WO2014040643A1

WO2014040643A1 - Method for energy saving in a cellular communication system

Info

Publication number: WO2014040643A1
Application number: PCT/EP2012/068135
Authority: WO
Inventors: George Koudouridis; Peter Legg; Henrik Lundqvist; Henrik Olofsson; Hong Li
Original assignee: Huawei Technologies Co., Ltd.
Priority date: 2012-09-14
Filing date: 2012-09-14
Publication date: 2014-03-20

Abstract

The present invention relates to a method for energy saving in a cellular communication system, said cellular communication system including one or more clusters of cells, wherein each cluster comprises one or more base stations and one or more mobile stations; said method comprising the steps of: exchanging, between base stations belonging to different clusters, energy saving status information associated with base stations of their respective cluster; and changing energy saving configuration for each of said one or more clusters in a sequence based on said energy saving status information. Furthermore, the invention also relates a method in a base station, a corresponding base station device, a computer program, and a computer program product.

Description

METHOD FOR ENERGY SAVING IN A CELLULAR COMMUNICATION SYSTEM

Technical Field

The present invention relates to a method for energy saving in a cellular communication system. Furthermore, the invention also relates a method in a base station, a corresponding base station device, a computer program, and a computer program product.

Background of the Invention

Cellular communication systems are designed for peak hour traffic despite the fact that the traffic activity in a certain area is high only during short periods of the day. Considering the traffic activity in a residential area, it is usually low during daytime when people have left homes for work while it increases in the evenings when people are at home. The opposite pattern prevails for the office area. In a heterogeneous network environment different infrastructures are differently utilised throughout the duration of a day, e.g. macro cells are serving traffic in a residential area during working hours while a vast amount of the traffic is carried by pico or femto cells during evenings and late hours. At the opposite end of the city in the working areas these indoor femto cells or outdoor pico cells that provide indoor coverage are not fully utilised at the end of business days or during evenings and nights. In all cases under-utilization of the radio infrastructure and radio resources is a waste of power and significant power savings could be achieved by reducing the number of radio resources that the network provides in time, space and frequency, e.g. by switching off a number of base stations at different tiers (micro, pico, femto etc.).

One of the simplest approaches to obtain energy efficiency is based on the activation of network resources on demand, thus avoiding to always power on all the resources that are necessary to serve the mobile users during peak traffic periods. This necessitates the implementation of a power on/off strategy that refers to the switching of radio infrastructure nodes and cells of a radio network. The radio network could be a heterogeneous network (HetNet) consisting of sites with different power transmission, coverage and capacity profiles. One of the key optimisation problems in such HetNet scenario is to maximize or maintain user throughput and coverage at a minimum of energy consumption cost.

One prior art solution proposes an iterative algorithm based on a simulated-annealing search. The algorithm is centralised meaning that it executes in a single point using information about all base stations and all mobile users in the network. The centralized approach has the following drawbacks:

• does not scale well to large networks with thousands of base station nodes and ten thousands of mobile users since (a) optimization complexity may be too great, (b) there will be more local maxima so there is a higher risk that the optimization algorithm chooses one of these, and (c) the gain matrix becomes very large and it is not stated how the gain matrix can be determined;

• network availability is at risk because the central computation entity represents a single point of failure.

There are other prior art solutions on energy savings in wireless networks. Some consider switching off a group of cells, whilst others focus on specific radio access technologies characteristics, e.g. UMTS and/or use different optimisation approaches. These solutions do not consider the trade-off between throughput and power consumption, and secondly do not consider the details of switching on/off individual cells.

Another drawback of the methods described above is that they are not feasible in practice. By switching group of cells according to a pattern it is not possible to flexibly address mobile users' needs in locations where non-uniform user distribution and non-uniform service demands occur. Moreover, the proposed optimisation algorithms are often computationally intractable and do not scale for large cellular networks.

Summary of the Invention

An object of the present invention is to provide a solution which mitigates or solves the drawbacks and problems of prior art solutions. More specifically, the present invention aims to provide a decentralized algorithm for energy saving and capacity optimization in a cellular communication system comprising multiple clusters of cells.

According to a first aspect of the invention, the above mentioned objects are achieved by a method for energy saving in a cellular communication system, said cellular communication system including one or more clusters of cells, wherein each cluster comprises one or more base stations and one or more mobile stations; said method comprising the steps of:

- exchanging, between base stations belonging to different clusters, energy saving status information associated with base stations of their respective cluster; and - changing energy saving configuration for each of said one or more clusters in a sequence based on said energy saving status information.

According to a second aspect of the invention, the above mentioned objects are achieved by a method in a base station, and in a corresponding base station device, for energy saving in a cellular communication system, said cellular communication system including one or more clusters of cells, wherein each cluster comprises one or more base stations and one or more mobile stations, and wherein said base station belongs to a first cluster; said method comprising the steps of:

- transmitting a first energy saving status information associated with one or more base stations of said first cluster to at least one base station belonging to a second cluster;

- receiving a second energy saving status information associated with one or more base stations of said second cluster; and

- determining energy saving configuration for base stations belonging to said first cluster based on said first and second energy saving status information.

The invention also relates to a computer program, having code means, which when run by processing means causes the processing means to execute the methods according to the present invention.

The present invention provides a solution which may be distributed between multiple clusters of base stations and therefore scales with acceptable complexity to address networks of any size. The present method (algorithm) runs off-line meaning that there are no real-time constraints to its execution. The multiple clusters together execute the distributed method and eventually this converges to a recommended network-wide energy saving configuration. The performance of the optimum configuration is expected to be equal to or only slightly inferior to the optimum configuration which would be determined by a fully centralised algorithm, but has the advantage over the fully centralised algorithm in the respect that it scales better to large networks. Further, the performance is expected to be better than a distributed algorithm that executes in isolated clusters that do not exchange information with each other.

According to preferred embodiments of the present invention joint optimisation of energy saving and network performance aspects which enables a network configuration that meets the policy needs of the operator, hence the operator can weigh the importance of energy saving and network performance and the configuration is tuned to the optimum point respecting this weighting.

Further applications and advantages of the invention will be apparent from the following detailed description.

Brief Description of the Drawings

The appended drawings are intended to clarify and explain different embodiments of the present invention in which:

• Fig. 1 illustrates a cellular communication network having a plurality of cells divided into multiple clusters;

• Fig. 2 illustrates the present invention; and

• Fig. 3 illustrates message exchanges and optimisation algorithm execution according to the invention.

Detailed Description of the Invention

In 3GPP RAN3, a method has been disclosed to allow a group of cells, i.e. a cluster comprising one or more base stations and one or more mobile stations, to cooperate to provide coverage compensation when one cell of the group is switched off. A cluster coordinator is defined and cells of the cluster send information to this cluster coordinator. The coordinator determines the on/off configuration of the cells and signals this back to each cell. Figure 1 shows an example network consisting of hexagonal macro cells and pico cells (marked with crosses) in which the network of cells is divided into one or more clusters of cells. The network is divided into four clusters in this particular example. Cluster A comprises 3 macro cells and 12 pico cells, and likewise clusters B and C. Cluster D consists of the remaining 12 macro cells.

The proposed 3 GPP RAN3 method defines the following mechanisms and logical entities:

Proposal 1 : Define Area Coverage Cluster (ACC) as the set of energy saving cells and compensating cells for which energy saving procedures are specified.

Proposal 2: ACC creation and management will be ruled and communicated from operator OAM, assigning Energy Saving Cell (ESC) or Compensation Cell (ESC) roles on the ACC.

Proposal 3: Define the ACCCO as ACC coordinator, to which cells information will be sent if needed, and from which commands or parameter set will be received for ESC algorithm control.

However, this method does not describe which parameter(s) that should be exchanged between the base station nodes. Additionally, it does not consider the best on/off configuration for all cluster of cells but the more narrow problem of which "energy saving" cells may be switched off; some cells are "compensation cells" and others are "energy saving" cells. Further, cell throughput or any other similar performance aspects are not addressed in this method.

Therefore, the present invention relates to a method for energy saving in a cellular communication system, wherein the cellular communication system includes one or more clusters of cells. Each cluster comprises one or more base stations and one or more mobile stations. The present method comprises the steps of: exchanging, between base stations belonging to different clusters, energy saving status information associated with base stations of their respective cluster; and changing energy saving configuration for each of the one or more clusters in a sequence based on the energy saving status information.

Often, the energy saving status information is exchanged between different cluster heads, which means that a base station in each cluster may be nominated as the cluster head and performs the utility calculations and optimization for its own cluster. The cluster heads then communicate with each other. So, each cluster head collects information about the energy saving status information for base stations of its own cluster and send this information to cluster heads of other clusters. Thereafter, the energy saving configuration is changed for each cluster in a given sequence based on the energy saving status information or on some other method. Exchanging the energy saving status information means that the energy saving status information is transferred between the base stations of the different clusters, e.g. by means of wired or wireless transmission which is well known to the skilled person. Figure 2 illustrates the present invention. The network in figure 2 is divided into 3 clusters, i.e. A, B and C. In each cluster one base station acts a cluster head and gathers all relevant information for that cluster (e.g. on/off status and mobile station measurements). The cluster heads are shaded in black in figure 2 and are labelled BS1, BS11, and BS21, respectively. In the step of the optimisation shown in figure 2 it is the turn of cluster A to run its algorithm and update the configuration of its cells. Clusters B and C (via their respective cluster heads BS11 and BS21) send relevant information to the head of cluster A, i.e. BS1. The minimum information exchanged between the cluster heads is the energy saving state of each cell in each cluster. In the next step of the optimisation roles are swopped, and e.g. cluster B runs the algorithm, and so on.

Figure 3 demonstrates message exchanges and optimisation algorithm execution by the heads of the clusters in figure 2. Cluster A begins and hence receives relevant information from clusters B and C before running its optimisation routine. Once the optimisation routine for cluster A has stopped any configuration update for cluster A identified by BS1 is shared to the other cluster heads BS11 and BS21. Next turn goes to cluster C, which starts after receiving measurements from cluster A and measurements and configuration data from cluster B. After cluster C has run its optimisation routine, it sends out updated configuration information and cluster B takes its turn. Here it is assumed that measurements are unchanged during the complete execution of the distributed algorithm, so that only energy saving configuration updates are needed. However, this is not true in general since measurement information may be updated as configurations change. For example, the energy consumption of the cluster may be updated after each energy saving state change. The clusters continue in sequence A-C-B- A-C-B, etc. At some point a stop condition is recognised by the cluster heads and the configuration for each cluster can be implemented and only at this point is the hardware adjusted, e.g. a cell is switched on/off, etc.

Cluster selection, i.e. which cells belonging to which cluster is an important aspect. Larger clusters should give better performance (fewer users adjacent to cells of other clusters) but complexity needs to be acceptable. Cluster boundaries should not bisect areas of dense mobile users since inter-cluster interference is likely to be high and this cannot be managed or controlled by a cluster where it occurs.

According to an embodiment of the invention the given sequence is round robin sequence scheduling. Alternatives to round robin sequence (scheduling) of clusters are possible. For example, "the cluster with the lowest utility goes next" sequence may be used. This scheduling can be enabled if all clusters of the system share their current utility to all other clusters. Yet another alternative is to use random assignment scheduling when deciding which cluster that should do the next change of energy saving configuration. In another embodiment of the invention, clusters take turns to tune their on/off configurations and share their current on/off status to other clusters. This is an off-line method wherein each cluster takes turn to update its on/off configuration but this is not implemented until the network optimization has finished. Typically, with a round robin scheduling, multiple rounds are required until the on/off configuration for the whole network converges to an optimum value. For example, in our example of four clusters A, B, C and D, in figure 1, clusters run their optimization in sequence A-B-C-D-AB-C-D-A-B-C-D, etc. When it is determined that cluster A can change its on/off optimization, then the on/off status for clusters B, C and D is passed to cluster A.

In another embodiment, a cluster (e.g. cluster A) calculates the cluster utility and tunes the energy saving status (e.g. on/off) of its cells to increase this utility. The cluster utility is calculated using the energy saving status of the cells belonging to the cluster and optionally other performance indicators captured over the cluster of cells. Knowledge of the on/off status of cells in other clusters is useful to determine the expected interference from them - the gain values of its mobile stations to these cells must also be known. Gain values may be calculated from its mobile station measurements of signal strength and the transmit powers of the cells in other clusters, or knowledge of the mobile station location.

In another embodiment, a cluster calculates the network utility and tunes its energy saving status to increase this utility. The network utility is calculated for all cells in the network. Knowledge of the on/off status of cells in other clusters is useful to determine the expected interference from them - the gain values of its mobile stations to these cells must also be known. Gain values may be calculated from its mobile station measurements of signal strength and the transmit powers of the cells in other clusters, or knowledge of the mobile station location. In addition, the mentioned cluster is required to calculate the impact of its on/off configuration on the throughput of users in other clusters, since this is embodied in the network utility. To do this the gain value from mobile stations in other cells to mentioned cluster cells is needed. Finally, to calculate the network utility the gain values of mobile stations in other clusters to cells of other clusters (i.e. B and C) is needed. Thus the complete gain matrix (all mobile stations, all cells) is required by cluster A. Essentially, this embodiment breaks the centralized optimisation into one in which the successor configurations of the optimization are constrained by the clustering. In each turn, one cell in one cluster may be toggled (on to off, or vice versa). Compared to the optimisation of the cluster utility, the network utility approach requires more signalling between clusters.

Methods may be required to terminate the optimization process in the network as mentioned above. For example, a simple rule could be applied to the network utility, such as utility has not been increased for 5 rounds. If cluster utility is employed termination could be enforced if each cluster shows no gain over its previous utility value. Another option would be to consider the energy saving state of each cluster and check for convergence to a steady state.

According to a preferred embodiment of the invention the present method further comprises the method steps of: exchanging, between base stations belonging to different clusters, measurement information associated with mobile stations of their respective cluster; and using the measurement information when changing the energy saving configuration for each of the one or more clusters of the cellular communication system. These measurements are useful in the method focussed on network utility to calculate path gain values, as explained above.

The measurement information may relate to one or more in the group comprising:

• radio signal strength;

• reference signal received power (RSRP), this is an LTE measurement and represents the signal strength of the cell specific reference symbols;

• pilot channel received signal code power (RSCP), this is the UMTS measurements on CCPCH channels which is also possible to use.

• path loss or path gain;

• channel quality measurement;

• channel quality indicator (CQI), this is a measurement undertaken by mobile stations in UMTS and LTE systems and reflects the quality of the downlink channel;

• signal-to-noise ratio (SNR) or signal-to-interference plus noise ratio (SINR);

• data throughput; and

• mobile station location, since knowledge of the location of the mobile station can be used to determine path gain values, e.g. using a previously established database. Knowledge of the bearer quality requirements such as the QoS Class Identifier (QCI) enable more elaborate utility values considering different UE classes or services. Moreover, the measurement information may be minimisation of drive testing (MDT) information. MDT has been introduced by 3 GPP as a means for a commercial mobile to log measurements (such as RSRP) and later to report these to the network.

It should further be realised that the above mentioned measurement information is measured by mobile stations and/or by base stations. When the mobile stations make the measurements it is the downlink case whilst when the base stations make the measurements it is the uplink case. Combination of the uplink and downlink cases may also be used. The exchange of measurement information between the base stations can be performed by broadcasting according to an embodiment.

Regarding the change of energy saving configuration this may involve setting or adjusting one or more in the group comprising:

• cell on/off mode;

• base station or base station component sleep/active mode, wherein in active state full operation is assumed while in sleep mode there is limited functionality in operation and power consumption is reduced;

• discontinuous transmission (DTX) operation whereby the cell transmission is halted for short periods of time thus saving power consumption;

• bandwidth reduction may be used for cells with low traffic levels such that the bandwidth of the power amplifier can be reduced so that its operation point is more efficient. Transmission of common channels and pilots across the full band is also not necessary;

• power amplifier operation may be adjusted to reduce power amplifier power consumption, e.g. by reducing the peak transmission power;

• transmission power which is directly correlated to power consumption; and

• antenna tilt adjustment can be used to adjust the coverage of a cell, thereby one cell may be switched off and its lost coverage compensated by up-tilting antennas of other cells. It has also been realised by the inventors that the measurement information may be tagged with information relating to one or more in the group comprising: base station antenna tilt, mobile station location, and mobile station quality of service class identifier (QCI). Measurement information may be tagged with the mobile station location to help the optimization engine to build up its own database of path gain values. Tagging using antenna tilt values can assist the cluster head to identify the antenna tilt value associated with different mobile station measurements. Alternatively, measurements can be tagged with a timestamp and the network state can be determined by information exchange between nodes. To allow antenna tilt tagging the base station should advertise the current tilt value (e.g. in system information) or signal mobile stations when it is changed (or at call set-up). Measurements may be gathered by direct signalling or by MDT signalling (including logged measurements). As for the base station antenna tilt information, this information can be broadcasted by base stations and read by mobile stations according to an embodiment of the information. This is necessary such that mobile stations can add the tilt value as a tag to measurements.

According to another embodiment of the invention the measurement information associated with mobile stations moving with a speed greater than a threshold speed is not exchanged. The basic idea is to ignore mobile stations that are moving rapidly. One general comment is that for on-line optimization, or any other optimization which does not use statistics collected over a longer period it would be good to use the speed of the mobile stations to understand which mobile stations are meaningful to take into account in the optimization algorithm. With fast moving mobile stations the data used in the optimization would be outdated quickly after the on/off status of the cluster is changed. Hence, the optimization could only be made based on mobile stations with low speed with the assumption that fast moving mobile stations will stay in a coverage layer, or the presence of fast mobile stations may need to be signalled to make sure that there is a coverage layer. The speed could be estimated either by MDT positioning, mobility state information (determined by a standardised method for LTE mobile stations), mobile station handover history, or any other suitable method. Alternatively, if e.g. the RSRP values are varying too much these values could be filtered in some way, either be excluded or just averaged depending on what would be best for the optimization. Based on the transition time/cost for switching on/off a cell some conclusion about a reasonable period between vector changes can be made and from that it would also be more clear for how long measurements should be collected to provide suitable statistics for the optimization. Considering the above reasoning one preferred speed threshold value is the speed for pedestrians. However, other threshold values may be used depending on the application.

The present method may according to yet another embodiment further comprise the steps of: exchanging, between the base stations belonging to different clusters, further information relating to one or more in the group comprising: cell antenna tilt, cell transmit power, and cell energy saving mode; and changing energy saving configuration for cells belonging to their cluster based on this further information. The base station can run an optimisation algorithm to determine the best energy saving configuration for its cluster, i.e. all cells in the cluster. The cell transmit power can be used to determine path loss values of mobile stations for which only signal strength measurements are available. The optimisation algorithm typically only adjusts the energy saving mode of one cell at a time, so it is important to know the current energy saving status of each cell. Cell antenna tilt can be used to determine which of several measurements from a mobile station tagged with different tilt values is currently applicable.

The present method may further comprise the step of: exchanging, between base stations belonging to different clusters, further information relating to cell related measurements, which preferably are one or more in the group comprising: cell energy consumption, cell throughput, and cell energy efficiency. Such measurements may be used to calculate the network utility. For example, if the utility includes network energy consumption then signalling of cell energy consumption for each cell in a cluster, given the current assumed configuration of the cluster, may be essential. If the network utility includes throughput parts then cell throughputs may be needed. If the network utility includes energy efficiency parts then energy efficiency values may be needed.

Yet another step in the present method may be that further information relating to cluster related measurements is exchanged between base stations belonging to different clusters. The cluster related measurements are one or more in the group comprising: cluster utility, and energy saving optimisation method. Cluster utility exchange may be beneficial in three respects. When cluster utility optimisation is used values of utility values in other clusters may be used to identify when the optimisation should stop. The cluster utility can be used to determine which cluster should run the optimisation next (e.g. using "cluster with lowest utility next" rule). When network utility is used the cluster utility may be a component in a network utility calculation. The optimisation method may be useful such that different clusters can employ the same techniques to give consistent behaviour across the clusters such that convergence to the final state is accelerated. The cluster utility is recalculated after changing the energy saving configuration according to yet another embodiment. When the cluster has its next chance to tune its configuration the recalculated utility value is the new baseline which should be increased by a new configuration. Changing the energy saving configuration may involve: minimising base station energy consumption. However, according to another preferred embodiment of the invention changing the energy saving configuration may involve joint optimisation of base station energy consumption AND one or more in the group comprising: mobile station energy consumption, mobile station quality of service, cell throughput and cell edge throughput. In the joint optimisation the algorithm considers more than one performance indicator, for example, energy consumption AND cell edge throughput. Clearly minimum energy consumption occurs with all cells switched off but cell edge throughput is then zero. The joint optimisation weights the performance indicators in some way to determine the best overall state. For example, a utility that weights energy consumption and cell edge throughput could be used. The objective is to maximise this utility value.

An example of joint optimisation of energy saving and throughput is given below. The configuration of the entire heterogeneous network can be defined in terms of a transmit indicator vector h and a power transmit vector p as given by h — ( i- , ... , h_K, h_K+1, ... , h_K+L) (1) ,

p = ( i, ... , if, if₊i, ... , if_+L) (2) , where p_c and h_c are the power transmit and the power on/off state configuration of cell c , respectively, and cell edge user throughput and energy consumption ratio are combined into a single aggregate objective (network utility) function / in which the utility value of the current working configuration (on/off vector) h_w is calculated relatively to a baseline reference configuration h_r as the weighted sum f(Pw. Pr = ^a■ ø- + (1 - a)^■ G (_p_w, Pr) (3),

where / corresponds to the objective function; coefficient a, 0< a < 1 is a weight indicating the significance of the terms. The first term is defined by the ratio of the outage (or cell edge) throughput of the current configuration O_w over the outage throughput of the baseline configuration O_r. The second term, corresponds to the Energy Reduction Gain (ERG) G and is defined based on the ratio of the ECR of current configuration E (p_w) over the ECR of the baseline configuration E(p_r) c*-*⁾⁼ⁱ-B¾ ⁽⁴⁾·

Given the above metrics the optimization problem is to

max f(p_w, p_r) (5).

In addition one or more constraints can be added such that the received signal strength for each mobile station is greater than threshold value(s). If cluster utility optimisation is followed then the utility is calculated over the subset of cells of the network that lie in the cluster. The present invention also relates to a method in a base station device for energy saving in a cellular communication system, and to the base station device as such. The cellular communication system is any suitable wireless system, e.g. defined by 3GPP standards, divided into one or more clusters of cells. The base station belongs to a first cluster and the method comprises the steps of: transmitting a first energy saving status information associated with one or more base stations of said first cluster to at least one base station belonging to a second cluster; receiving a second energy saving status information associated with one or more base stations of the second cluster; and determining energy saving configuration for base stations belonging to the first cluster based on the first and second energy saving status information.

Hence, this method is the application of the method above in a base station. It should also be noted that the method in a base station and the base station device may be modified, mutatis mutandis, according to different embodiments of the method in the cellular communication system.

Moreover, the above described base stations and mobile stations have suitable functions and are therefore arranged to fulfil the requirements of different wireless communication standards such as 3GPP standards, WiMax, etc. Hence, the base stations may e.g. be eNBs, and the mobile stations may be UEs in a LTE system.

Furthermore, as understood by the person skilled in the art, any method according to the present invention may also be implemented in a computer program, having code means, which when run by processing means causes the processing means to execute the steps of the method. The computer program is included in a computer readable medium of a computer program product. The computer readable medium may comprises of essentially any memory, such as a ROM (Read-Only Memory), a PROM (Programmable Read-Only Memory), an EPROM (Erasable PROM), a Flash memory, an EEPROM (Electrically Erasable PROM), or a hard disk drive.

Finally, it should be understood that the present invention is not limited to the embodiments described above, but also relates to and incorporates all embodiments within the scope of the appended independent claims.

Claims

1. Method for energy saving in a cellular communication system, said cellular communication system including one or more clusters of cells, wherein each cluster comprises one or more base stations and one or more mobile stations; said method comprising the steps of:

- exchanging, between base stations belonging to different clusters, energy saving status information associated with base stations of their respective cluster; and

- changing energy saving configuration for each of said one or more clusters in a sequence based on said energy saving status information.

2. Method according to claim 1, wherein said sequence is round robin sequence.

3. Method according to claim 1, further comprising the step of:

- exchanging, between base stations belonging to different clusters, measurement information associated with mobile stations of their respective cluster; and

- using said measurement information when changing the energy saving configuration for each of said one or more clusters.

4. Method according to claim 3, wherein said measurement information relates to one or more in the group comprising: radio signal strength, reference signal received power (RSRP), pilot channel received signal code power (RSCP), channel quality measurement, channel quality indicator (CQI), signal-to-noise ratio (SNR), signal-to-interference plus noise ratio (SINR), data throughput, and mobile station location.

5. Method according to claim 3, wherein said measurement information is measured by mobile stations and/or by base stations.

6. Method according to claim 5, wherein said measurement information is tagged with information relating to one or more in the group comprising: base station antenna tilt, mobile station location, and mobile station quality of service class identifier (QCI).

7. Method according to claim 6, wherein the base station antenna tilt information is broadcasted by base stations and measured by mobile stations.

8. Method according to claim 3, wherein said measurement information is minimisation of drive testing (MDT) information.

9. Method according to claim 3, wherein measurement information associated with mobile stations moving with a speed greater than a threshold speed is not exchanged between the different clusters.

10. Method according to claim 1, further comprising the step of: - exchanging, between said base stations belonging to different clusters, further information relating to one or more in the group comprising: cell antenna tilt and cell transmit power.

11. Method according to claim 3, wherein said step of exchanging involves:

- broadcasting, by said base stations belonging to different clusters, said measurement information.

12. Method according to claim 1, further comprising the step:

- exchanging, between said base stations belonging to different clusters, further information relating to cell related measurements.

13. Method according to claim 12, wherein said cell related measurements are one or more in the group comprising: cell energy consumption, cell throughput, and cell energy efficiency.

14. Method according to claim 1, further comprising the step:

- exchanging, between said base stations belonging to different clusters, further information relating to cluster related measurements.

15. Method according to claim 14, wherein said cluster related measurements are one or more in the group comprising: cluster utility, and energy saving optimisation method.

16. Method according to claim 15, wherein cluster utility is recalculated after changing the energy saving configuration.

17. Method according to claim 1, wherein said energy saving status information relates to one or more in the group comprising: cell on/off mode, base station or base station component sleep/active mode, discontinuous transmission (DTX) operation, bandwidth reduction, power amplifier operation, transmission power, and antenna tilt.

18. Method according to claim 1, wherein changing the energy saving configuration involves setting or adjusting one or more in the group comprising: cell on/off mode, base station or base station component sleep/active mode, discontinuous transmission (DTX) operation, bandwidth reduction, power amplifier operation, transmission power, and antenna tilt.

19. Method according to claim 1, wherein changing the energy saving configuration involves: minimising base station energy consumption.

20. Method according to claim 1, wherein changing the energy saving configuration involves joint optimisation of base station energy consumption and one or more in the group comprising: mobile station energy consumption, mobile station quality of service, cell throughput and cell edge throughput.

21. Method in a base station for energy saving in a cellular communication system, said cellular communication system including one or more clusters of cells, wherein each cluster comprises one or more base stations and one or more mobile stations, and wherein said base station belongs to a first cluster; said method comprising the steps of:

22. Computer program, characterised in code means, which when run by processing means causes said processing means to execute said method according to any of claims 1-21.

23. Computer program product comprising a computer readable medium and a computer program according to claim 22, wherein said computer program is included in the computer readable medium, and comprises of one or more from the group: ROM (Read-Only Memory), PROM (Programmable ROM), EPROM (Erasable PROM), Flash memory, EEPROM (Electrically EPROM) and hard disk drive.

24. Base station arranged for communicating in a cellular communication system, said cellular communication system including one or more clusters of cells, wherein each cluster comprises one or more base stations and one or more mobile stations, and said base station belongs to a first cluster of cells; said base station further being arranged to:

- transmit a first energy saving status information associated with one or more base stations of said first cluster to at least one base station belonging to a second cluster;

- receive a second energy saving status information associated with one or more base stations of said second cluster; and

- determine energy saving configuration for base stations belonging to said first cluster based on said first and second energy saving status information.