the learning process. Learning process is further divided in two branches by checking the topological position of the cell B (cell ) regarding the cell A;

(i) if RNC determines that cells A and B do not belong to the same topological object, then the relation Δ(Β->Α) is used to merge topological objects having member cell B and A respectively; similar to examples on Figures 5, 6; or

(ii) if RNC determines that cells A and B are belonging to the same topological object, i.e. sub-tree; newly calculated relation Δ (B->A) is used for fine tuning of the synchronization topological object, as is explained bellow.

The phase drift δ may occur between newly recorded relation Δ(Β->Α) and actually obtained relation calculating the closed loop via referent cell U, as shown on Figure 9. Phase drift may occur for various reasons, for instance upon restarting RBS that transmits the cell or similar event that causes that the initial conditions recorded by the RNC were obsolete.

It is possible that cells A and B are not close to the reference cell U and measurement reports - being physical data - may introduce error in frame number counting, or cell being rather unstable in time .

In order to perform fine tuning of some synchronization topological object, we may define phase drift as: δ = (Δ(Β^ϋ) - Δ(Α->ϋ)) - Δ(Β^Α) were Δ(Β- ϋ) and Δ (A- U) are information from synchronization topological object and the Δ (B->A) is from the measurement report, and update previously recorded values {Δ(Β->ϋ), Δ(Α->ϋ)} with the new values {Δ(Β-»ϋ)', Δ (A->U) ' } defined as:

Δ(Β-»ϋ)'= (Δ(Β- υ) - α·δ/2) mod 256

Δ(Α- ϋ)'= (Δ(Α^υ) + α-δ/2) mod 256 where α is the tuning speed factor having values within the range [0.0 - 1.0] .

Smaller a will cause the topological object to update in more iterations. So, a "shifted" cell B, where cells A and B are belonging to the same topological object, will finally be tuned with few iterations that will brought the topological object in optimal operational condition.

STEP 3A : In case that such cell B in STEP 2 with the SC Y does not exist the guessing is performed via following algorithm; RNC will check the synchronization topological object in order to establish whether the cell B' characterized with: and Δ(Β'- ϋ) = A(W->U)= Round (Δ (W" A) + Δ(Α-»ϋ)) mod 256 exists. That will ultimately identify the cell B' within the topological object using Δ() and SC.

If such a cell B' is found, it will be communicated to the handover (HO) function to execute the handover towards the cell B' .

Skilled person in the art should note that the Δ() values in the synchronization topological object are kept as decimal values, while in the guessing stage, the integer part of Δ(ΐΌυ) is compared against the integer values of the Δ() values. With this approach, the accuracy of merging of topological objects is achieved while the guessing algorithm operates efficiently in the integer d^main. Industrial Applicability

As is apparent from the above description, exemplary embodiments of the present invention improve the performance of a handover and increase a handover success probability.

Moreover, exemplary embodiments of the present invention improve the performance of cell recognition not only via SC but also via synchronization topological object based on the introduced Δ relations that enlarge the characterization space within WCDMA network .

Although the exemplary embodiments of the present invention have been disclosed for illustrative purposes, the present invention is not limited to the above described embodiments. Those skilled in the art will appreciate that various modifications, additions, and substitutions are possible, without departing from the scope of the present invention and the full scope of equivalents thereof.

Abbreviations

WCDMA - Wideband Code Division Multiple Access

UE - user equipment

MR - Measurement Report

SC - :scrambling code

RNC - Radio Network Controller

NCELL - neighbour relation

HCell - home cell, first member of the pair NCELL

NCell - neighbouring cell, second member of the pair NCELL

SFN - system frame number

CFN - connection frame number

RBS - radio base station

DPCCH - Dedicated Physical Control Channel

P-CCPCH - Primary Physical Common Control Channel References

1 - RBS A

2 - RBS B

3 - UE

4 - moving direction of UE

5 - uplink transmission

6 - downlink pilot signal to RBS A

7 - downlink pilot signal to RBS B

8 - serving RNC

9 - link between 8 and 1

Claims

1. A method of operating data-processing system for automatic identification of neighbors in WCDMA network where said method comprising steps of:

(I) receiving measurement report by a RNC obtained from user equipment serviced by the cell A being characterized by its scrambling code containing reports of some set of cells Bi characterized by their individual scrambling codes ;

(II) detecting the actual transmitting cells of received signals, based on the reported scrambling codes and querying the available monitored set data for cells with the same scrambling code;

(III) forming a sub-tree ST_j where said RNC extracts a corresponding tree relations

among all sub-tree ST_j constituents, defined as:

A(Ci->U) = (A(Ci)-A(U) ) for each C_L from the set {A, B_if B₂, ...}, where U is an arbitrary reference cell from the set {A, Bi, B₂, ...}, and constructed sub-tree ST_j is being defined by the cell set {A, Bi,..., B_n}_j and set of relations A(Ci->U) ; where A(Cj_.) and Δ(ϋ) represent corresponding SFN-CFN time differences expressed in frames and defined by the 3GPP 25.215 specification; and

(IV) merging of any previously formed synchronization topological objects or newly recorded sub-tree S i from step (III) to existing synchronization topological objects; sub-trees, branches, islands or trees; handled by the serving RNC, using data relations obtained from step (III) to form larger synchronization topological objects; wherein said larger synchronization topological objects have each cell listed only once and characterized by used scrambling code and Δ relations with the neighboring cells forming said topological object; by maintaining the values of relations Δ between the said topological cell neighbors constant; and

(V) providing simplification of each synchronization topological object by selecting referent cell U to which each other cells in the synchronization topological object are referenced; and performing summation of all interconnected cell relations A(Cj_.->C_j) along the connecting path between any desired cell M and the selected referent cell U within the synchronization topological object, using shortest path possible and counting the directions of the inter-relations that influence a +/- sign in the sum; and where Δ(Μ->ϋ) is being defined via relation:

A(M- U) = ((Δ(Μ-»0ι) + A(Ci->C₂) + ... + Δ (C_n" U) ) mod 256 where said steps (I)-(V) are executed independently or as a part of the decision algorithm stored also in said data-processing system applied to data extracted from measurements report or part of the data from the measurement report taken in step (I) .

2. A method as defined in claim 1, wherein the decision algorithm further comprising the steps of:

(A) isolating particular scrambling code Y of unknown cell form the measuring report obtained in step (I) defined in claim 1 where service cell A has scrambling code X, and referencing the cell W to the root of the synchronization object U by: A(W- U) = Round (Δ (W- A) + Δ (A->U) ) mod 256

(B) performing step (II) defined in claim 1 to check if there is any cell B in monitored set with the scrambling code Y.

A method as defined in claim 2, wherein the decision algorithm further executing the following steps in case that the cell B in step (B) exists, identifying cell B as an unknown cell W from the step (A) :

(C) if RNC determines that cells A and B do not belong to the same topological object, then the relation Δ(Β->Α) is used to merge topological objects having member cell B and A respectively; or

(D) if RNC determines that cells A and B are belonging to the same topological object, newly calculated relation Δ(Β->Α) is used for fine tuning of the synchronization topological object.

A method as defined in claim 3, wherein the decision algorithm further executing the following steps in order to perform fine tuning of the synchronization topological object in step (D) :

(E) calculation of the phase drift δ: δ = (Δ(Β- υ ) - Δ(Α->ϋ)) - Δ(Β->Α) were Δ(Β- υ) and Δ(Α->υ ) are information from synchronization topological object and the Δ (B- A) is from the measurement report taken in step (I) defined in claim 1; and

(F) updating previously recorded values { Δ ( Β->υ ) , Δ (Α~>υ)} with the new values {Δ( Β->ϋ)', Δ( Α- ϋ)'} defined as: Δ(Β-»ϋ)'= (Δ(Β- υ ) - ·δ/2) mod 256

Δ(Α->υ ) ' = (Δ(Α-»ϋ) + α·δ/2) mod 256 where α is the tuning speed factor having values within the range [0.0 - 1.0].

A method as defined in claim 2, wherein the decision algorithm further executing the following steps in case that the cell B in step (B) with the scrambling code Y does not exists:

(C) RNC will check the synchronization topological object in order to establish whether the cell B' characterized with :

SC = Y and Round (Δ (B' - U) ) =

= A(W->U)= Round (Δ (W- A) + Δ (A->U) ) mod 256 exists, and

(D) if such a cell B' is found, it will be communicated to the handover function to execute the handover towards the cell B' .

A data-processing system for automatic identification of neighbors in WCDMA network comprising means for carrying out the method of claims 1-5.