WO2023225798A1

WO2023225798A1 - Method and control system for controlling power amplifier based on traffic load information

Info

Publication number: WO2023225798A1
Application number: PCT/CN2022/094488
Authority: WO
Inventors: Bengt THOMAS LEJON; Haiying CAO
Original assignee: Telefonaktiebolaget Lm Ericsson (Publ); Haiying CAO
Priority date: 2022-05-23
Filing date: 2022-05-23
Publication date: 2023-11-30

Abstract

A control system (300) and method therein for controlling a power amplifier (320) in a transmitter (310) are disclosed. The control system (300) obtains traffic load information for the transmitter (310) and controls the power amplifier (320) based on the traffic load information by adjusting a supply voltage and/or a bias voltage of the power amplifier (320). The supply voltage and/or a bias voltage of the power amplifier (320) may be adjusted by a hysteresis function based solution, a machine learning trained traffic load model based solution, a prewarning signal or a RAN traffic characteristic based solution to balance the power capability and efficiency of the power amplifier to the instantaneous traffic load of the transmitter (310) and by that triggering of linearization fault handlings may be minimized and the transmitter efficiency may be maximized.

Description

METHOD AND CONTROL SYSTEM FOR CONTROLLING POWER AMPLIFIER BASED ON TRAFFIC LOAD INFORMATION

TECHNICAL FIELD

Embodiments herein relate to a method and control system for controlling a power amplifier. In particular, the embodiments relate to controlling a power amplifier comprised in a radio transmitter to optimize radio transmitter efficiency, and a radio unit comprising the power amplifier in a radio access network.

BACKGROUND

In a typical wireless communication network, wireless devices, also known as wireless communication devices, mobile stations, stations (STA) and/or user equipment (UE) , communicate via radio access network (RAN) to one or more core networks (CN) . The RAN covers a geographical area which is divided into service areas or cell areas, which may also be referred to as a beam or a beam group, with each service area or cell area being served by a radio network node such as a RAN node, a Wi-Fi access point or a radio base station (RBS) , which in some networks may also be denoted as for example, “NodeB” or “eNodeB” or “gNB” . A radio network node such as RAN node or RBS may also be referred as a cellular site. A service area or cell area is a geographical area where radio coverage is provided by the radio network node.

A RAN node consists of a baseband unit (BU) , a radio unit (RU) or remote radio unit (RRU) and an antenna or an antenna array. In a RRU, more than 60%power is consumed by power amplifier (PA) . Thus, the efficiency of PA is important to the overall efficiency of the RRU. One of the most effective methods to improve PA efficiency is to reduce its supply voltage. However, within a certain supply voltage range, by reducing PA’s supply voltage, the power capability decreases and the nonlinearity of PA increases so that the spectral performance degrades and eventually the transmission of radio signals is not possible anymore when the supply voltage is too low. Therefore, it would be desirable to make an optimum trade-off between the efficiency and linearity for PA design.

A PA is usually designed with some margins or power headroom for the supply voltage to cover different circumstances. Such power headroom is usually used to cover power amplifier batch variation, temperature drift, aging and so forth. As the available power is proportional to the supply voltage and due to the linear operation of the power amplifier, the efficiency of the power amplifier drops with increased power headroom. A way to improve radio efficiency during operation is to optimize the supply voltage of PA to minimize the power headroom. Therefore, this power headroom can be utilized to increase efficiency by increasing the supply voltage.

Linearization fault handling may be used to reset the supply voltage to its maximum voltage level based on linearization error. The linearization error is a measurement on the difference of an output signal from a PA to an input signal to the PA. A linearization fault handling may be triggered by a fast degradation in the linearization error estimate. When a linearization fault handling is triggered, the supply voltage of PA is reset to its maximum voltage level to restore the PA capability to generate maximum power. A linearization fault may happen, e.g. during a rapid increasement of traffic load if the supply voltage is lower than what is needed to deliver the wanted power for the traffic load. Triggering of linearization fault handling causes stress to PA, degrades PA lifetime and even radio performance can be affected, such as re-transmission or even dropped calls.

WO2018223256 discloses a PA controller and method for controlling PA. The PA controller lowers the supply voltage as long as the linearization error estimate is not increasing and increases the supply voltage if the linearization error estimate is increasing.

An issue of controlling PA is to be able to follow power changes during operation. If the supply voltage is too low to allow for an increased power, there will be spectral issues and unstable operation of the control loops in the transmitter systems. In this case, linearization fault handling might happen. If the supply voltage is too high, then efficiency is degraded, especially when the traffic load is low, setting high supply voltage will result in low efficiency.

SUMMARY

It is therefore an object of embodiments herein to provide an improved method for controlling power amplifier to optimize radio transmitter efficiency and minimize triggering of linearization fault handling.

A control method which incorporates traffic load information or RAN traffic load statistics information is proposed according to embodiments herein to improve radio transmitter efficiency, increase the stability of the control system and reduce linearization faults.

According to a first aspect of embodiments herein, the object is achieved by a control system and method therein for controlling a power amplifier. The power amplifier is comprised in a transmitter of a radio unit in a radio access network (RAN) . The method comprises obtaining traffic load information for the transmitter and controlling the power amplifier based on the traffic load information by adjusting a supply voltage and/or a bias voltage of the power amplifier.

There are different alternatives to control the power amplifier based on the traffic load information.

According to some embodiments herein, the supply voltage may be adjusted such that a slope of reducing supply voltage over time is less than a slope of decreasing traffic load over time and a slope of increasing supply voltage over time is larger than a slope of increasing traffic load over time.

According to some embodiments herein, the supply voltage may be adjusted such that the supply voltage is increased faster when the traffic load is increasing and decreased slower when the traffic load is decreasing.

That means a hysteresis function may be implemented in the control system for controlling the power amplifier. With the hysteresis function, it takes longer time to reduce the voltage when the voltage is to be decreased and allows faster recovery of the voltage to the highest level when the voltage is to be increased. The hysteresis time may be seconds to minutes or to several tens of minutes. This control method may handle a generalized behaviour of random traffic load and that the high traffic load density can vary over time with a ramp up and ramp down behaviour. The control system with a hysteresis function is mainly controlling the voltage when the traffic load is in a more steady state condition and less control in a ramp up and ramp down scenario. The reason for a slow lowering of the voltage is due to that the risk of linearization fault handling increases with lower voltage.

According to some embodiments herein, the supply voltage and/or bias voltage may be adjusted based on traffic load model. The traffic load model may be trained by machine learning to the actual traffic load for the transmitter over time such that the model is optimized and unique for the transmitter. Then the optimized and unique traffic load model is used to control the supply/bias voltage of the PA’s to minimize the number of linearization fault handlings.

This method allows an improved total efficiency for the transmitter since the traffic load model models when the traffic load is at a steady state e.g., a low traffic load state during night and a high traffic load state in the morning for the specific sector that the transmitter operates in and predicts when in time the ramping between the low and high traffic load states occurs and by that less margin may be put to the supply voltage of the PA over time. This method also has an advantage of being a radio unit standalone solution without adding extra complexity to coordinate with baseband unit and core network.

According to some embodiments herein, the supply voltage and/or bias voltage may be adjusted based on a prewarning signal generated based on the traffic load information for the transmitter from a baseband unit in the RAN. The baseband unit has the knowledge of actual traffic in advance and then may generate the prewarning signal in time for the controller to be able to change the supply voltage and/or bias voltage before the actual traffic signal arrives to the PA. That means the prewarning signal may be generated in advance based on traffic load changes such that the supply voltage and/or bias voltage is adjusted in time before the traffic load changes happen. In this way, linearization faults may be prevented.

According to some embodiments herein, the supply voltage and/or bias voltage may be adjusted based on traffic load patterns generated based on accumulated traffic load statistics of a RAN over a certain time period.

According to a second aspect of embodiments herein, the object is achieved by a baseband unit and method therein for generating a prewarning signal based on traffic load information of a transmitter. The prewarning signa is used for controlling a power amplifier comprised in the transmitter. The method comprises obtaining traffic load information for the transmitter and comparing a traffic load to be handled by the transmitter with a traffic load capacity of the transmitter. If the traffic load to be handled by the transmitter is larger than the traffic load capacity of the transmitter, generating a prewarning signal to increase a supply voltage or a bias voltage of the power amplifier. If the traffic load to be handled by the transmitter is lower than the traffic load capacity of the transmitter, generating a prewarning signal to decrease a supply voltage or a bias voltage of the power amplifier.

Embodiments herein use knowledge about the traffic behavior in a cellular network and the limitations in the function of a linearization system of power amplifiers combined with either a hysteresis function based solution, a machine learning trained traffic load model, a prewarning signal based solution or a RAN traffic characteristic based solution to balance the PA power capability and efficiency to the instantaneous traffic load of the transmitter and by that triggering of linearization fault handlings may be minimized and radio transmitter efficiency may be maximized.

Therefore, the embodiments herein provide an improved method for controlling power amplifier to optimize radio transmitter efficiency and minimize triggering of linearization fault handling.

BRIEF DESCRIPTION OF THE DRAWINGS

Examples of embodiments herein are described in more detail with reference to attached drawings in which:

Figure 1 is a schematic overview of a wireless cellular communication network comprising one or more RANs;

Figure 2 is a schematic overview of a wireless cellular communication network illustrating cells served by base stations and examples of different type of service areas and examples of different traffic patterns of the service areas;

Figure 3 is a schematic block diagram showing a control system for controlling PA according to embodiments herein;

Figure 4 is a flow chart illustrating a method for controlling PA according to embodiments herein;

Figure 5 are diagrams showing control system behaviors;

Figure 6 is an example machine learning system according to embodiments herein;

Figure 7 is a diagram showing control system behavior according to an embodiment herein;

Figure 8 is a diagram showing control system behavior according to an embodiment herein; and

Figure 9 is a flow chart illustrating a method for generating a prewarning signal according to the embodiments herein.

DETAILED DESCRIPTION

Embodiments herein relate to cellular communications networks in general. Figure 1 is a schematic overview depicting a communication network 100. The communication network 100 may be a wireless communications network comprising one or more RANs, and one or more CNs. The communication network 100 may use a number of different technologies, such as Wi-Fi, Long Term Evolution (LTE) , LTE-Advanced, New Radio (NR) , Wideband Code Division Multiple Access (WCDMA) , Global System for Mobile communications/enhanced Data rate for GSM Evolution (GSM/EDGE) , Worldwide Interoperability for Microwave Access (WiMax) , or Ultra Mobile Broadband (UMB) , just to mention a few possible implementations.

Network nodes operate in the wireless communication network 100 such as a first network node 110 and a second network node 120. The first and

second network node

110, 120 may be any of RAN node, such as gNB, eNB, en-gNB, ng-eNB, gNB etc. The first network node 110 provides radio coverage over a geographical area, a service area 11, which may also be referred to as a beam or a beam group. The second network node 120 provides radio coverage over a geographical area, a service area 12, which may also be referred to as a beam or a beam group.

The first and

second network nodes

110 and 120 may be a transmission and reception point e.g. a radio access network node such as a Wireless Local Area Network (WLAN) access point or an Access Point Station (AP STA) , an access controller, a base station, e.g. a radio base station such as a NodeB, a gNB, an evolved Node B (eNB, eNode B) , a base transceiver station, a radio remote unit, an Access Point Base Station, a base station router, a transmission arrangement of a radio base station, a stand-alone access point or any other network unit capable of communicating with a wireless communication device within the service area served by the respective first and

second network nodes

110 and 120 depending e.g. on the radio access technology and terminology used.

In the wireless communication network 100, one or more

wireless communication devices

130, 131 such as a UE, a mobile station or a wireless terminal communicates via one or more RANs to one or more CNs. It should be understood by the skilled in the art that “wireless communication device” is a non-limiting term which means any terminal, wireless communication terminal, user equipment, Machine Type Communication (MTC) device, Device to Device (D2D) terminal, or node e.g. smart phone, laptop, mobile phone, sensor, relay, mobile tablets or even a small base station communicating within a cell.

The first and

second network nodes

110 and 120 each comprises a

RU

111, 121, a

BU

112, 122, and an antenna or antenna array 113, 123. The first and

second network nodes

110 and 120 may communicate with the

wireless communication device

130, 131 with Downlink (DL) transmissions to the

wireless communication device

130, 131 and Uplink (UL) transmissions from the

wireless communication device

130, 131. For example, during UL transmissions, signals from the UE 130 reaches a CN 140 after being received by the RU 111 and transformed into a digital format by the BU 112, and during DL transmissions, data from the CN 140 is received by the BU 112 and transformed into radio signals and transmitted by the RU 111 to the UE 130.

As discussed in the background, an issue of controlling PA is to be able to follow power changes during operation. Another issue is that the traffic load is varying in a random way and the information of the actual power level needed in a certain time is unknown in advance by the radio transmitter and also that a significant time is needed to adjust the supply voltage of PA due to large capacitors that need to be charged or discharged.

There are different communication traffic behaviors in a wireless cellular communication network depending on different type of areas covered by the communication network 100. Figure 2 shows a conceptual cellular communication network 200 where cells served by network nodes or radio base stations, indicated by

cellular sites

211, 212, 213, 214 in a three-sector configuration are shown. Examples of different type of areas and typical traffic behavior over time in different areas are also shown to illustrate that different sectors have different traffic depending on the type of the area. For example, the service area of the network node 211 covers a city center where there has a peak traffic during daytime. A commuter area covered by the

network nodes

212, 214 may have two distinct peaks of traffic, one in the morning and one in the afternoon but have low traffic in between compared to a residential area covered by the

network nodes

213, 214 that has low traffic during the day and high traffic during the evening and into the night because people are at work and school during the day.

From traffic behavior it is seen that there are often peak hours with high traffic load and low traffic load in between for many hours that has short bursts of high traffic load. The traffic loads are also very different in both peak and average load over time for different base station sites. Depending on the actual traffic load situation, it may only need to increase the supply voltage to some certain values but not to the maximum level. As even with 100%traffic load, the supply voltage may still be low due to the power headroom level. When linearization error is too high, it then can handle power changes. It is also important to minimize triggering of linearization fault handling to improve radio transmitter performance and PA lifetime as this is causing stress to a PA, degrade PA’s lifetime and even radio performance can be affected, such as re-transmission or even dropped calls.

The principle of the embodiments herein is that the radio transmitter may be more stable without resetting the supply voltage to the maximum leveland the radio transmitter efficiency may be improved, if the traffic behavior or traffic load information at each site or network node is known or obtained and a change in needed power in a PA is predicted in advance. The supply and/or bias voltage of the PA can be adapted based on traffic load information so that the PA can deliver the needed power before the high-power situation occurs. If this is not possible to do, it is likely that the PA control system is pushed into a saturated state so that a linearization failure occurs.

According to embodiments herein, a control method which incorporates traffic load information of the radio transmitter or RAN traffic statistics information is proposed to improve power consumption of a PA and reduce radio linearization faults.

Figure 3 shows a control system 300 in a radio transmitter TX 310 for controlling a PA 320 according to embodiments herein. The control system 300 comprises a PA controller PAC 330, a first direct current to direct current (DC/DC) converter DC/DC 1, a second DC/DC converter DC/DC 2, and a Digital Pre-distorter Linearization DPL 340. The DPL 340 receives a feedback signal FB from the PA 320 and a DL signal DLS from a baseband unit BU 350. The DPL 340 generates a linearization error estimate signal LES and a linearization fault signal LF to the PAC 330. As an example, the PA 320 is shown as a metal-oxide-semiconductor field-effect transistor (MOS FET) or a Gallium Nitride (GaN) based transistor with a supply voltage Vdd and a gate bias voltage Vgb. The radio transmitter TX 310 is a part of a RU and is configured to transmit radio signals in a

communication network

100, 200.

According to the embodiments herein, the power amplifier PA 320 is controlled based on the traffic load information obtained for the radio transmitter TX 310 by adjusting the supply voltage Vdd and/or bias voltage Vgb of the PA 320. The PAC 330 controls the first and second DC/DC converters which provide the supply and bias voltages to the PA 320.

The linearization error estimate signal LES is a measurement on the difference of the signal after the PA 320 to an original input signal. The DPL 340 estimates the PA 320 nonlinearity and adds an additional correction to the original input signal to compensate against the PA 320 nonlinearity. If the PA 320 nonlinearity is below an upper limit, the additional correction can secure the radio performance. If the PA 320 nonlinearity exceeds the upper limit, the DPL 340 will fail to compensate and the radio performance will no longer meet the requirement.

The linearization fault signal LF is generated by a fast degradation in the linearization error estimate. The linearization fault signal LF is then used to trigger a reset of the Vdd to the PA 320 to its maximum voltage value to restore the capability of generating maximum power. This may happen, as an example, during a rapid increasement of traffic load.

As mentioned hereinbefore, the tradeoff between the efficiency and linearity is critical for the PA design. In a traditional radio transmitter, a few volts are kept as a voltage margin to cover all variations in mass production and from aging effects, thereby achieving acceptable linearity. However, this is not favorable for energy efficiency because of the following reasons. Firstly, the voltage margin cannot be eliminated during production since the aging behavior is not deterministic. Secondly, in mass production, the voltage margin is employed in all products. But in fact, the worst case can only be reached by a small portion of the PA products. For most PAs, the margin results in energy wastes. Thirdly, during operation, only a small portion of carrier setups needs a full voltage. Furthermore, considering the traffic behavior where there is often low traffic for many hours between short period of high traffic peaks, so voltage margin is purely waste for most of times.

There are different alternatives to control the PA 320. As shown in Figure 3, a hysteresis function HF 331 may be implemented in the PAC 330 when adjusting the supply voltage Vdd of the PA 320. A traffic load model TM 332 may be generated based on the traffic load information and the PAC 330 may control the PA 320 using the generated traffic load model. A prewarning signal PWS 333 of a traffic change may be generated and the PAC 330 may control the PA 320 using the prewarning signal PWS 333. A RAN traffic pattern RAN TP 334 may be generated based on RAN traffic information and the PAC 330 may control the PA 320 based on the generated traffic pattern RAN TP 334.

A method for controlling a power amplifier 320 in a radio transmitter 310 dynamically based on traffic load information according to embodiments herein will be described in detail with reference to Figure 4. The method comprises the following actions which may be performed in any suitable order.

Action 410

The control system 300 obtains traffic load information for the transmitter 310.

To obtain traffic load information for the transmitter 310, the control system 300 may measure traffic load of the transmitter 310 over time or receive the traffic load information of the RAN from a core network CN 140 in the communication network 100.

The traffic load may be measured or represented by the number of active Physical Resource Blocks (PRBs) , the integration of a baseband signal in frequency domain over a frequency band, the sum of all resource blocks over the frequency range of a carrier, or the integration of a baseband signal in time domain over a certain time period e.g. a symbol time and measured with a certain bandwidth that covers all carriers that are transmitted by the power amplifier. An increased traffic load means a high output power is needed from the PA and a decreased traffic load means a low output power is needed from the PA.

Action 420

The PA controller 330 controls the power amplifier 320 based on the traffic load information by adjusting a supply voltage Vdd and/or a bias voltage Vgb of the power amplifier PA 320.

The different alternatives to control the PA 320 will be described in detail by the following actions.

Action 421

According to some embodiments herein, a hysteresis function HF 331 may be implemented when adjusting the supply voltage Vdd of the PA 320. The hysteresis function HF 331 means that the PA controller 330 is slow to lower the voltage to improve efficiency and fast to raise it to improve output power capability when a user requires data that requires higher output power. In this way, the number of linearization faults maybe reduced when traffic load in the network is decreasing or at a low level with sudden short time high traffic load pulses. This embodiment may put the radio transmitter in an optimized efficiency mode when traffic load is stable like during the night but not during the first time period of the low traffic load period thereby reducing the number of linearization faults when traffic load is in a transition from low to high or high to low that may have a more random pattern with short traffic peaks. This embodiment may also allow for a more optimized Vdd level for the traffic load at a certain time as the control system 300 for the Vdd can be in a normal operation during a longer time period.

In other words, the PAC 330 is hysteresis on lowering the voltage so that it takes longer time to reduce the voltage and allow fast recovery of the voltage to the highest level. The hysteresis time may be seconds to minutes or to several tens of minutes. This may be used to single users with high traffic load but seldom, for example one person that starts a high resolution YouTube clip or movie that is buffered, the full bandwidth, e.g. all PRBs of the network is used during that time, or a single user that starts speed test. This is to reduce the number of linearization fault handlings when traffic in the network is decreasing or at a low level with sudden short time high traffic pulses. Raising supply voltage faster can improve output power capability when a user requires data that requires higher output power. This embodiment may handle a generalized behaviour of random traffic and that the high traffic load density can vary over time with a ramp up and ramp down behaviour and that the control system 300 by a hysteresis is mainly controlling the voltage when the traffic load is in a more steady state condition and less control in a ramp up and ramp down scenario. The reason for a slow lowering of the voltage is due to that the risk of linearization faults increase with lower voltage. Maximum voltage is always safe for all traffic load conditions but with lower efficiency.

Figure 5 (a) is a diagram showing a system behavior of a control system without hysteresis function. The supply voltage Vdd is shown with dotted line, and the traffic load or output power of the radio transmitter is shown with solid line. As can be seen, the supply voltage is reset to the maximum as long as a linearization fault happens, as indicated by reference numbers 511-515. Figure 5 (b) is a diagram showing an improved system behavior of the control system with hysteresis function according to embodiments herein. As can be seen, the number of resetting the supply voltage to maximum is reduced from 5 to 3, as indicated by reference numbers 521-523. With hysteresis function implement in the control system, the supply voltage Vdd is decreased slower than the traffic load decreasing, as indicated by

reference numbers

524, 525, and the supply voltage is successfully controlled such that less linearization faults happen and resetting the supply voltage to maximum is avoided, as indicated by

reference numbers

526, 527, which improves the power efficiency.

Therefore, according to some embodiments herein, the supply voltage may be adjusted such that a slope of reducing supply voltage over time is less than a slope of decreasing traffic load over time and a slope of increasing supply voltage over time is larger than a slope of increasing traffic load over time. Alternatively, the supply voltage of the power amplifier may be adjusted by increasing the supply voltage faster when the traffic load is increasing and decreasing the supply voltage slower when the traffic load is decreasing.

Action 422

According to some embodiments herein, the control system 300 may generate a traffic load model TM 332 for the transmitter 310 based on the measured traffic load over time or the received traffic load information.

The traffic information in the RAN network may include several sectors or sites. The traffic model TM 332 may be generated and trained by machine learning based on the traffic load information for the transmitter 310 over time such that the traffic model TM 332 is optimized and unique for the transmitter 310. The unique traffic model for this cellular sector that the transmitter 310 operates in is modelled for this transmitter 310 so that the operation and efficiency is optimized for each cellular sector individually.

The traffic model may be a mathematical function with parameters generated based on the traffic load information for the transmitter 310 over time.

Different traffic load models for different time scales may be generated or trained for the transmitter 310 to match an actual adjustment time for adjusting the supply voltage and/or the bias voltage of the power amplifier PA 320.

The PA controller 330 may control the power amplifier PA 320 based on the traffic load information by adjusting the supply voltage and/or the bias voltage of the power amplifier using the generated traffic load model.

Figure 6 shows an example machine learning system 600 to generate a traffic load model, e.g. the a traffic load model TM 332 shown in Figure 3. The machine learning system 600 comprises a power detector 610 to detect a power of a downlink signal DLS. The power of a downlink signal DLS represents the traffic load and that signal power is to be handled by the PA 320. The traffic model TM 332 may be trained to a specific traffic pattern needed at a cellular site with an expected supply voltage needed over time. The traffic model TM 332 may be a Neural Network or a mathematical function with parameters describing the same. The power of the DL signal is detected and a training controller 620 trains a traffic model over time such as per hour, day, week, season so that the PA controller 330 can have an assumed value on the supply/bias voltage depending on the time of the day at that individual site optimized to historical traffic and adjust the Vdd/Vgb based on the historical traffic data. The time scale of the traffic load may be defined in the model as such that different models for different time scales may be generated to match the actual PA adjustment time in the control system 300. Depending on the implementation methods and actual hardware used, the adjustment time maybe varied. This may work as many cellular sites have similar traffic between days. This may remove many linearization faults and enable a more optimized Vdd level to the level of traffic load as the model is improving to fit the local traffic condition and changes over time.

The traffic load model TM 332 may be generated and trained using any type of machine learning methods, such as Supervised Machine Learning, Unsupervised Machine Learning, Reinforcement Machine Learning, Semi-Supervised Machine Learning etc.

Figure 7 is a diagram showing a system behavior of the control system 300 optimized by machine learning. The traffic model TM 333 trained by machine learning can predict traffic load changes over time and the supply voltage is adjusted based on the traffic model TM 333. As can be seen, the supply voltage follows the traffic load changes over time but not when there are traffic peaks during a short period. For example, the supply voltage has not been reset to maximum when there are traffic peaks during a short period, as indicated by

reference number

711, 712.

Action 423

According to some embodiments herein, the PA controller 330 may receive a prewarning signal PWS 333. The prewarning signal PWS 333 may be generated by the baseband unit BU 350 based on the traffic load information for the transmitter 310. The PA controller 330 may control the PA 320 based on the traffic load information by adjusting the supply voltage Vdd and/or the bias voltage Vgb of the power amplifier PA 320 using the prewarning signal PWS 333.

The baseband unit BU 350 knows about the actual traffic in advance and may then generate a prewarning signal in time for the PA controller 330 to be able to change the voltage before the actual signal arrives to the PA 320 and by that prevent a linearization fault.

The timing of the prewarning signal is critical. By using the prewarning signal, a more optimized supply voltage level adapted to the traffic condition can be present and by this embodiment the ideal maximized efficiency can be reached by the transmitter TX 310 with no linearization faults but with an added complexity to add a function to calculate the total power of carrier signals in advance. Alternatively, the transmitter 310 may also inform the baseband unit BU 350 on how much traffic can be allocated in a time slot to avoid linearization fault.

Figure 8 is a diagram showing a system behavior of the control system 300 when using the prewarning signal PWS 333 to adjust the supply voltage of PA 330. As can be seen, the supply voltage totally follows the traffic changes over time. There is no resetting of the supply voltage to maximum even when traffic peaks happen during a short period.

Action 424

According to some embodiments herein, the control system 300 may receive accumulated traffic load statistics of the RAN node 110/120 over a certain time period from the core network CN 140 in the communication network 100 and generate traffic load patterns for the transmitter TX 310 from the received accumulated traffic load statistics over a certain time period. This embodiment is indicated by RAN TP 334 in Figure 3.

The PA controller 330 may adjust the supply/bias voltage of the power amplifier PA 320 based on the traffic load information by adjusting the supply/bias voltage of the power amplifier PA 320 based on the traffic load patterns RAN TP 334.

According to some embodiments herein, the controlling of the power amplifier PA 320 by adjusting a supply voltage and/or a bias voltage of the power amplifier PA 320 may be further based on the performance of the transmitter TX 310. When the supply voltage and/or a bias voltage of the power amplifier PA 320 is to be adjusted based on the traffic load information, the method may further comprise the following actions:

Action 430

The control system 300 determines the performance of the transmitter TX 310 by determining error vector magnitude (EVM) or operating band unwanted emissions of the transmitter TX 310.

Action 432

The control system 300 compares the performance of the transmitter TX 310 with a performance threshold region comprising a first threshold Th1 and a second threshold Th2.

Action 434

The control system 300 adjusts the supply voltage and/or the bias voltage in response to a performance comparison result.

The control system 300 may increase the supply voltage and/or the bias voltage if the performance of the transmitter TX 310 is less than the first threshold Th1.

The control system 300 may decrease the supply voltage and/or the bias voltage if the performance of the transmitter TX 310 is larger than the second threshold Th2.

The control system 300 may keep the supply voltage and/or the bias voltage unchanged if the performance of the transmitter is within the performance threshold region.

A method performed in a

baseband unit BU

112, 122 for generating a prewarning signal PWS 333 based on the traffic load information of a transmitter TX 310 will be described with reference to Figure 9. The method comprises the following actions.

Action 910

The

baseband unit BU

112, 122 obtains or detects incoming traffic load for the transmitter TX 310. The

baseband unit BU

112, 122 may detect the incoming traffic load by integrating the baseband signal in frequency domain over a certain frequency range and a certain time period as defined by the symbol time, or summing all the resource blocks or counting the number of active resource blocks compared to the total number of resource blocks, or integrating the baseband signal in time domain over a certain time period e.g. the symbol time and over a certain bandwidth e.g. the carrier bandwidth.

Action 920

The

baseband unit BU

112, 122 compares the traffic load to be handled by the transmitter TX 310 with a current traffic load capacity of the transmitter TX 310. The traffic load capacity of the transmitter TX 310 may be defined by a mathematical function versus the actual PA voltage. For example, Traffic load capacity=A+B＊V^2, where V is PA voltage, A and B are constants that may be tuned during the design of PA which model the measured maximum power versus PA voltage.

Action 931

The

baseband unit BU

112, 122 generates a prewarning signal to increase a supply voltage and/or a bias voltage of the power amplifier PA 320, if the traffic load to be handled by the transmitter TX 310 is larger than the current traffic load capacity of the transmitter TX 310. The prewarning signal may be a signal indicating how much the supply/bias voltage should increase.

Action 932

The

baseband unit BU

112, 122 generates a prewarning signal to decrease a supply voltage and/or a bias voltage of the power amplifier PA 320, if the traffic load to be handled by the transmitter is lower than the current traffic load capacity of the transmitter can handle. The prewarning signal may be a signal indicating how much the supply/bias voltage should decrease.

Action 940

The

baseband unit BU

112, 122 generates a prewarning signal to inform the baseband unit itself to spread the traffic load such that the large data packet should be less than the current traffic load capacity of the transmitter TX 310 can handle.

When generating the prewarning signal, the

baseband unit BU

112, 122 needs to keep the delay Td and ramping time Tr of the supply voltage in mind, which may be preconfigured or signaled from the transmitter 310 beforehand. The delay time is defined as the time between the PA controller send a signal to increase the supply voltage to the supply voltage starts to ramp up and the ramping time Tr is defined as the time for the power or supply voltage of the PA ramps up from a minimum power percentage to 100%. For example, it takes Tr=34 ms for a certain PA model to ramp up from 10%to 100%power, e.g. Vdd changes from 14.6 V to 24 V i.e., an average raise rate of 2.68%/ms.

The

baseband unit BU

112, 122 estimates or predicts traffic load characteristics and calculates a minimum power percentage with a probability Pm. The minimum power percentage reflects the incoming traffic loads that the transmitter should be able to handle in the near future, i.e. within the sum of the delay time Td and ramping time Tr, Td+Tr. The probability Pm is a tradeoff between traffic load impact and power saving level. The probability Pm is the prediction or estimation done in the

baseband unit

112, 122. Because it is not 100%sure what kind of traffic load is coming in, so there is a probability that the

baseband unit

112, 122 may estimate the traffic load wrong. For example, if it estimates there is 80%of the chance that there is 100%traffic load is coming in, as such it may generate a prewarning signal to increase the supply/bias voltage to a level that corresponds to 100%traffic load. However, only 60%traffic load is actually coming in. In this case, less power is saved compared to if the supply/bias voltage is not increased to 100%traffic load.

When the traffic load to be handled by the transmitter TX 310 is larger than the current traffic load capacity of the transmitter TX 310 can handle, there are several options to deal with the situation, as described in

Actions

931 and 940, i.e. to increase the power of PA 330 or spread the traffic load.

To spread the traffic load, the

baseband unit BU

112, 122 may enforce a number of Physical Resource Block (PRB) limit for each carrier signal.

The PRB limit may be based on the calculated minimum power percentage, while the transmitter power in the end can decrease to a level which is higher than this minimum power percentage, because Vdd may be impacted by some other factors as well, e.g. temperatures, digital pre-distortion status etc.

The PRB limit may be based on the minimum power percentage signaled from the transmitter 310 so that the true or current transmitter power can be reflected. However this requires a feedback loop so longer delay in the control loop.

If the estimated or predicted traffic load increases and Pm drops to an unacceptable level, the

baseband unit BU

112, 122 may generate a prewarning signal to inform the PA controller to control the PA to “power ramp up to a certain level” . Ideally, after informing the PA controller, the

baseband unit BU

112, 122 needs to hold the current number of PRB limit for a time period Td, then apply the ramping slope Tr as the increase rate of PRB allocation limit.

The power of the PA 330 can ramp down very slowly which may take several or tens of seconds from 100%to 10%. Therefore, too quick decrease of the number of PRB allocation is not needed and can waste power resource. Ramping slope may be either signaled from the transmitter in real time or preconfigured at the

baseband unit BU

112, 122. The

baseband unit BU

112, 122 can follow the ramping curve so that the number of PRB decrease rate is not quicker than the actual power decrease rate.

Since the power of the PA 330 can ramp down very slowly, the power saving is therefore not expected to benefit from short term traffic variation, e.g. ms or sec level. Instead, the main target is to address long term traffic change e.g., difference between the busy and non-busy hours.

Simulations have been done for a PA controlled according to the method disclosed in WO2018223256 and the method according to embodiments herein for all traffic scenarios with 10%-100%traffic load. Embodiments herein can achieve lower Vdd values and thus lower power consumption under traffic scenarios with 10%-85%traffic load. Particularly, embodiments herein can achieve 4.2%power saving with 50%traffic load and 10.5%power saving with 10%traffic load compared to the solution proposed in WO2018223256.

To summarize, embodiments herein improve the stability of the PA control system 300 and efficiency of the transmitter TX 310. That is when the PA control system uses prior information such as traffic load or signal power, the PA control system can achieve a faster convergence and the supply voltage Vdd can reach the lowest possible level without triggering linearization fault handlings.

When using the word "comprise" or “comprising” it shall be interpreted as non-limiting, i.e. meaning "consist at least of" .

The embodiments herein are not limited to the above described preferred embodiments. Various alternatives, modifications and equivalents may be used. Therefore, the above embodiments should not be taken as limiting the scope of the invention, which is defined by the appended claims.

Claims

A method for controlling a power amplifier (320) , wherein the power amplifier (320) is comprised in a transmitter (310) of a radio unit (111, 121) in a radio access network, RAN (110, 120) , the method comprises:

obtaining (410) traffic load information for the transmitter (310) ; and

controlling (420) the power amplifier (320) based on the traffic load information by adjusting a supply voltage and/or a bias voltage of the power amplifier (320) .
The method according to claim 1, wherein adjusting the supply voltage of the power amplifier based on the traffic load information comprises adjusting (421) the supply voltage such that a slope of reducing supply voltage over time is less than a slope of decreasing traffic load over time and a slope of increasing supply voltage over time is larger than a slope of increasing traffic load over time.
The method according to any one of claim 1, wherein adjusting the supply voltage of the power amplifier based on the traffic load information comprises increasing the supply voltage faster when the traffic load is increasing and decreasing the supply voltage slower when the traffic load is decreasing.
The method according to claim 1, wherein obtaining (410) traffic load information for the transmitter (310) comprises measuring traffic load of the transmitter (310) over time, and the method further comprising:

generating (422) a traffic load model for the transmitter (310) based on the measured traffic load over time.
The method according to claim 1, wherein obtaining (410) traffic load information for the transmitter (310) comprises receiving the traffic load information of the RAN (110, 120) from a core network (140) in a communication network (100) ; and the method further comprising:

generating (422) a traffic load model (332) for the transmitter (310) based on the received traffic load information.
The method according to any one of claims 4-5, wherein the traffic load model (332) is generated and trained by machine learning based on the traffic load information for the transmitter (310) over time such that the model is optimized and unique for the transmitter (310) .
The method according to any one of claims 4-6, wherein the traffic load model (332) is a mathematical function with parameters generated based on the traffic load information for the transmitter (310) over time.
The method according to any one of claims 4-7, wherein controlling (420) the power amplifier based on the traffic load information comprises adjusting (422) the supply voltage and/or the bias voltage of the power amplifier (310) using the generated traffic load model (332) .
The method according to any one of claims 4-8, wherein different traffic load models for different time scales are generated and trained for the transmitter (310) to match an actual adjustment time for adjusting the supply voltage and/or the bias voltage of the power amplifier (310) .
The method according to claim 1, wherein obtaining (410) traffic load information for the transmitter (310) comprises:

receiving a prewarning signal (333) generated based on the traffic load information for the transmitter (310) from a baseband unit (350) in the RAN (110, 120) ; and wherein

controlling (420) the power amplifier based on the traffic load information comprises adjusting (423) the supply voltage and/or the bias voltage of the power amplifier based on the prewarning signal.
The method according to claim 10, further comprising:

transmitting traffic load capacity of the transmitter (310) to the baseband unit (350) .
The method according to claim 1, wherein obtaining traffic load information for the transmitter comprises:

receiving accumulated traffic load statistics of the RAN (110, 120) over a certain time period from a core network (140) in a communication network (100) ;

generating traffic load patterns (334) for the transmitter (310) from the received accumulated traffic load statistics over a certain time period; and wherein

adjusting the supply voltage and/or bias voltage of the power amplifier (320) based on the traffic load information comprises adjusting (424) the supply voltage and/or bias voltage of the power amplifier (320) based on the traffic load patterns (334) .
The method according to any one of claims 1-12, wherein controlling (420) the power amplifier (310) by adjusting a supply voltage and/or bias voltage of the power amplifier is further based on a performance of the transmitter (310) .
The method according to claim 13, wherein when the supply voltage and/or bias voltage of the power amplifier (310) is to be adjusted based on the traffic load information, the method further comprises:

determining (430) the performance of the transmitter (310) by determining error vector magnitude or operating band unwanted emissions of the transmitter (310) ;

comparing (432) the performance of the transmitter (310) with a performance threshold region comprising a first threshold (Th1) and a second threshold (Th2) ;

adjusting (434) the supply voltage and/or bias voltage in response to a performance comparison result by:

increasing the supply voltage and/or bias voltage if the performance of the transmitter (310) is less than the first threshold (Th1) ;

decreasing the supply voltage and/or bias voltage if the performance of the transmitter (310) is larger than the second threshold (Th2) ;

keeping the supply voltage and/or bias voltage unchanged if the performance of the transmitter (310) is within the performance threshold region.
A method performed in a baseband unit (350) for generating a prewarning signal (333) based on traffic load information of a transmitter (310) , wherein the prewarning signa (333) is used for controlling a power amplifier (320) comprised in the transmitter (310) , the method comprising:

obtaining (910) incoming traffic load information for the transmitter (310) ;

comparing (920) a traffic load to be handled by the transmitter (310) with a current traffic load capacity of the transmitter (310) ;

If the traffic load to be handled by the transmitter (310) is larger than the traffic load capacity of the transmitter (310) ,

generating (931) a prewarning signal to increase a supply voltage and/or bias voltage of the power amplifier (320) ;

If the traffic load to be handled by the transmitter is lower than the traffic load capacity of the transmitter (310) ;

generating (932) a prewarning signal to decrease a supply voltage and/or bias voltage of the power amplifier (320) .
The method according to 15, further comprising generating (940) a prewarning signal to inform the baseband unit (350) to spread the traffic load.
A control system (300) for controlling a power amplifier (320) , wherein the power amplifier (320) is comprised in a transmitter (310) of a radio unit (111, 121) in a radio access network, RAN (110, 120) , the control system (300) is configured to:

obtain traffic load information for the transmitter (310) ; and

control the power amplifier (320) based on the traffic load information by adjusting a supply voltage and/or a bias voltage of the power amplifier (320) .
A control system (300) according to claim 17 is further configured to perform the method according to any claims of 2-14.
A baseband unit (350) for generating a prewarning signal (333) based on traffic load information of a transmitter (310) , wherein the prewarning signa (333) is used for controlling a power amplifier (320) comprised in the transmitter (310) , the baseband unit (350) is configured to:

obtain incoming traffic load information for the transmitter (310) ;

compare a traffic load to be handled by the transmitter (310) with a current traffic load capacity of the transmitter (310) ;

If the traffic load to be handled by the transmitter (310) is larger than the traffic load capacity of the transmitter (310) ,

generate a prewarning signal to increase a supply voltage and/or bias voltage of the power amplifier (320) ;

If the traffic load to be handled by the transmitter is lower than the traffic load capacity of the transmitter (310) ;

generate a prewarning signal to decrease a supply voltage and/or bias voltage of the power amplifier (320) .
A radio unit (111, 121) comprises a control system (300) according to any claims of 17-18.