WO2016155776A1

WO2016155776A1 - Selection of a serving cell taking into consideration beamforming gains

Info

Publication number: WO2016155776A1
Application number: PCT/EP2015/056907
Authority: WO
Inventors: David ASTELEY; Henrik Asplund; Niklas JALDÉN
Original assignee: Telefonaktiebolaget Lm Ericsson (Publ)
Priority date: 2015-03-30
Filing date: 2015-03-30
Publication date: 2016-10-06

Abstract

The present invention relates to a method for selecting serving cell for a terminal UE in a communication system comprising multiple network nodes 21, 22. Each network node is provided with one or more array antennas, each array antenna is used for communication purposes in one or more cells, and each cell provides coverage in a service area 21a, 22a. The method comprises: transmitting a reference signal 31; 71; 85 for each cell, measuring (in the terminal UE) the reference signal 32 transmitted for at least one cell, and compiling reports 33; 72; 86 indicating power and/or quality of each received reference signal. The method further comprises: estimating the achievable beamforming gains 34; 73, 76; 81-83 for each of the at least one cell, adjusting the reports 35; 77; 88, 88a based on the achievable beamforming gains, and selecting serving cell 78, 78a; 88, 88a for the terminal based upon the adjusted reports.

Description

SELECTION OF A SERVING CELL TAKING INTO CONSIDERATION

BEAMFORMING GAINS

Technical Held

The disclosure relates to a method for selecting serving cell for a terminal in a communication system according to the preamble of claim 1. The invention also relates to a network node in a communication system.

Background

Advanced antenna systems may be used to significantly enhance

performance of wireless communication systems in both uplink (UL) and downlink (DL). In the downlink, there are three basic approaches for utilizing advanced antenna systems: for diversity, multiplexing and beamforming.

With beamforming, the radiation pattern may be controlled by transmitting a signal from a plurality of elements with an element specific gain and phase, also known as element weights. In this way, radiation patterns with different pointing directions and beam widths in both elevation and azimuth

directions may be created.

With so called user equipment (UE) specific beamforming, or terminal specific beamforming, (narrower) beams may be formed to specific users in order to increase the received signal power while at the same time controlling interference generated to other UEs receiving data transmission.

However, UE specific beamforming is not the only form of beamforming. In mobile broadband systems such as LTE, a cell-specific reference signal (CRS) is transmitted for each cell. This signal is used by UEs both for

measurements to select a node to communicate with, as well as a

demodulation reference signal for data to be received by both single and multiple UEs served by the node. Often, the area where a specific cell- specific reference signal is received with highest power (as compared to cell- specific reference signals transmitted from other nodes) is referred to as a cell, and beamforming of the cell-specific reference signal may therefore be referred to as "cell shaping". Since the CRS is aimed for all users in the cell, the cell shape is typically wider than the UE specific beam, and is typically optimized to match the average channel characteristics. Being wider, the peak gain of the cell specific beam is lower than the beam of the UE specific beam.

The process in which a terminal is assigned to a certain cell, out of several possible candidates, is commonly referred to as cell selection. Cell selection is commonly done on a much slower basis than on 1 ms transmission time interval (ΤΤΊ) level. Typically, the terminal (i.e. user equipment) measures reference signal received power (RSRP) and reference signal received quality (RSRQ) on the CRS signals and reports this to the network via the network node managing the current serving cell for the terminal.

When UE specific precoding is used (including spatial multiplexing), the terminal measurement reports of RSRP and RSRQ are not necessarily directly representative of the actual link quality, as data transmission is performed with a different precoding, i.e. using different element weights, over the antenna array compared to the precoding used for transmission of the CRS signal. This typically gives higher antenna gain during data transmission compared to the CRS transmissions. Cell selection, as of today, based on such reports will lead to degraded link and system performance since the terminal may not always be served by the cell providing the best link for data transmissions. This is especially true when the different involved cells have different UE-specific beamforming capabilities, e.g. are using different numbers of transmit antennas.

Furthermore, terminals may not know which precoders that are used in a certain candidate cell, nor the number of antennas available for terminal specific beamforming at the site, and can hence not estimate the achievable precoding gain to adjust the measurements reports for selection of best serving cell accordingly. Further, it is possible to restrict the set of available precoders in a cell, so-called codebook subset restriction, e.g. in order to avoid creating interference in undesirable directions. This restriction may not be known to a terminal not connected to this particular cell.

In LTE Rel- 1 1 and Rel- 12 there are solutions based on CSI-RS signaling that could be used in order to get terminal reports including multiple antenna benefits. However, for Rel- 1 1 multiple CSI-processes is a UE capability and for Rel- 12 there may be requirements on network synchronization. Since UE specific precoding may benefit all LTE terminals (for all releases), solutions are needed also for legacy terminals (Rel. 8, 9, and 10) as well as for non- synchronized networks.

Summary

One object with the present invention is to provide a method for selecting a serving cell for a terminal in a communication system that improves the quality of the data communication link compared to prior art techniques. The object may be achieved with a method for selecting serving cell for a terminal (UE) in a communication system comprising multiple network nodes. Each network node is provided with one or more array antennas, each array antenna is used for communication purposes in one or more cells, and each cell provides coverage in a service area. The method

comprises: transmitting a reference signal for each cell,

in the terminal (UE), measuring the reference signal transmitted for at least one cell,

compiling reports indicating power and/ or quality of each received reference signal,

estimating the achievable beamforming gains for each of the at least one cell,

adjusting the reports based on the achievable beamforming gains, and selecting serving cell for the terminal based upon the adjusted reports.

The object may also be achieved by a network node for selecting serving cell for a terminal (UE) in a communication system comprising multiple network nodes. Each network node is provided with one or more array antennas, each array antenna is used for communication purposes in one or more cells, and each cell provides coverage in a service area. The network node is configured to: transmit a reference signal for each cell,

- receive reports from the terminal indicating power and / or quality of the reference signal received for at least one cell, said terminal is configured to measure the reference signal transmitted for the at least one cell and compile said reports,

estimate and/ or collect achievable beamforming gains for each of the at least one cell,

adjust the reports based on the achievable beamforming gains, and select serving cell for the terminal based on the adjusted reports.

An advantage with the present invention is that variations, and also limitations, in the data communication link can be anticipated and taken into account when selecting serving cell for a terminal.

Another advantage with the present invention is that it may be implemented in synchronized as well as non- synchronized communication networks.

Further objects and advantages will be apparent for a skilled person from the detailed description and the drawings. Brief description of the drawings

Fig. 1 illustrates cell selection according to prior art.

Fig. 2 illustrates cell selection using the method according to the invention. Fig. 3 shows a flow chart embodying the method of the invention.

Fig. 4 shows a flow chart of an embodiment for determining beamforming gains for each cell.

Fig. 5 shows a flow chart of an embodiment for adjusting reports based on achievable beamforming gains.

Fig. 6 shows a flow chart of an embodiment for determining terminal specific beamforming gains.

Fig. 7 shows a first example for selecting serving cell in a communication system. Fig. 8 shows a second example for selecting serving cell in a communication system.

Fig.9 shows a generic apparatus of the invention. Detailed description

The invention will be described in relation to LTE communication networks, but may be implemented in other types of communication systems, that use a reference signal for each cell to determine the channel quality for data transmission for a terminal in the cell. The channel quality for data

transmission is determined by estimating the achievable beamforming gains in an LTE communication network, but the concept is applicable to any other system where the cell selection is performed on information (of beamforming type) than what is later used for data transmissions.

A cell-specific offset (CSO) has been introduced to allow load balancing in the network by allowing a weaker, but less loaded, cell to serve some users that would otherwise be served by a stronger but possibly overloaded cell.

Recently, the use of CSO has also been proposed to improve situations with cells with different transmit powers, aka Heterogeneous Networks. The CSO can be used to increase the uptake of small cells with lower transmit powers that may, especially on the uplink, be better suited to serve some users than a more high power macrocell.

The method described in connection with figures 3-9 is based on the concept that transmitted reference signal(s), cell-specific reference signal(s) in LTE, are measured in each terminal and reports indicative of received power and/ or received quality is generated in the terminal. These reports are adjusted, e.g. by adaptively adjusting the cell-specific offset of the cells, based on estimates of achievable UE specific beamforming gains (compared to the cell specific beamforming used when transmitting the cell-specific reference signal) in order to improve the link rate on data transmissions.

The invention may be applied in LTE communication networks. CQI

(Channel Quality Indication) report is an important element of LTE and has significant impact on the system performance. There are two types of CQI report in LTE: periodic and aperiodic. The periodic CQI report is carried by PUCCH. But if the UE needs to send UL data in the same subframe as the scheduled periodic CQI report, the periodic CQI report will use the PUSCH, together with UL data transmission. This is because a UE can't transmit on both PUCCH and PUSCH simultaneously. In this case, the periodic PUCCH resource will be idle. Since periodic CQI report brings in the "always on" signaling overhead, the report granularity is relatively rough. In order to get more detail CQI report, the eNB can trigger aperiodic CQI report when needed. The aperiodic CQI report is transmitted on PUSCH, together with UL data or alone.

SINR is a measure of signal quality as well but it is not defined in the 3 GPP specs but defined by the UE vendor. It is not reported to the network. SINR is primarily used by operators, and the LTE industry in general, as it better quantifies the relationship between RF conditions and Throughput. UEs typically use SINR to calculate the CQI (Channel Quality Indicator) they report to the network. It is a common practice to use Signal-to-Interference Ratio (SINR) as an indicator for network quality. It should be however noted that 3GPP specifications do not define SINR and therefore UE does not report SINR to the network. SINR is still internally measured by most UEs and recorded by drive test tools.

Figure 1 illustrates cell selection in a communication system 10 according to prior art. The communication system comprises in this example two network nodes 1 1 and 12, each provided with an array antenna (not shown) used for communication purposes in a respective cell. Each cell is configured to provide coverage in a service area 1 1a and 12a. Four terminals UE1 -UE4 are present in the service areas covered by the array antennas of the network nodes 1 1 and 12.

The cell selection according to prior art comprises transmitting a reference signal for each cell, which reference signals are configured to be detected by any terminal within the coverage area of each respective transmitted reference signal. Thereafter, in each terminal UE1 -UE4, the transmitted reference signal for at least one cell is measured. Normally each terminal detects multiple reference signals, with the purpose to select one of the cells as serving cell. Reports are compiled in each terminal indicating power and/ or quality of each received reference signal, and the serving cell for each terminal is selected based upon the reports compiled by each respective terminal. This normally occurs in the network node providing coverage to the terminal in consensus with other network nodes associated with the compiled reports, as indicated by the arrow 17. For communication

purposes, separate beams 13- 16 are used for data transmission with each respective terminal UE1 -UE4.

In an LTE communication network, each of the terminals UE1 -UE4

measures a cell-specific reference signal (CRS) transmitted for each of the two cells and compiles reports indicating power (in LTE called RSRP) and/ or quality (in LTE called RSRQ) that are sent to the respective serving cells for consideration by the network when determining which terminals should be served by which cells. In the state of art cell selection each terminal would be served by the cell from which it detects the reference signal with the highest quality (in LTE the highest RSRP and/ or RSRQ) unless the transmit powers of the two cells are different, or one of the cells is overloaded.

As an example, a first reference signal is transmitted in the service area 1 1a of a first cell, which is detected by UE1 , UE2, UE3 and UE4. A second reference signal is transmitted in the service area 12a of a second cell, which is detected by UE1 , UE2, UE3 and UE4. UE1 only measure the first reference signal, since the power of the second reference signal is below a predetermined level, and the compiled report(s) regarding quality is sent to the network node 1 1 providing cell coverage for terminal UE1. The selection of serving cell is trivial since the report(s) only include a quality indication for the current serving cell. On the other hand, terminals UE2, UE3 and UE4 compile reports regarding power and / or quality in each terminal based on both the first and second reference signal. UE2 and UE3 send their reports to network node 1 1 and UE4 sends its reports to network node 12, and the selection of serving cell is based on the highest RSRP and/ or RSRQ. In this example UE1 uses beam 13, UE2 uses beam 14 and UE3 uses beam 15 for data transmission in service area 1 1a; and UE4 uses beam 16 for data transmission in service area 12a.

However, the prior art method has a drawback when one or both cells are configured for terminal specific beamforming. An improvement in achievable beamforming gain for such a cell is expected for data transmission for each terminal compared to when transmitting the reference signal. This difference can be due to the used transmission mode or precoding codebook,

environmental impact such as angular spread, or even different terminal capabilities. The invention relates to determining estimates of the base stations' achievable beamforming gains for each cell, and using these estimates to adjust the reports (such as RSRP and/ or RSRQ) from the terminals used in the cell selection process, or using the estimates to adjust the cell threshold with a cell-specific offset. The estimated received signal power (RSRP adjusted with achievable beamforming gains) may be determined in a network node eNB, and communicated to some central network node, or between separate eNB to decide on which node is best for serving each terminal. Figure 2 illustrates cell selection in a communication system 20

implementing a method according to the invention. The communication system 20 comprises in this example two network nodes 21 and 22, each provided with an array antenna (not shown) used for communication purposes in a first and second cell, respectively. The first node 21 is provided with an array antenna having X antenna elements and the second node 22 is provided with an array antenna having Y antenna elements. In this example the number of antenna elements in the second array antenna is larger than the number of antenna elements in the first array antenna, i.e. Y>X and thus the beamforming gain in the second cell is higher than the beamforming gain in the first cell. The uptake for the second cell (more antenna elements) is larger at the expense of the uptake for the first cell (less antenna elements), and an array antenna with more antenna elements can therefore generate longer and narrower beams compared to an array antenna with less antenna elements. The first cell is configured to provide coverage in a service area 21a and the second cell is configured to generate coverage in a service area 22a. Four terminals UE1 -UE4 are present in the service areas covered by the array antennas of the network nodes 21 and 22.

The prior art cell selection (described in connection with figure 1) would not associate any cell selection offsets with respect to additional beamforming gain expected for data transmission (payload data) on either of the cells. Each reference signal is transmitted using a set of beamforming weights from each array antenna, and a different set of weights is used for terminal specific downlink data transmission. This difference may be used to estimate the achievable beamforming gains for each cell and the reports compiled from the received reference signal may be adjusted to take this into

consideration.

The inventive method comprises:

1) Transmitting a reference signal for each cell, which reference signals are configured to be detected by any terminal within the coverage area of each respective transmitted reference signal.

2) In each terminal UE1 -UE4, the transmitted reference signal for at least one cell is measured. Normally each terminal detects multiple reference signals, with the purpose to select one of the cells as serving cell. 3) Reports indicating power and/ or quality of each received reference signal are compiled in each terminal.

4) Estimate the achievable beamforming gains for each of said at least one cell. This can be performed in different ways as illustrated in connection with figures 3, 4 and 6. 5) The reports are adjusted based on the achievable beamforming gains to create a cell selection offset that can be used to select the most appropriate cell for each terminal.

6) Selecting serving cell for the terminal based upon the adjusted reports. This is performed by the network, and may be performed in the network node providing coverage to the terminal in consensus with other network nodes associated with the compiled reports, as indicated by the arrow 27, or by a central node 28 (which may collect the reports from each terminal and adjust the reports, or may collect adjusted reports from the network node providing coverage to the terminal, before making a selection of serving cell for the terminal). For communication purposes, separate beams 23- 26 are used for data transmission with each respective terminal UE1-UE4.

Similarly to the example in figure 1 , a first reference signal is transmitted in the service area 21a of a first cell, which is detected by UEl , UE2, UE3 and UE4. A second reference signal is transmitted in the service area 22a of a second cell, which is detected by UEl, UE2, UE3 and UE4.

UEl only measure the first reference signal, since the power of the second reference signal is below a predetermined level, and the compiled report(s) regarding quality is sent to the network node 21 providing cell coverage for the terminal. The selection of serving cell is trivial since the report(s) only include a quality indication for the current serving cell.

On the other hand, terminals UE2, UE3 and UE4 compile reports regarding power and / or quality in each terminal based on both the first and second reference signal. The achievable beamforming gains for each cell are estimated (as described in connection with figure 3, 4 and 6) and is used to adjust the reports compiled in the terminals.

As mentioned above, the adjustment of the reports may be performed in the network node providing coverage for the terminal, i.e. network node 21 for UEl , UE2 and UE3; and network node 22 for UE4, or in the central node 28 for all terminals UE1-UE4. UE2 and UE3 send their reports to network node 21 and UE4 sends its reports to network node 22, and the selection of serving cell is based on the adjusted reports that indicate the channel quality for data transmission for each terminal. The selection of serving cell may be performed in the network node providing coverage to the terminal in consensus with other network nodes associated with the compiled reports, or may be performed in the central network node 28.

In addition the trigger point for transmission of reports from each terminal to the network node providing coverage to the terminal may be changed. In order for this to work, a cell-specific offset CSO is determined based on the beamforming gains for each cell and the cell-specific offset CSO is

communicated to each terminal in addition to the reference signal for each cell. Each terminal is configured to receive the cell-specific offset CSO and reference signal for each cell in the respective service area. For instance, if UEl receives a cell-specific offset related to the second cell provided by network node 22, UEl may compile reports regarding power and/or quality for the second reference signal if the CSO triggers UEl , and thereafter transmit the reports to network node 21. In this example UEl uses beam 23 and UE3 uses beam 25 for data

transmission in service area 21a; and UE2 uses beam 24 and UE4 uses beam 26 for data transmission in service area 22a. The channel quality for data transmission is estimated to be higher for UE2 when connected to the serving cell covering service area 22a compared to using beam 14 for communication purposes as described in the prior art example in figure 1.

One reason for this may be that the array antenna of the second node 22 is using more antenna elements (and therefore can generate longer and narrower beams) compared to the array antenna of the first node 21 , which may be provided with less antenna elements than the array antenna of the second node 22. The characteristics of the array antennas may also be different that may have an impact on the beamforming gain of each

respective cell.

Figure 3 shows a flow chart embodying a method for selecting serving cell for a terminal (UE) in a communication system comprising multiple network nodes as described in connection with figure 2. Each network node is provided with one or more array antennas, and each array antenna is used for communication purposes in one or more cells, each cell provides coverage in a service area. Thus when implementing a method for selecting serving cell for a terminal in a communication system comprising multiple network nodes, the network node is configured to: transmit a reference signal for each cell,

- receive reports from a terminal indicating power and / or quality of the reference signal received for at least one cell, the terminal is configured to measure the reference signal transmitted for the at least one cell and compile the reports,

estimate the beamforming gains for one or more cells associated with the network node; and collect achievable beamforming gains for each of said at least one cell not associated with the network node,

The network node may further be configured to: - identify other network nodes associated with the compiled reports, and

select the serving cell for the terminal based on the adjusted reports in consensus with the identified other network nodes.

In addition, the network node may be configured to determine terminal (UE) specific beamforming gains, and also use the terminal specific beamforming gains to adjust the reports.

The method is initiated in 30 and comprises the following steps:

31 : transmitting a reference signal for each cell, i.e. the array antenna configured to be used for communication purposes in a cell transmits a non- common reference signal in that cell (in LTE the non-common reference signal is called cell-specific reference signal, CRS).

32: in the terminal (UE), measuring the reference signal for at least one cell. The terminal has to be able to receive at least one reference signal to be able to communicate with the network, and it should be noted that each terminal will typically only be able to receive reference signals transmitted in the network from neighboring cells.

33: compiling reports indicating power and/ or quality of each received reference signal. Thus, separate reports for each cell are compiled in the terminal. These reports may be transmitted to the network node providing coverage for the terminal, or may be transmitted to a central node in the communication network.

34: estimating the achievable beamforming gains for each of the at least one cell. This step may be performed in the network node provided with the array antenna that is used for communication purposes in each cell. The

achievable beamforming gains for each cell may be distributed between the network nodes (also including a central node, if present) in the

communication system in order to provide information to perform the next step of the method.

The knowledge of beamforming gains may also be used to determining a cell- specific offset CSO for each cell. The CSO may be communicated for each cell to the terminal (UE), and, in the terminal (UE), consider the

communicated cell-specific offset to decide if to send the reports to the network node providing cell coverage for the terminal (UE) .

35: adjusting the reports based on the achievable beamforming gains. The adjustments may be performed in the network node providing coverage to the terminal (with knowledge of achievable beamforming gains for the other network nodes associated with the compiled reports from step 33), or the adjustments may be performed in a central node (with knowledge of achievable beamforming gains for the other network nodes associated with the compiled reports from step 33), see description in connection with figure 5. 36: selecting serving cell for the terminal based upon the adjusted reports. This step may be performed in a central node in the communication system. It should be appreciated that the central node can be a distributed network node, virtual network node or a single dedicated network node. However, it is also possible to implement this step locally and let the network node that provide coverage to the terminal select the serving cell in consensus with the other network nodes that are associated with the compiled reports. The selection process also includes notifying the network nodes of the selection irrespectively of where the decision has been made regarding serving cell for the terminal.

Figure 4 shows a flow chart of an embodiment for estimating achievable beamforming gains (step 34 in figure 3) for each of the at least one cell further comprises determining an average beamforming gain in the service area for each of said at least one cell. This may be achieved by empirical mapping or not, which is decided in step 40. If empirical mapping is not used then the flow continues to step 42 wherein the average beamforming gain may be determined based on pre-calculated (or planned) expected average beamforming gain on cell level.

The pre-calculated expected average beamforming gain may be based on uplink (UL) channel measurements for the array antenna used to provide coverage to each of said at least one cells. This can be based on the number of antenna subarrays used and their configuration and / or the

characteristics of the array antenna. A subarray comprises one or more antenna elements connected by a distribution network. This is a rather straightforward approach based on: the observation that the theoretical antenna gain of an antenna array is proportional to the number of transmit antenna subarrays, and aggregating UL channel measurements and knowledge about available set of beams for data communication and cell specific beam used to transmit the reference signal. These UL channel measurements can be channel estimates over multiple subarrays in an antenna array.

These channel estimates may be combined with the possible terminal- specific downlink (DL) precoders as well as the beamforming weights used for the reference signal, forming estimates of the achievable beamforming gain as well as the possible terminal- specific beamforming gain. The difference between these two estimates determines a cell selection offset in comparison with a hypothetical reference cell having no terminal- specific beamforming. Thus, the reference signal is transmitted using a set of beamforming weights from each array antenna (i.e. using a cell specific beam), and the pre- calculated average beamforming gain may further be based on: set of available weights for beamforming for terminal specific downlink data transmission, and

- the set of beamforming weights used for transmitting the reference signal.

On the other hand if empirical mapping is selected to be used in step 40, the flow continues to step 41 , wherein the step of determining the average beamforming gain may be based on an empirical mapping from reports regarding power to channel quality.

The empirical mapping may be determined by aggregating terminal reports regarding power and quality (in LTE communication networks: RSRP and CQI).

The RSRP reports quantify the received signal power for the cell- specific reference signals (CRSs) while the CQI reports quantify the channel quality for the data transmission. Thus, two cells with identical cell-specific reference transmissions but different terminal- specific beamforming of data transmissions may determine different mappings between the two measures. The offset in these mapping will be a measure of the cell selection offset between the two cells.

The flow thereafter continues from 41 and 42 to step 35, figure 3.

In addition to what has been described above, the step of estimating achievable beamforming gains for the at least one cell may comprise determining beamforming gains in multiple angular segments for each of the at least one cells, and adjust the reports based on the achievable

beamforming gains in the angular segment where the terminal is located.

The beamforming gains may be determined by estimating average

beamforming gains over each angular segment for each cell, and the average beamforming gain may comprise creating spatial channel characteristics based on reports regarding power over the angular segments, and

aggregating terminal- specific measurements over said spatial channel characteristics to identify in which angular segment the terminal is located. Note that the average beamforming gain, described above, may be divided into smaller sectors for more detailed mapping. It could for example be such that the average beamforming gain is higher in a few specific areas / angular segments due to for example environmental effects such as angular spread. Thus a direction of arrival estimate together with an average beamforming gain may provide additional gains to the cell average.

The cell selection offsets determined using any of the embodiments above may be stored for future reference on a cell-by-cell basis in e.g. the

corresponding eNB, or aggregated in a central network node. Whenever the cell association of a terminal is determined, the applicable cell selection offsets may be retrieved from the eNBs or the central node and applied to the measurements reported by the terminal.

Other possible embodiments are aggregating information on results of handovers triggered by based beamforming gain offsets. Such information must then be signaled between eNBs or from centralized system nodes. Such information may provide additional robustness on average cell selection offset values, either on average cell level or on cell portion level.

Figure 5 shows a flow chart of an embodiment for adjusting and triggering reports based on achievable beamforming gains (step 35 and 36 in figure 3) for each of the cells, or each segment of the cells. If a change in level for triggering reports compiled in a terminal is used, then a decision is made in step 50 and the flow continues to step 51 in which a cell-specific offset CSO is determined based on the achievable beamforming gains for each cell. The cell-specific offset CSO is communicated to the terminal for the at least one cell in step 52, and in the terminal UE, considering the communicated cell- specific offset to decide if and when to send the reports to a network node providing cell coverage for the terminal. The flow thereafter continues to step 54. On the other hand, if the level for triggering reports compiled in a terminal is not changed the flow continues directly to step 54 in which the reports are sent from the terminal UE to a network node within said communication system. In step 55, the estimated achievable beamforming gains for the at least one cell is shared with the network node and the reports (e.g. RSRP and / or RSRQ in LTE) are adjusted in the network node based on the achievable beamforming gains for each cell. Note that the network node may be the network node providing coverage for the terminal or a central node as described below.

If the adjustment of the reports was made in a central node, step 56, the flow continues to step 57 and the selection of serving cell is made in the central node configured to select the serving cell for the terminal based on the adjusted reports. On the other hand, if the adjustment of the reports was made in network node providing coverage to the terminal the flow continues to step 58. The selection of serving cell may in this case be performed in the central node or in the network node providing coverage to the terminal. If a decision is made in step 58 to make the selection in the central node, the adjusted reports are communicated from the network node to the central network node, which is configured to select the serving cell for the terminal based on the adjusted reports received from the network node providing coverage for the terminal.

However, if a decision is made in 58 to select serving cell in the network node providing coverage to the terminal, the flow continues to step 59, where other network nodes associated with the compiled reports are identified. The serving cell is selected for the terminal UE in the network node based on the adjusted reports in consensus with the identified other network nodes.

Figure 6 shows a flow chart of an embodiment for determining terminal specific beamforming gains in addition to the step of estimating achievable beamforming gains for each of the at least one cell (step 34 in figure 3). The reference signal for each cell is transmitted using a set of beamforming weights from each array antenna, and terminal specific beamforming gains may be determined by:

60: triggering uplink transmission from each terminal. This may be

implemented in LTE using RACH, PUSCH, PUCCH and/ or aperiodic/ periodic SRS.

61 : configuring network nodes to listen to the uplink transmissions.

61 : for each cell, estimating the terminal specific beamforming gains based on:

an uplink channel estimate,

the set of beamforming weights for reference signal, and

set of available weights for beamforming for data transmission.

Figure 7 shows a first example 70 for selecting serving cell in a

communication system. This example comprises three network nodes A, B and C, one terminal UE and an optional central node (optional components and functionalities are indicated by dashed lines). Each node is provided with an array antenna that provides coverage in a service area and is used for communication purposes in a respective cell A, B and C.

71 : a reference signal is transmitted for each cell, CRS A, CRS B and CRS C. The arrows illustrate that the reference signal from Node A and node B are received above a predetermined level at the terminal UE. The reference signal CRS C from node C is weak and the terminal will therefore not compile a report for cell C.

72: reports for cell A and cell B (Report A, Report B) are compiled indicating power and/ or quality of each received reference signal.

73: Node B is estimating its achievable beamforming gains for cell B. This is also performed in Node C, but since the terminal does not compile any report for cell C, this has been omitted in the figure.

74: The knowledge of the achievable beamforming gains for cell B is shared with Node A, which collects information regarding beamforming gains for neighboring cells associated with the reports transmitted from the terminal in 75.

75: Report A and report B are transmitted to the network node providing coverage to the terminal (in this example Node A). 76: The achievable beamforming gains is estimated for cell A, since Node A has to be informed of the achievable beamforming gains for cell A and cell B to make necessary adjustments of the reports received from the terminal UE.

77: The reports A and B are adjusted based on the achievable beamforming gains for cell A and cell B. In one embodiment the selection of serving cell may be performed in the network node providing coverage to the terminal, or in an alternative embodiment the selection of serving cell may be performed in a central node (dashed lines).

78: The serving cell is selected in the network node that provides coverage to the terminal, i.e. Node A. 79: The other nodes associated with the reports (i.e. Node B) and the terminal are notified regarding the selection. Cell A or Cell B may be selected as serving cell and if the serving cell is changed, hand-over has to be performed.

77a: Alternatively, the adjusted reports are sent to the central node for selecting serving cell.

78a: The serving cell is selected in the central node.

79a: The nodes associated with the reports (i.e. Node A and B) and the terminal are notified regarding the selection. Cell A or Cell B may be selected as serving cell for the terminal UE and if the serving cell is changed, hand- over has to be performed.

Figure 8 shows a second example 80 for selecting serving cell in a

81 : Achievable beamforming gains are estimated for cell A.

82: Achievable beamforming gains are estimated for cell B.

83: Achievable beamforming gains are estimated for cell C. 84: The knowledge of beamforming gains for cell B and C is shared with network node A that provides coverage to the terminal UE. The beamforming gains from all cells (A-C) may also be shared with a central node, if present.

85: a reference signal is transmitted for each cell, CRS A, CRS B and CRS C together with a cell-specific offset for the cells, CSO A- C. The arrows illustrate that the reference signal from Node A, Node B and Node C are received above a predetermined level at the terminal UE when taking the cell-specific offset into consideration. The reference signal CRS C from node C is weak, but the CSO for cell C triggers the terminal to compile a report for cell C.

86: reports for cell A, cell B and cell C (Report A, Report B, Report C) are compiled indicating power and/ or quality of each received reference signal.

87: Reports A, B and C are transmitted to the network node providing coverage to the terminal (in this example Node A). In one embodiment the adjustment of the reports and the selection of serving cell may be performed in the network node providing coverage to the terminal, or in an alternative embodiment the adjustment of the reports and selection of serving cell may be performed in a central node (dashed lines).

88: The reports A, B and C are adjusted based on the achievable

beamforming gains for cell A, cell B and cell C, and the serving cell is selected in the network node that provides coverage to the terminal, i.e. Node A.

89: The other nodes associated with the reports (i.e. Node B and Node C) and the terminal are notified regarding the selection if necessary. Cell A, Cell B or Cell C may be selected as serving cell and if the serving cell is changed, hand-over has to be performed. 87a: Alternatively, the reports and knowledge of the beamforming gains for cell A-C (if not already shared with the central node) are sent to the central node.

88a: The reports A, B and C are adjusted based on the achievable

beamforming gains for cells A-C that previously was sent to the central node, and the serving cell is selected in the central node based on the adjusted reports.

89a: The nodes associated with the reports (i.e. Node A, B and C) and the terminal are notified regarding the selection if necessary. Cell A, Cell B and Cell C may be selected as serving cell for the terminal UE and if the serving cell is changed, hand-over has to be performed.

Fig.9 shows a generic apparatus 90 of the invention configured to be used for selecting serving cell for a terminal in a communication system

comprising multiple network nodes. Each network node is provided with one or more array antennas, each array antenna is used for communication purposes in one or more cells, and each cell provides coverage in a service area.

Each network node is configured to transmit a reference signal for each cell 91 , and the terminal is configured to measure the transmitted reference signal for at least one cell 92 and compile reports indicating power and/ or quality of each received reference signal 93.

The generic apparatus comprises:

A first module 91 configured to transmit a reference signal for each cell.

A second module 92 configured to measure, in the terminal, the reference signal for at least one cell.

A third module 93 configured to compile reports indicating power and/ or quality of the reference signal received for the at least one cell, A fourth module 94 for estimating the achievable beamforming gains for the at least one cell.

A fifth module 95 for adjusting the reports based on the achievable beamforming gains. A sixth module 96 for selecting serving cell for the terminal based upon the adjusted reports.

Abbreviations

CRS - Cell-specific Reference Signal

CSI - Channel State Information CSI-RS - Channel State Information - Reference Signal

CSO - Cell-Specific Offset

CQI - Channel Quality Indicator

DL - DownLink eNB - evolved Node B LTE - Long Term Evolution

PUCCH - Physical Uplink Control Channel

PUSCH - Physical Uplink Shared Channel

RACH - Random Access Channel

RSRP - Reference Signal Received Power RSRQ - Reference Signal Received Quality

SRS - Sounding Reference Signal

TTI - Transmission Time Interval UE - User Equipment UL - UpLink

Claims

1. A method for selecting serving cell for a terminal (UE) in a

communication system comprising multiple network nodes (21 , 22), each network node is provided with one or more array antennas, each array antenna is used for communication purposes in one or more cells, each cell provides coverage in a service area (21a, 22a), said method comprises: transmitting a reference signal (31 ; 71 ; 85) for each cell,

in the terminal (UE), measuring the reference signal (32) transmitted for at least one cell, and

- compiling reports (33; 72; 86) indicating power and/or quality of each received reference signal,

wherein said method further comprises:

estimating the achievable beamforming gains (34; 73,76; 81-83) for each of said at least one cell,

- adjusting the reports (35; 77; 88, 88a) based on the achievable beamforming gains, and

selecting serving cell (78, 78a; 88, 88a) for the terminal based upon the adjusted reports.

2. The method according to claim 1 , further comprises: - determining a cell-specific offset (51) based on the achievable beamforming gains for each cell,

communicating the cell-specific offset (52; 85) for each cell to said terminal (UE), and

in the terminal (UE), considering the communicated cell-specific offset (53) to decide if to send reports to a network node providing cell coverage for the terminal (UE).

3. The method according to claim 1 or 2, further comprises: sending said reports (54; 75; 87) from the terminal (UE) to a network node within said communication system,

sharing the estimated achievable beamforming gains (74; 84, 87a) for said at least one cell with the network node, and

in said network node, adjusting the reports (55; 77; 88, 88a) based on the achievable beamforming gains for each cell.

4. The method according to claim 3, wherein the method further comprises: - selecting said network node to be the network node providing cell coverage for the terminal (UE).

5. The method according to claim 4, wherein the method further comprises: communicating the adjusted reports from said network node to a central network node (77a) configured to select the serving cell for the terminal based on the adjusted reports from said network node.

6. The method according to any of claims 2-5, wherein the method further comprises: identifying other network nodes associated with the compiled reports, and

selecting the serving cell (78) for the terminal (UE) in the network node based on the adjusted reports in consensus with the identified other network nodes.

7. The method according to any of claims 1-6, wherein said step of estimating the achievable beamforming gains for each of said at least one cell further comprises: determining an average beamforming gain in the service area for each of said at least one cell.

8. The method according to claim 7, wherein said step of determining the average beamforming gain is based on pre-calculated expected average beamforming gain (42).

9. The method according to claim 8, wherein said pre-calculated expected average beamforming gain is based on the number of antenna subarrays and / or the characteristics of the array antenna used to provide coverage to each of said at least one cells.

10. The method according to claim 8, wherein said pre-calculated expected average beamforming gain is based on uplink channel

measurements for the array antenna providing coverage to each of said at least one cells.

1 1. The method according to claim 10, wherein the reference signal is transmitted using a set of beamforming weights from each array antenna, and said pre-calculated average beamforming gain further is based on set of available weights for beamforming for terminal specific downlink data transmission and said set of beamforming weights used for transmitting the reference signal.

12. The method according to claim 6, wherein said step of determining the average beamforming gain is based on an empirical mapping from reports regarding power to channel quality (41).

13. The method according to claim 12, wherein said empirical mapping is determined by aggregating terminal reports regarding power and quality for data transmission.

14. The method according to any of claims 1 - 13, wherein said step of estimating the achievable beamforming gains for said at least one cell further comprises: determining beamforming gains in multiple angular segments for each of said at least one cells, and adjust the reports based on the

achievable beamforming gains in the angular segment where the terminal is located.

15. The method according to claim 14, wherein each of said at least one cells is divided into multiple angular segments, said step of determining beamforming gains in angular segments comprises: estimating average beamforming gains over each angular segment for each cell.

16. The method according to claim 15, wherein the step of estimating average beamforming gain over each angular segment comprises: creating spatial channel characteristics based on reports regarding power over the angular segments, and aggregating terminal- specific measurements over said spatial channel characteristics to identify in which angular segment the terminal is located.

17. The method according to any of claims 1 - 16, wherein said step of estimating the achievable beamforming gains for said at least one cell further comprises:

determining terminal specific beamforming gains.

18. The method according to claim 17, wherein the reference signal is transmitted using a set of beamforming weights from each array antenna, said terminal specific beamforming gains is determined by: triggering uplink transmission (60) from the terminal, and

configuring network nodes to listen to the uplink transmissions (61), and

for each cell, estimating the terminal specific beamforming gains (62) based on an uplink channel estimate, said set of beamforming weights for reference signal, and set of available weights for beamforming for data transmission.

19. The method according to any of claims 1 - 18, wherein communication system is an LTE communication network and the reference signal is selected to be a cell-specific reference signal (CRS).

20. A network node (21 , 22) for selecting serving cell for a terminal (UE) in a communication system (20) comprising multiple network nodes, each network node is provided with one or more array antennas, each array antenna is used for communication purposes in one or more cells, each cell provides coverage in a service area (21a, 22a), said network node is

configured to: transmit a reference signal (71) for each cell, receive reports (75) from the terminal (UE) indicating power and/ or quality of the reference signal received for at least one cell, said terminal is configured to measure the reference signal transmitted for said at least one cell and compile said reports, estimate (76) and/ or collect (74) achievable beamforming gains for each of said at least one cell, adjust the reports (77) based on the achievable beamforming gains, and select serving cell (78) for the terminal based on the adjusted reports.

21. The network node according to claim 20, wherein said network node further is configured to: identify other network nodes associated with the compiled reports, and

select the serving cell for the terminal (UE) based on the adjusted reports in consensus with the identified other network nodes.

22. The network node according to claim 20 or 21 , wherein said network node further is configured to: determine terminal specific beamforming gains.

23. The network node according to any of claims 20-22, wherein said network node is a distributed network node, virtual network node, or a single dedicated network node.