TECHNICAL FIELD

[0001] The present disclosure relates to a method of controlling operation of a group of access points being configured to serve a plurality of wireless communication devices, and a device performing the method.

[0002] The project leading to this application has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No 101013425.

BACKGROUND

[0003] In some wireless radio communication deployments, a plurality of access points (AP) are arranged in a confined environment such as within a single integrated circuit or accommodated within a single housing. In such an environment, cooling conditions may be poor which causes the APs to become more sensitive to temperature as compared to a deployment where a plurality of APs are setup under good cooling conditions by means of e.g. heat sinks, fans or even wind.

[0004] Distributed Multiple Input Multiple Output (D-MIMO), also known as cell-free massive MIMO, is a key technology candidate for physical layer in 6 ^th generation (6G) wireless radio communication.

[0005] In D-MIMO, service antennas are distributed geographically and operated (e.g. phase-coherently) as a group. In a typical D-MIMO architecture, multiple antenna panels (also known as remote radio heads, transmission points, or access points) are interconnected and configured to cooperate in decoding data to/from a given wireless communication device, e.g. a smart phone, tablet, gaming console, connected vehicle, etc.

[0006] In D-MIMO, the multiple antenna panels are commonly arranged in a so- called radio stripe where a group of AP ICs e.g. are interconnected inside a cable along a length of the cable. In such a setup, a high level of miniaturization is attained and cooling equipment in the form of e.g. heat sinks or fans is undesired (and sometimes not possible to implement). Nevertheless, high operating temperature and potential overheating of the APs is an issue that needs to be resolved. SUMMARY

[0007] One objective is to solve, or at least mitigate, this problem in the art and provide an improved method of controlling operation of a group of access points being configured to serve a plurality of wireless communication devices.

[0008] This objective is attained in a first aspect by a method of controlling operation of a group of access points (APs) being configured to perform radio communication for serving a plurality of wireless communication devices (WCDs). The method comprises determining a thermal state of each AP in the group, and assigning a task to be performed by at least one of said each AP, among a set of tasks to be performed by the group of APs upon serving the WCDs, the assigned task being a task resulting in a power consumption when being performed at the at least one of said each AP that causes the determined thermal state of the at least one of said each AP to comply with a predetermined thermal condition.

[0009] This objective is attained in a second aspect by an access point control device configured control operation of a group of APs being configured to perform radio communication for serving a plurality of WCDs, said access point control device comprising a processing unit and a memory, said memory containing instructions executable by said processing unit, whereby the device is operative to determine a thermal state of each AP in the group, and to assign a task to be performed by at least one of said each AP, among a set of tasks to be performed by the group of APs upon serving the WCDs, the assigned task being a task resulting in a power consumption when being performed at the at least one of said each AP that causes the determined thermal state of the at least one of said each AP to comply with a predetermined thermal condition.

[0010] Advantageously, by taking into account a thermal state such as AP temperature upon assigning tasks to the APs, the access point control device ensures that the tasks to be assigned does not cause power consumption at an AP resulting in the thermal state not complying with a predetermined thermal condition, such as e.g exceeding a temperature threshold value that the AP temperature is not allowed to exceed. [0011] In an embodiment, the thermal state comprises a measured temperature of each AP and the predetermined thermal condition being a temperature threshold value that should not be exceeded for the condition to be complied with.

[0012] In an embodiment, the access point control device further determines a channel state of a communication channel to be established between the at least one of said each AP and at least one of the WCDs to be served, wherein the power consumption being the result of a selected task is determined by further taking into account the channel state of the communication channel to be established.

[0013] In an embodiment, the channel state comprises an expected channel quality of the communication channel to be established.

[0014] In an embodiment, the access point control device further determines the expected channel quality by instructing one or more of the APs in the group to measure a channel quality indicator (CQI) of a channel established with the respective WCD.

[0015] In an embodiment, the access point control device further, for each AP, measures a temperature when in idle mode, and measures a temperature when subjected to at least one load scenario when in operative mode; wherein the determining of a thermal state of each AP in the group comprises determining, from the measured idle temperature and the measured load temperature, expected heat dissipation capability for each AP.

[0016] In an embodiment, the access point control device further determines, from the measured idle temperature and the measured load temperature, the expected heat dissipation capability for each AP by taking into account a thermal mass constant and heat dissipation capability constant of each AP in the group.

[0017] In an embodiment, the access point control device further determines a channel state of communication channels to be established to serve the WCDs, computes a utility measure of each AP in the group based on the determined channel state and thermal state, the utility measure increasing with a higher channel quality and decreasing with a higher temperature, and assigns to an AP in the group which has not yet been assigned and having a highest utility measure, the task of serving a selected WCD of the plurality of WCDs if power consumption at the assigned AP for serving the selected WCD does not cause an increase in temperature resulting in the utility measure decreasing below a predetermined utility threshold value, and proceeding with the assigning until all WCDs are served by at least one AP.

[0018] In an embodiment, the access point control device evaluates, in case the utility measure decreases below the predetermined utility threshold value, an AP in the group which has not yet been assigned and having a next-highest utility measure for assignment.

[0019] In an embodiment, the access point control device determines whether or not a quality measure of the channel established with the selected WCD exceeds a quality threshold value, and if so proceeding with the assigning until all WCDs are served by at least one AP.

[0020] In an embodiment, the access point control device evaluates, in case the quality measure of the channel established with the selected WCD does not exceed a quality threshold value, an AP in the group which has not yet been assigned and having a next-highest utility measure for assignment.

[0021] In a third aspect, a computer program is provided comprising computerexecutable instructions for causing the access point control device of the second aspect to perform the method of the first aspect when the computer-executable instructions are executed on a processing unit included in the device.

[0022] In a fourth aspect, a computer program product is provided comprising a computer readable medium, the computer readable medium having the computer program according to the third aspect embodied thereon.

[0023] Generally, all terms used in the claims are to be interpreted according to their ordinary meaning in the technical field, unless explicitly defined otherwise herein. All references to "a/an/the element, apparatus, component, means, step, etc." are to be interpreted openly as referring to at least one instance of the element, apparatus, component, means, step, etc., unless explicitly stated otherwise. The steps of any method disclosed herein do not have to be performed in the exact order disclosed, unless explicitly stated.

BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Aspects and embodiments are now described, by way of example, with reference to the accompanying drawings, in which: [0025] Figure la illustrates a top view of an example of an AP in the form of an IC arranged on a printed circuit board utilized in D-MIMO;

[0026] Figure ib illustrates a side view of the AP of Figure la;

[0027] Figure 2 illustrates a D-MIMO system implemented in the form of a radio weave;

[0028] Figure 3 illustrates a D-MIMO system implemented in the form of a plurality of interconnected radio stripes;

[0029] Figure 4 shows a flowchart illustrating a method of controlling operation of a group of APs being configured to serve a plurality of wireless communication devices according to an embodiment;

[0030] Figure 5 shows a flowchart illustrating a method of controlling operation of a group of APs being configured to serve a plurality of wireless communication devices according to another embodiment;

[0031] Figure 6 shows a flowchart illustrating a method of controlling operation of a group of APs being configured to serve a plurality of wireless communication devices according to a further embodiment;

[0032] Figure 7 shows a flowchart illustrating a method of controlling operation of a group of APs being configured to serve a plurality of wireless communication devices according to yet an embodiment; and

[0033] Figure 8 illustrates an AP control device according to an embodiment.

DETAILED DESCRIPTION

[0034] The aspects of the present disclosure will now be described more fully hereinafter with reference to the accompanying drawings, in which certain embodiments of the invention are shown.

[0035] Figure la illustrates an example of an AP 10 in the form of an IC arranged on a printed circuit board (PCB) 11 utilized in D-MIMO as previously discussed in a top view while Figure lb illustrates the AP 10 in a side view. In this example, the AP 10 comprises four antenna elements I2a-i2d and an external interface in the form of four ports I3a-i3d for transmitting data to and receiving data from other components and possibly to connect to a power supply. The AP 10 may further be equipped or interconnected via a set of interconnectors 15 with a component having data processing capability, such as digital signal processor 14, which in its turn is connected via another set of interconnectors 16 with the PCB 11.

[0036] Figure 2 illustrates a D-MIMO system 20 implemented in the form of a radio weave where a plurality of APs loa-iol are interconnected and further connected to a managing controller in the form of a central processing unit 21 (CPU). In this particular illustration, the D-MIMO system 20 serves four wireless communication devices (WCDs) 22-25 commonly referred to as user equipment (UE). However, as is understood, a D-MIMO system may in practice serve hundreds or even thousands of UEs and is commonly deployed in e.g. shopping malls or a smaller section of a dense city street. The APs of Figure 2 are sometimes also referred to as antenna processing units (APUs).

[0037] Figure 3 illustrates a D-MIMO system 20 implemented in the form of a plurality of interconnected radio stripes. In this implementation, APs loa-iod are interconnected inside a cable 30 thus forming a so-called radio stripe 31. Each radio stripe may be interconnected with one or more further radio stripes to form the D- MIMO system 20, which again typically comprises a CPU 21 for managing the radio stripes. As is understood, an access point control device comprising the CPU 21 may be arranged close the APs, for instance in the same housing. Alternatively, the access point control device comprising the CPU 21 may be arranged remotely from the APs. For instance, the APs maybe arranged at different location inside a building such as a galleria, while the access point control device is connected to the APs and performs the function of a supervision node,

[0038] Thus, in a typical D-MIMO architecture, multiple APs are interconnected and configured to cooperate in decoding data to/from a given UE. Each AP may in turn comprise multiple antenna elements that are configured to operate phase- coherently together. One way of operation is in time-division duplexing (TDD), relying on reciprocity of a propagation channel, whereby uplink pilots transmitted by the UEs are used to obtain both uplink and downlink channel responses simultaneously. This type of TDD operation is usually called reciprocity-based operation. A D-MIMO system may also operate in a frequency-division duplexing (FDD) mode. In FDD the uplink and downlink channels can typically not be assumed to be reciprocal and hence additional downlink pilot signals and feedback to the APs may be required to obtain the required channel state information.

[0039] To make deployment of a large number of distributed MIMO access points small, simple and cost efficient, various solutions have been proposed, such as the radio weave and the radio stripe implementations of Figures 2 and 3. With these implementations, it is difficult to provide good cooling conditions since cooling components such as fans or heat sinks are bulky and space-demanding.

[0040] As is understood, the APs heat up during operation and heat dissipation becomes a problem, in particular for APs that are heavily loaded by serving many UEs, and for APs that are mounted in locations that are either subject to significant inbound heat radiation or transfer. This may occur for instance if an AP is located in strong sunshine, close to the ceiling in a building with inadequate air conditioning, etc., or installed in such a way that heat dissipation is difficult - for example, if the AP is embedded or located in the immediate proximity of material that provides high thermal insulation.

[0041] Figure 4 shows a flowchart illustrating a method of controlling operation of a group of APs being configured to serve a plurality of UEs, such as APs loe-ioh serving UEs 22-25 as illustrated in Figure 2. Selection of which AP(s) should serve each UE can be performed in different ways, one common way is to select APs based on the average channel gain (or path-loss) to each UE. For example, an AP maybe assigned to serve a UE if the average channel gain between said UE and said AP is above a threshold, in order to ensure that a channel of sufficient quality will be set up. In this exemplifying embodiment, the CPU 21 is the device performing the steps of the method. For brevity, control of only the four APs loe, lof, 10g, loh will be illustrated while in practice, the CPU 21 may consider all APs loa-iol when assigning tasks to the APs for serving the UEs.

[0042] A group of APs is thus a plurality of APs where each AP in the group may be taken into consideration for serving one or more UEs. For instance, it maybe envisaged that the average channel gain between the four APs loe-h and the UEs 22- 25 to be served exceeds a predetermined threshold value, while for instance AP 10a has an average channel gain being below the predetermined threshold value and therefore AP 10a will not be included in the group for consideration. [0043] The APs loe-ioh will, upon performing radio communication for serving the UEs 22-25, perform a task of transmitting and/or receiving radio signals resulting in heat up of the APs. For instance, analog components such as power amplifiers (operated during radio signal transmission) and low noise amplifiers (operated during radio signal reception) of the APs loe-ioh will heat up in addition to e.g. AP microprocessor(s) heating up during digital signal processing tasks being performed. As is understood, heat generated by analog components is typically greater than heat generated by digital components.

[0044] In a first step S101, the CPU 21 determines a thermal state of each AP loe- loh in the group. For instance, a temperature sensor (not shown) maybe arranged at each AP in order to locally determine the temperature of each AP loe-ioh in the group.

[0045] The temperature may be measured momentarily at each AP or may be measured over time to determine a trend of the temperature at each AP, or even estimated.

[0046] In this exemplifying embodiment, a current temperature at each AP loe- loh is measured and it is assumed that the temperature of first, second and third AP loe, lof, 10g is at temperature Ti while the temperature of fourth AP loh is at a higher temperature T2.

[0047] In a second step S102, the CPU 21 takes the thermal state - in this example the measured temperature - of each AP loe-ioh into account upon assigning tasks to the group for serving the UEs 22-25 in order to avoid causing overheating of any one of the APs loe-ioh.

[0048] That is, taking the thermal state under consideration, the CPU 21 will determine whether or not a task to be assigned to an AP in the group will result in a power consumption at the AP that causes the thermal state of the AP to comply with a predetermined thermal condition or not.

[0049] It may be envisaged that step S102 comprises assigning, to each AP loe- loh, a task to be performed among a set of tasks to be performed by the group of APs loe-ioh upon serving the UEs 22-25, the selected task being a task resulting in a power consumption when being performed at said each AP loe-ioh that causes the thermal state of said each AP loe-ioh to comply with a predetermined thermal condition, wherein a thermal state of each AP (loe-ioh) in the group has been determined.

[0050] In this exemplifying embodiment, the thermal state of an AP is considered to comply with a predetermined thermal condition if the measured temperature of the AP does not exceed a predetermined temperature threshold value Tt.

[0051] It is assumed that the task of the group of APs loe-ioh is to each serve a corresponding UE 22-25, while in practice a UE may well be served by a plurality of APs.

[0052] In this example, it is determined that while the temperature of the first, second and third AP loe-iog is at Ti, which is well below the predetermined temperature threshold value Tt, the temperature of the fourth AP loh is at T2 which in this example is just under Tt.

[0053] Thus, a task assigned to the APs cannot be allowed to cause a power consumption at the respective AP resulting in the thermal state not complying with the predetermined thermal condition, i.e. the assigned task cannot cause an increase in power consumption at the AP that causes the temperature of the AP to exceed Tt.

[0054] As a result, the CPU 21 concludes that since the temperature T2 of the fourth AP loh already is high, and thus approaches Tt, there is a risk that the temperature T2 of the fourth AP loh will exceed Tt (and possibly be overheated) if the fourth AP loh is given the task of fully serving one of the UEs since the associated power consumption of the fourth AP loh will further cause the temperature T2 to increase.

[0055] It is assumed that the power consumption of the three APs loe-iog is not expected to increase to such an extent that Ti > Tt when being assigned the task of fully serving a UE.

[0056] Even though the CPU 21 could come to the conclusion that the fourth AP loh should not be assigned the task to serve a UE at all (if temperature T2 is sufficiently high), it is in this exemplifying embodiment assumed that the first AP loe and the second AP lof are assigned the task of serving the first UE 22 and second UE 23, respectively, while the third AP 10g is assigned the task of serving the third UE 24 and the fourth UE 25 in downlink (DL) while the fourth AP loh is assigned the less power consuming task of serving the third UE 24 and the fourth UE 25 in uplink (UL). Thus, all tasks have been assigned by the CPU 21 and no AP temperature exceeds Tt, i.e. the thermal state of each AP complies with the predetermined thermal condition. The CPU will thus send an instruction to each AP loe-ioh in the group indicating the task assigned to said each AP loe-ioh.

[0057] UL processing typically requires more digital processing than DL, generating heat primarily in digital components of the APs, while DL processing requires both digital processing as well as analog power amplification in power amplifiers that generate heat in both digital and analog components. By allocating UL and DL processing to different sets of APs for the same UE, tasks causing heat generation can be distributed over several APs more evenly, thereby avoiding overheating of any individual AP.

[0058] It may even be envisaged that e.g. voltage and/ or clock frequencies of digital circuits such as digital signal processors (DSPs), application specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), complex programmable logic devices (CPLDs), etc., is dynamically adapted to a task assigned to each AP. A lower voltage in a digital circuit typically result in reduced leakage losses while a lower clock frequency result in less switching losses, and lower losses translate to lower temperature increase in the APs. By dynamically adjusting the voltage and clock, known as dynamic voltage and frequency scaling (DVFS), a more fine-grained control of the heat generation in an AP can be achieved.

[0059] In some embodiments the voltage and/ or clock-frequency of digital circuits (e.g. ASICs, DSPs, CPUs, accelerator cores, etc) is dynamically adapted to the workload assigned to each AP. A lower voltage in a digital circuit results in reduced leakage losses while a lower clock frequency results in less switching losses. Lower losses translate to lower temperature increase in the APs. By dynamically adjusting the voltage and clock (known as dynamic voltage and frequency scaling, DVFS) a more fine-grained control of the heat generation in an AP can be achieved.

[0060] Advantageously, by assigning tasks to the group of APs loe-ioh while taking into account the thermal state of each AP, overheating issues may be resolved or at least mitigated. [0061] Hence, this results in a reduced risk of overheating and thermal wear-out of the APs and reduced use of cooling means such as e.g. fans and heat sinks, while enabling miniaturization of the APs.

[0062] Figure 5 illustrates a further embodiment, where not only the thermal state in the form of current AP temperature is taken into account but further a channel state, in this exemplifying embodiment in the form of an expected channel quality C.

[0063] As is understood, the worse the channel quality, the more processing is required to be performed at an AP, which ultimately results in a higher power consumption and hence higher temperature. For instance, if a channel established between the first AP loe and the first UE 22 is poor, the first AP loe may have to increase transmission power, or generally assign further resources such as increasing bandwidth, using a more processing-heavy encoding scheme, retransmitting dropped data, etc.

[0064] The channel state may in an embodiment be determined by having one or more of the APs loe-ioh measure a channel quality indicator (CQI) of a channel established with the respective UE 22-25.

[0065] Hence, in this embodiment, the expected channel state is determined in step Sioia and taken into account by the CPU 21 upon determining power consumption at each AP in order to ensure that the power consumption does not increase to such an extent that the thermal state of one or more APs no longer complies with the predetermined thermal condition when the tasks are being assigned in step S102. As is understood, the determining of the thermal state in step S101 maybe performed before, after or even in parallel with step S101 of determining channel state.

[0066] Hence, as shown in Figure 5, it is assumed in step S102 that (a) the expected channel quality Ci of a channel established between the first AP loe and the first UE 22 exceeds a predetermined channel quality threshold value Ct, (b) the expected channel quality C2 of a channel established between the second AP lof and the second UE 23 exceeds the predetermined channel quality threshold value Ct, and (c) the expected channel quality C3 of a channel established between the third AP 10g and the third UE 24 exceeds the predetermined channel quality threshold value Ct. It is noted that in practice, a different channel quality threshold value Ct may be set for one more of the APs.

[0067] However, the expected channel quality C4 of a channel established between the fourth AP loh and the fourth UE 25 is below the predetermined channel quality threshold value Ct, the conclusion being that the fourth AP loh is at risk of being overheated - the temperature already being at T2 described previously to be just under Tt - if the fourth AP loh would be assigned the task of handling UL communication of the fourth UE 25, as was the case in the embodiment of Figure 4, since the lower channel quality C4 likely would result in a processing-heavier burden being imposed on the fourth AP loh with a corresponding increase in temperature T2. In other words, a processing-heavy task assigned to the fourth AP loh may cause the thermal state of the fourth AP loh to no longer comply with the predetermined thermal condition, i.e. T2 > Tt.

[0068] As a result, the CPU 21 will in this exemplifying embodiment in step S102 assign to the third AP 10g the task of taking over the handling the UL of the fourth UE 25 from the fourth AP loh and thus to fully handle the fourth UE 25, while the fourth AP loh will continue handling the UL for the third UE 24 only.

[0069] As is understood, thermal state and/ or channel state is in practice determined and reassessed repeatedly, thereby causing tasks to be continuously reassigned by the CPU 21 among the group of APs loe-ioh.

[0070] As previously mentioned, it is envisaged that a trend in the thermal state of the group of APs is determined and taken into account rather than just assessing instant temperature.

[0071] Figure 6 illustrates a flowchart of the method according to this embodiment where in a procedure carried out initially, and/or on a regular basis, the CPU 21 determines the thermal state of each AP loe-ioh by first ensuring that each AP is idle, waits for a given time period and then determines a temperature of each AP in step Siooa; this is referred to as AP ambient/idle temperature. As is understood, if the idle time period is sufficiently long, the idle temperature will effectively be at the ambient temperature of the AP.

[0072] Thereafter, the CPU 21 subjects each AP loe-ioh to a load while each AP is operative. This may be a potentially fictitious load in that each AP loe-ioh performs a particular task causing a certain load for a given time period during which the CPU 21 again acquires the measured temperature of each AP in step Sioib; this is referred to as AP load temperature. As is understood, when the APs are subjected to this load scenario, they are not necessarily operating to serve the UEs 22-25, but may perform any appropriate computational task resulting in a certain load.

[0073] Using the measured AP idle temperature and the measured AP load temperature, the thermal state of each AP loe-ioh is determined in step S101, this time in the form of heat dissipation capability.

[0074] A mathematical model for assessing the heat dissipation capability of an AP will be discussed in the following how the AP temperature varies with time.

[0075] At each instant of time t, each AP has an AP temperature T _AP (t) and an ambient temperature T _amb , that is assumed to be constant (or at least varies sufficiently slowly to be considered constant).

[0076] The CPU 21 may compute a thermal state model for each AP loe-iog in the group describing how the AP temperature T _AP (t) varies with time and load.

[0077] In the model described below in equation (1), it is assumed that the heat P _d (t) dissipating from the AP at time t is proportional to the difference between the AP temperature and the ambient temperature, T _AP (t) - T _amb . In other words, the heat dissipation capability of an AP may be determined by computing the difference between the measured load temperature and the measured ambient/idle temperature.

[0078] Equation (1) hence constitutes a first-order approximation of actual physical conditions, which is valid for small deviations between the AP temperature and the ambient temperature.

[0079] The actual electrical power consumed in an AP at time t is denoted P _e (t) and temperature change rate of an AP is proportional to the difference between the power consumed P _e (t) and the power dissipated, which in turn is proportional to T _AP (t) ~ T _a mb( - The temperature T _AP (t of an AP may thus be mathematically expressed by means of a differential equation: where and c ₂ are constants representing thermal mass and heat dissipation capability, respectively, of an AP.

[0080] Specifically, c _± quantifies at which rate the AP heats up when electrical power is consumed, and c ₂ quantifies at which rate the AP cools off by releasing heat to the environment. The solution to equation (1) is given by equation (2) below, i.e. by computing the convolution between the consumed power P _e (t) and the heat dissipation impulse response e ^-C2(t-T ) _:

[0081] As is understood, numerous models may be envisaged and computed to determine thermal state in the form of heat dissipation capability of an AP over time and depending on the load to which the AP is subjected, as well as ambient temperature. Equations (1) and (2) in the above provide an example of how the thermal state of an AP may be modelled.

[0082] Again with reference to Figure 6, where it was determined in step S101 using the measured AP idle temperature T _am b and the measured AP load temperature TAP, the thermal state of each AP loe-ioh is determined in step S101, this time in the form of heat dissipation capability Pa . In equation (1) hereinabove, Pa is represented by c ₂ in equation (1) hereinabove). A heat dissipation pulse response can be defined as e~ ^C2(J: ~ ^T and by c ₂ can be estimated from measurements.

[0083] Thus, it is determined whether or not the heat dissipation capability Pai, Pa2, Pag, Paa, of each respective AP loe-ioh complies with the predetermined thermal condition when tasks are being assigned in step S102. In this embodiment, the predetermined thermal condition is considered to be complied with if the heat dissipation capability Pa of each AP loe-iog exceeds a predetermined heat dissipation capability threshold value Pt. Further in this example, the predetermined thermal condition also stipulates that neither one of previously discussed Ti or T2 may exceed Tt, and as previously Ti is far below Tt while T4 is just below Tt leaving the fourth AP loh with a less power consumption margin than the remaining APs in the group. [0084] Thereafter, as previously described the expected channel state may optionally be determined in step Sioia - in the form of channel quality C - and taken into account by the CPU 21 upon determining power consumption at each AP in order to ensure that the power consumption does not increase to such an extent that the thermal state of one or more APs no longer complies with the predetermined thermal condition when the tasks are being assigned in step S102.

[0085] In this embodiment, the channels to be established with the second UE 23, third UE 24 and fourth UE 25 are all determined to have a quality C2, C3, C4 exceeding the channel quality threshold value Ct, while the channel to be established with the first UE 22 is poorer, i.e. Ci < Ct. Setting up a channel with the first UE 22 will thus require a higher power consumption at the AP serving the first UE 22, which should be carefully considered by the CPU 21 when assigning processing tasks in step S102.

[0086] As shown in Figure 6, it is assumed in step S102 that (a) the heat dissipation capability Pdi, Pd2, Pd3 of the first AP loe, the second AP lof and the third AP 10g, respectively, all exceeds the predetermined heat dissipation capability threshold value Pt with a high margin, while the heat dissipation capability Pd4 of the fourth AP loh is just above the predetermined heat dissipation capability threshold value Pt, the conclusion being that the fourth AP loh is at risk of being overheated if the fourth AP loh would be assigned a processing-heavy task

[0087] In step S102, due to the poor quality of the channel established with the first UE 22, the first AP loe and the second AP lof will share the serving of the first UE 22 by having the first AP loe handle UL communication and the second AP lof handling DL communication.

[0088] Further, since the fourth AP loh has a relatively low power dissipation capability Pd4 (i.e. just above Pt), the CPU 21 will conclude that the fourth AP loh cannot fully handle a UE, but will be assigned to handle DL communication for the fourth UE 25, as T2 otherwise likely would rise above Tt, wherein the predetermined thermal condition no longer is complied with.

[0089] Finally, the third AP 10g is assigned the remaining tasks of fully serving the second UE 23 and the third UE 24 and UL communication for the fourth UE 25. In this context, it maybe envisaged that Pd3 not only exceeds Pt with a sufficient margin, but exceeds Pt with a great margin thereby indicating that the third AP log may be assigned a processing-heavy task due to high-quality channel conditions. Thus, the third AP log is assigned a more power consuming task than the remaining APs in the group due to a low temperature Ti and high power dissipation capability Pdi, the rationale being that its temperature Ti still will not rise above Tt, The predetermined thermal condition will thus be complied with even if a burdening task is assigned to the third AP 10g.

[0090] In a further embodiment to be discussed in the following, step S101 may further include determining coefficients Ci and c ₂ of equations (1) and (2).

[0091] As previously described, in an initial step or on regular basis, the CPU 21 interacts with each AP loe-ioh in the group in order to acquire measured temperatures T _amb and T _AP (t) . This may be performed by first idling the APs for a predetermined period of time and then apply a load during a time period and at several occasions during that time period measure T _AP (t). Based on these measurements, and c ₂ can then be estimated.

[0092] For example, assuming that the CPU 21 subjects the APs loe-ioh to a load that is known to result in a power consumption equal to P _e (t) for 0 < t < T.

Furthermore, assuming that before time t = 0 the CPU 21 has controlled each AP to be idle for an extended time such that the AP is at ambient temperature at time t = 0. At time t = 0, the CPU 21 then inquires the AP for T _amb . Subsequently, during the time window from 0 to T, the AP is inquired N times to report its temperature (such as at time instances t ₂ , ..., t _w ) upon being subjected to a load, which results in the load temperature 7)ip(t _n ) being acquired.

[0093] Since T _amb hence is known and P _e (t) is known (it represents the load imposed on the AP), the two parameters that remain to determine are and c ₂ . The CPU 21 may estimate these parameters from the temperature measurements, for instance by performing a regression fit. An example of a regression criterion is the least-squares criterion of equation (3): [0094] The CPU 11 may use any standard optimization method to minimize the regression criterion with respect to and c ₂ and thereby obtain estimates of these coefficients.

[0095] Figure 7 shows a flowchart illustrating assignment of tasks to APs for serving a plurality of UEs according to an embodiment.

[0096] Similar to the embodiment described with reference to Figure 5, the CPU 21 determines a thermal state of each AP loe-ioh in the group in step S101 by measuring a current temperature of each AP loe-ioh is measured and it is assumed that the temperature of the first, second and third AP loe, lof, 10g is at temperature Ti while the temperature of the fourth AP loh is at a higher temperature T2.

[0097] Further, a channel state of channel(s) to be established with each UE 22- 25 is determined in step Sioia, for instance by determining a channel quality C based on a pilot signal being sent between the APs loe-ioh and the UEs 22-25.

[0098] A utility measure may be defined for each AP/, I = 1, ... , L, based on the current temperature T of each AP loe-iog and the expected channel quality C of each channel being established. As an example, this utility measure may be defined as in equation (4):

Utility; = cq x CS _{l k} — _l x TS _l (4)

[0099] where ai and Pi are optional weighting factors related to the channel state (CS) for each channel k and the thermal state (TS) for each AP 1, respectively. It is noted that factors ai and Pi may vary among the APs.

[00100] In this example, APs with high temperature will obtain a small utility measure since remaining thermal “headroom” for these APs is smaller, i.e. they are closer to overheating. Similarly, channels with high quality will result in a higher utility measure.

[00101] With reference to Figure 7, for UEk, for instance the first UE 22, the utility measure of equation (4) is computed for each APi in step Sioib.

[00102] Thereafter, in step S102, the APi with the highest computed utility measure in step Sioib is assigned a task - in this example the first AP loe is assumed to be assigned the task of serving first UE 22. [00103] Optionally in step S102, the CPU 21 further determines whether or not the task of serving the first UE 22 is expected to result in a power consumption of the first AP loe that causes the thermal state of the first AP loe to comply with a predetermined thermal condition. In this case, the temperature is not allowed to increase to such an extent that Utilityi < Tu, where Tu is a predetermined utility measure threshold which the utility measure is not allowed to fall below for any given AP to be assigned. If so, the AP with the next highest utility measure, e.g. second AP lof, will be evaluated for assignment and thus assessed with respect to Tu, and so on.

[00104] If it is not possible to assign another AP to the first UE 22, the process proceeds with assessing assignment for the second UE 23.

[00105] In optional step S103, it is determined whether or not a quality measure Q of the established communication channel between the first AP loe and the first UE 22 exceeds a predetermined quality threshold TQ. If so, it is determined in step S104 whether or not further UEs are to be assigned an AP, else the assigning process terminates. At this particular stage, only the first UE 22 have been assigned an AP, and thus the assigning continues with the second UE 23 until all four UEs 22-25 i ⁿ the group have been assigned an AP.

[00106] If in step S103 it is determined that the quality measure Q of the established communication channel between the first AP loe and the first UE 22 does not exceed the predetermined quality threshold TQ, step S102 is repeated where a non-assigned AP with a next-highest computed utility measure is assessed and potentially assigned to serve the first UE 22. The quality Q maybe determined based on e.g. CQI, pathloss, signal-to-interference-and-noise ratio (SINR), etc.

[00107] As is understood, should step S103 not be included in the assigning process of Figure 7, the CPU 21 would after having assigned to the first AP loe the task of serving the first UE 22 in step S103 proceed directly to step S104 and concluded that there are three remaining UEs 23-25 to consider and thus proceeding with the assigning of step S102 until all UEs 22-25 are served by at least one AP loe- loh.

[00108] Figure 8 illustrates an AP control device 40 configured to control operation of the group of APs loe-ioh being configured to serve the plurality of UEs 22-25, where the steps of the method performed by the AP control device 40 in practice are performed by CPU 21 embodied in the form of one or more microprocessors arranged to execute a computer program 17 downloaded to a storage medium 18 associated with the microprocessor, such as a Random Access Memory (RAM), a Flash memory or a hard disk drive, or contained in an internal memory of the CPU 21. The CPU 21 is arranged to cause the AP control device 40 to carry out the method according to embodiments when the appropriate computer program 17 comprising computer-executable instructions is downloaded to the storage medium 18 and executed by the CPU 21. The storage medium 18 may also be a computer program product comprising the computer program 17. Alternatively, the computer program 17 may be transferred to the storage medium 18 by means of a suitable computer program product, such as a Digital Versatile Disc (DVD) or a memory stick. As a further alternative, the computer program 17 may be downloaded to the storage medium 18 over a network. The CPU 21 may alternatively be embodied in the form of a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field-programmable gate array (FPGA), a complex programmable logic device (CPLD), etc. The AP control device 40 may further comprise a communication interface 19 (wired or wireless) over which it is configured to transmit and receive data.

[00109] The device 40 of Figure 8 may be provided as a standalone device or as a part of at least one further device. For example, the device 40 maybe provided in a node of a core network, or in an appropriate device of a radio access network (RAN), such as in a radio base station, in an internet server, etc. Alternatively, functionality of the device 40 may be distributed between at least two devices, or nodes. These at least two nodes, or devices, may either be part of the same network part (such as the core network) or may be spread between at least two such network parts.

[00110] Thus, a first portion of the instructions performed by the device 40 may be executed in a first device, and a second portion of the of the instructions performed by the device 40 may be executed in a second device; the herein disclosed embodiments are not limited to any particular number of devices on which the instructions performed by the device 40 maybe executed.

[00111] Hence, the method according to the herein disclosed embodiments are suitable to be performed by a device 40 residing in a cloud computational environment. Therefore, although a single processing circuitry in the form of a CPU 21 is illustrated in Figure 8, the processing circuitry maybe distributed among a plurality of devices, or nodes. The same applies to the computer program 17. Embodiments may be entirely implemented in a virtualized environment.

[00112] These aspects may, however, be embodied in many different forms and should not be construed as limiting; rather, these embodiments are provided by way of example so that this disclosure will be thorough and complete, and to fully convey the scope of all aspects of invention to those skilled in the art. Like numbers refer to like elements throughout the description.

[00113] The aspects of the present disclosure have mainly been described above with reference to a few embodiments and examples thereof. However, as is readily appreciated by a person skilled in the art, other embodiments than the ones disclosed above are equally possible within the scope of the invention, as defined by the appended patent claims.

[00114] Thus, while various aspects and embodiments have been disclosed herein, other aspects and embodiments will be apparent to those skilled in the art. The various aspects and embodiments disclosed herein are for purposes of illustration and are not intended to be limiting, with the true scope and spirit being indicated by the following claims.