Wireless communication device and distortion compensation method

Abstract

To sufficiently compensate for nonlinear distortion generated in a power amplifier provided for each of a plurality of antenna elements.SOLUTION: A wireless communication device includes: a plurality of power amplifiers that are provided corresponding to a plurality of antenna elements and amplify respective signals; a distortion compensation unit that performs distortion compensation of a transmission signal using inverse functions corresponding to nonlinear distortion generated in the plurality of power amplifiers; and a control unit that operates by switching between a first mode in which an individual inverse function for each power amplifier is set in the distortion compensation unit and a coefficient of the individual inverse function is updated and a second mode in which updating of the coefficients of the individual inverse functions is stopped and an integrated inverse function obtained by weighted addition of the individual inverse functions is set in the distortion compensation unit.SELECTED DRAWING: Figure 2

Description

[Technical field]

[0001] The present invention relates to a wireless communication device and a distortion compensation method. [Background technology]

[0002] In recent years, there is an increasing need for a power amplifier that operates with high efficiency in terms of miniaturization of a transmission device, reduction of operating costs, and environmental issues. When a power amplifier is operated with high efficiency, in a region where the input power is relatively large, the input power is not linearly amplified and nonlinear distortion occurs. Therefore, an amplification method that also uses digital predistortion, which adds a distortion with the inverse characteristic of the nonlinear distortion generated in the power amplifier to the transmission signal in advance, is sometimes used. The distortion that is added to the transmission signal in advance in the digital predistortion is called a distortion compensation coefficient, and is read from a lookup table or calculated using a polynomial. The distortion compensation coefficient is then appropriately updated so as to sufficiently compensate for the nonlinear distortion that varies depending on the environment, such as temperature.

[0003] On the other hand, in order to reduce interference with destinations other than the destination of the transmitted signal and improve the capacity of the communication system, beamforming may be performed to form a directional beam. When beamforming is performed, a phase difference is set in the signals transmitted from the multiple antenna elements constituting the array antenna. Each of the multiple antenna elements is provided with, for example, a phase shifter to control the phase of the signal, and the phase-controlled signal is amplified by a power amplifier corresponding to each antenna element.

[0004] In this way, even when a power amplifier is provided for each of the multiple antenna elements, it has been proposed to perform digital predistortion on the transmission signal collectively. That is, after compensating for distortion of the transmission signal using a distortion compensation coefficient, the transmission signal is split and amplified by a power amplifier for each antenna element, and then transmitted from each antenna element. In this case, the distortion compensation coefficient is updated, for example, by the following method.

[0005] The first method combines feedback signals fed back from multiple antenna elements and updates the distortion compensation coefficients so that the error between the combined feedback signal and the transmission signal is reduced. The distortion compensation coefficients are updated using, for example, the Least Means Square (LMS) algorithm.

[0006] The second method involves switching feedback signals fed back from multiple antenna elements over time to calculate the error with the transmitted signal, and updating the distortion compensation coefficients to reduce this error. In this case, the distortion compensation coefficients are updated using, for example, the LMS algorithm. [Prior art documents] [Patent documents]

[0007] [Patent Document 1] JP 2019-154024 A [Patent Document 2] JP 2016-100835 A [Non-patent literature]

[0008] [Non-Patent Document 1] S. Lee et al., "Digital Predistortion for Power Amplifiers in Hybrid MIMO Systems with Antenna Subarrays", 2015 IEEE VTC, May 2015 Summary of the Invention [Problem to be solved by the invention]

[0009] However, the above-mentioned method of updating the distortion compensation coefficients has a problem in that the distortion compensation coefficients are not updated appropriately, and nonlinear distortion occurring in the transmission signal in the power amplifier may not be sufficiently compensated for.

[0010] Specifically, when feedback signals from multiple antenna elements are multiplexed, if the electrical lengths of the feedback paths do not exactly match, the accuracy of updating the distortion compensation coefficients based on the multiplexed feedback signals is low. In other words, if the electrical lengths of the feedback paths are different, phase differences, delay time differences, and gain differences occur in the feedback signals in the feedback paths, and the feedback signals are not properly multiplexed. As a result, the accuracy of the distortion compensation coefficients updated based on the error between the feedback signals and the transmission signals is reduced.

[0011] When feedback signals from multiple antenna elements are switched over time, problems due to phase differences, delay time differences, and gain differences of the feedback signals do not occur, but the feedback signals for which the error with the transmission signal is calculated have different timings for each antenna element. Therefore, the error between the feedback signals and the transmission signal with the same timing for all power amplifiers cannot be obtained, and the calculation accuracy of the distortion compensation coefficient using the LMS algorithm, for example, decreases.

[0012] Therefore, it is conceivable to calculate the error between the feedback signal and the transmission signal using only the feedback signal from any one of the antenna elements, and update the distortion compensation coefficient so as to reduce this error. However, in this case, although the nonlinear distortion in the power amplifier that outputs the feedback signal can be effectively compensated for, the nonlinear distortion in the other power amplifiers is not sufficiently compensated for, making it difficult to sufficiently suppress the adjacent channel leakage ratio (ACLR) in all power amplifiers.

[0013] The disclosed technology has been developed in consideration of these points, and aims to provide a wireless communication device and a distortion compensation method that can sufficiently compensate for nonlinear distortion that occurs in a power amplifier provided for each of a plurality of antenna elements. [Means for solving the problem]

[0014] In one aspect, a wireless communication device disclosed in the present application has a plurality of power amplifiers provided corresponding to a plurality of antenna elements, each of which amplifies a signal; a distortion compensation unit that performs distortion compensation of a transmission signal using an inverse function corresponding to nonlinear distortion generated in the plurality of power amplifiers; and a control unit that operates by switching between a first mode in which an individual inverse function for each power amplifier is set in the distortion compensation unit and coefficients of the individual inverse functions are updated, and a second mode in which updating of the coefficients of the individual inverse functions is stopped and an integrated inverse function obtained by weighting and adding the individual inverse functions is set in the distortion compensation unit. Effect of the Invention

[0015] According to one aspect of the wireless communication device and distortion compensation method disclosed in the present application, it is possible to sufficiently compensate for nonlinear distortion occurring in a power amplifier provided for each of a plurality of antenna elements. [Brief description of the drawings]

[0016] [Figure 1] FIG. 1 is a diagram illustrating an example of a communication system according to an embodiment. [Diagram 2] FIG. 2 is a block diagram showing a configuration of an RU according to an embodiment. [Diagram 3] FIG. 3 is a flow diagram showing the operation of the learning mode. [Figure 4] FIG. 4 is a flow diagram illustrating the operation of the linear combination mode. [Diagram 5] FIG. 5 is a diagram for explaining input and output signals. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0017] Hereinafter, an embodiment of a wireless communication device and a distortion compensation method disclosed in the present application will be described in detail with reference to the drawings. Note that the present invention is not limited to this embodiment.

[0018] Fig. 1 is a diagram illustrating an example of a communication system according to an embodiment. In the communication system illustrated in Fig. 1, a plurality of RUs (Radio Units) 100 are connected to a CU / DU (Central Unit / Distributed Unit) 10, and the RUs 100 and UEs (User Equipment) 20 communicate with each other wirelessly.

[0019] The CU / DU 10 is a device that performs baseband processing on signals, for example, encoding information to generate a transmit baseband signal and transmitting it to the RU 100, and decoding a receive baseband signal received from the RU 100. The CU / DU 10 may be configured separately as a CU and a DU.

[0020] The RU 100 is connected to the CU / DU 10 by wire, performs wireless transmission processing on a transmission baseband signal generated by the CU / DU 10, performs wireless reception processing on a received signal from the UE 20 to generate a received baseband signal, and transmits the received signal to the CU / DU 10. The RU 100 also has multiple antenna elements, and performs beamforming by assigning antenna weights to each of the multiple antenna elements when communicating wirelessly with the UE 20. The RU 100 also performs digital predistortion to compensate for nonlinear distortion occurring in a power amplifier provided for each antenna element. In digital predistortion, inverse functions of nonlinear distortion occurring in each power amplifier are linearly combined, and distortion compensation of the transmission signal is performed by the linearly combined inverse function. The configuration and operation of the RU 100 will be described in detail later.

[0021] The UE 20 is a user terminal device such as a mobile phone or a smartphone, and performs wireless communication with the RU 100.

[0022] Fig. 2 is a block diagram showing a configuration of an RU 100 according to an embodiment. The RU 100 shown in Fig. 2 includes a communication interface unit (hereinafter abbreviated as "communication I / F unit") 110, a processor 120, a memory 130, a D / A (Digital / Analog) conversion unit 140, a beamforming processing unit (hereinafter abbreviated as "BF processing unit") 150, a power amplifier (PA) 160, a PA switching unit 170, and an A / D (Analog / Digital) conversion unit 180. Note that Fig. 2 shows processing units related to the process of transmitting a signal to the UE 20, and omits the illustration of processing units related to the process of receiving a signal from the UE 20.

[0023] The communication I / F unit 110 is an interface that is connected to the CU / DU 10 by wire, and transmits and receives baseband signals to and from the CU / DU 10. Specifically, the communication I / F unit 110 receives a transmission baseband transmitted from the CU / DU 10, and transmits a received baseband signal to the CU / DU 10.

[0024] The processor 120 includes, for example, a central processing unit (CPU), a field programmable gate array (FPGA), or a digital signal processor (DSP), and generally controls the entire RU 100. Specifically, the processor 120 includes an inverse function control unit 121, a coefficient update unit 122, and a distortion compensation unit 123.

[0025] The inverse function control unit 121 controls an inverse function used in distortion compensation by the distortion compensation unit 123. The inverse function is a function that performs an inverse transformation of a nonlinear distortion occurring in the PA 160, and is expressed by a lookup table or a polynomial. The inverse function control unit 121 operates by switching between two modes, a learning mode and a linear combination mode, and sets an inverse function in the distortion compensation unit 123 in each mode. Specifically, in the learning mode, the inverse function control unit 121 sequentially sets individual inverse functions corresponding to each of the PAs 160 in the distortion compensation unit 123. That is, in order to learn individual inverse functions related to each of the PAs 160, the inverse function control unit 121 sets individual inverse functions corresponding to each of the PAs 160 in the distortion compensation unit 123, and causes distortion compensation using the individual inverse functions to be performed.

[0026] The individual inverse function F for the i-th PA160 i -1 If the transmit baseband signal at a discrete time n is x(n), then (x) can be expressed by the following equation (1).

number

[0027] In equation (1), Q is the memory depth and K is the order of nonlinear distortion. i,kq is a coefficient updated in the learning mode. In the learning mode, the inverse function control unit 121 sets an individual inverse function of the above equation (1) for each PA 160 in the distortion compensation unit 123.

[0028] In addition, in the linear combination mode, the inverse function control unit 121 calculates a weight for each PA 160 based on the gain of each PA 160, linearly combines the individual inverse functions for each PA 160 using the calculated weight, and derives a final integrated inverse function for distortion compensation. Then, the inverse function control unit 121 sets the integrated inverse function for distortion compensation in the distortion compensation unit 123, and causes the distortion compensation unit 123 to perform distortion compensation of the transmission baseband signal.

[0029] The weight of the i-th PA160 is k iThen, the integrated inverse function F for distortion compensation of the nonlinear distortion in the M PAs 160 is -1 can be expressed by the following equation (2).

number

[0030] In equation (2), the weight k i is a value based on the gain of the i-th PA 160 and satisfies the following condition:

number

[0031] Also, F i -1 denotes the individual inverse function for the i-th PA 160, and M denotes the number of antenna elements (i.e., the number of PAs 160). Therefore, the integrated inverse function F for distortion compensation is -1 is the individual inverse function F for each PA160 i -1 is obtained by weighted sum of

[0032] The coefficient update unit 122 updates the coefficient of the individual inverse function in the learning mode of the inverse function control unit 121. Specifically, the coefficient update unit 122 updates the coefficient of the individual inverse function by using, for example, an LMS algorithm so that an error between a feedback signal from any one of the PAs 160 and a transmission baseband signal is reduced.

[0033] In the learning mode of the individual inverse function for the i-th PA 160, the coefficient update unit 122 calculates the coefficient a i,kq Update. a i,kq (j+1)=a i,kq (j)+μ i ε i (n)·x * (n) ...(3)

[0034] However, in formula (3), μ i is the step size, and εi (n) indicates the error between the feedback signal from the i-th PA 160 and the transmit baseband signal. Also, x * (n) is the complex conjugate of the transmitted baseband signal x(n).

[0035] In addition, when the inverse function control unit 121 is in the linear combination mode, the coefficient update unit 122 stops updating the coefficients of the individual inverse functions.

[0036] The distortion compensation unit 123 performs distortion compensation of the transmission baseband signal using the inverse function set by the inverse function control unit 121. That is, in the learning mode of the inverse function control unit 121, the distortion compensation unit 123 performs distortion compensation of the transmission baseband signal using an individual inverse function (the above formula (1)) related to any one of the PAs 160. Also, in the linear combination mode of the inverse function control unit 121, the distortion compensation unit 123 performs distortion compensation of the transmission baseband signal using an integrated inverse function (the above formula (2)) for final distortion compensation. Then, the distortion compensation unit 123 outputs the distortion-compensated signal after distortion compensation to the D / A conversion unit 140.

[0037] The memory 130 includes, for example, a random access memory (RAM) or a read only memory (ROM), and stores various information when the processor 120 executes processing.

[0038] The D / A conversion unit 140 performs D / A conversion on the distortion-compensated signal output from the distortion compensation unit 123. The analog distortion-compensated signal obtained by the D / A conversion is up-converted by an up-converter to become a radio frequency transmission signal.

[0039] The BF processor 150 performs beamforming to set a phase difference between the signals of each antenna element by adding antenna weights of each of the antenna elements to the transmission signal, and controls the direction of the directional beam. Then, the BF processor 150 outputs the signal of each antenna element to the PA 160 of each antenna element.

[0040] A PA 160 is provided for each antenna element and amplifies a signal for each antenna element. The signal amplified by the PA 160 is transmitted from each antenna element and fed back to the PA switching unit 170. Furthermore, when the signal is amplified by the PA 160, nonlinear distortion occurs according to the power of the signal. However, this nonlinear distortion is suppressed when the distortion compensation unit 123 performs distortion compensation on the transmission baseband signal.

[0041] The PA switching unit 170 switches and outputs the feedback signal fed back from the PA 160 for each antenna element. Specifically, the PA switching unit 170 outputs the feedback signal from each PA 160 for a predetermined time in the learning mode of the inverse function control unit 121. The feedback signal output from the PA switching unit 170 is down-converted by the down-converter to become a feedback signal of a baseband frequency.

[0042] The A / D conversion unit 180 A / D converts the down-converted feedback signal and outputs the digital feedback signal to the coefficient update unit 122.

[0043] Next, the operation of the RU configured as above will be described separately for the learning mode and the linear combination mode of the inverse function control unit 121. Fig. 3 is a flow diagram showing the operation of the inverse function control unit 121 in the learning mode.

[0044] In the learning mode, the PA switching unit 170 switches the PA 160 that outputs the feedback signal so that the feedback signal is output from any one of the PAs 160 (step S101). Here, switching is performed so that the feedback signal of the first PA 160 is output. As a result, the feedback signal from the first PA 160 is fed back to the processor 120.

[0045] Furthermore, when the PA switching unit 170 switches the PA 160 that is the output source of the feedback signal, the inverse function control unit 121 acquires the gain of the switched PA 160 (step S102). At the same time, the inverse function control unit 121 sets an individual inverse function corresponding to the switched PA 160 in the distortion compensation unit 123 (step S103). That is, the setting is made so that distortion compensation is performed by an individual inverse function that performs an inverse conversion of the nonlinear distortion in the first PA 160. This individual inverse function may be expressed using a lookup table that stores a distortion compensation coefficient in association with the power of the transmission baseband signal, or may be expressed using a polynomial that calculates a distortion compensation signal from the transmission baseband signal.

[0046] In this state, when a transmission baseband signal is received by the communication I / F unit 110, the transmission baseband signal is distortion-compensated by the distortion compensator 123 using an individual inverse function corresponding to the first PA 160. The distortion-compensated signal is then D / A converted and up-converted, amplified by each PA 160, and transmitted from the antenna element. The signals amplified by each PA 160 are fed back to the PA switching unit 170, and the signal amplified by the first PA 160 is output from the PA switching unit 170.

[0047] The feedback signal from the first PA 160 is down-converted and A / D converted, and input to the coefficient update unit 122. Then, the coefficient of the individual inverse function corresponding to the first PA 160 is updated by the coefficient update unit 122 (step S104). Specifically, for example, the LMS algorithm shown in the above formula (3) is used to update the error ε 1 (n) to reduce the current coefficient a 1,kq (j) New coefficient a 1,kq(j+1) is calculated. The calculated new coefficient is notified to the inverse function control unit 121, and the inverse function control unit 121 updates the coefficient of the individual inverse function set in the distortion compensation unit 123. As a result, the distortion compensation unit 123 performs distortion compensation using the individual inverse function of the new coefficient.

[0048] When the coefficient of the individual inverse function is updated in this manner, the inverse function control unit 121 measures the time since the PA 160 that outputs the feedback signal is switched, and determines whether or not a predetermined time has elapsed since the PA 160 was switched (step S105). If the result of this determination is that the predetermined time has not elapsed (step S105 No), the coefficient of the individual inverse function related to the first PA 160 is continuously updated (step S104). On the other hand, if the predetermined time has elapsed since the PA 160 was switched (step S105 Yes), it is determined that the coefficient of the individual inverse function related to this PA 160 has been updated for a sufficient time and converged. Then, the inverse function control unit 121 stores the individual inverse function whose coefficient has converged in the memory 130 in association with the identification information of the PA 160. At this time, information related to the gain of the PA 160 is also stored in the memory 130.

[0049] Furthermore, when the coefficient of the individual inverse function for the current PA 160 has been sufficiently updated, it is determined whether or not the update of the coefficient of the individual inverse function has been completed for all PAs 160 (step S106). If the result of this determination is that there is still a PA 160 for which the update of the coefficient of the individual inverse function has not been completed (step S106 No), the PA switching unit 170 switches the output source of the feedback signal to the next PA 160 (step S101). Then, in the same manner as above, the coefficient of the individual inverse function for the switched PA 160 is updated.

[0050] On the other hand, when the update of the coefficients of the individual inverse functions for all PAs 160 is completed (Yes in step S106), the inverse function control unit 121 determines that the learning mode is completed. That is, the coefficients of the individual inverse functions for all PAs 160 are sufficiently updated, and the individual inverse functions and gains corresponding to each PA 160 are stored in the memory 130, completing the learning mode. Then, the inverse function control unit 121 transitions to the linear combination mode.

[0051] FIG. 4 is a flow diagram showing the operation of the inverse function control unit 121 in the linear combination mode.

[0052] In the linear combination mode, the inverse function control unit 121 reads out the gain of each PA 160 from the memory 130 and calculates the weight of the individual inverse function for each PA 160 (step S201). i Then, the weight k of the individual inverse function for PA160 i is calculated, for example, by the following formula (4).

number

[0053] That is, weights are calculated according to the ratio of gains of M PAs 160. The weights of the individual inverse functions do not necessarily have to be calculated according to the ratio of gains, and may be calculated according to the ratio of values ​​of powers of gains, for example. Regardless of how the weights of the PAs 160 are calculated, the sum of the weights of the individual inverse functions for the PAs 160 is 1, as described above.

[0054] When the weights of the individual inverse functions for each PA 160 are calculated, the inverse function control unit 121 linearly combines the individual inverse functions for each PA 160 (step S202). Specifically, as shown in the above formula (2), the individual inverse functions for each PA 160 are weighted and added by the weights, so that the final integrated inverse function for distortion compensation is derived. The integrated inverse function is set in the distortion compensation unit 123, and the subsequent transmission baseband signal is distortion compensated by the integrated inverse function (step S203). That is, the distortion compensation coefficients read from the lookup table for each PA 160 are linearly combined and applied to the transmission baseband signal, or the distortion compensation signal is calculated from the transmission baseband signal by a polynomial obtained by linearly combining the polynomials for each PA 160.

[0055] The distortion-compensated signal obtained by the distortion compensation is D / A converted and up-converted, and the BF processing unit 150 executes beamforming to assign an antenna weight to the signal of each antenna element (step S204). Then, the signal of each antenna element is amplified by the PA 160 for each antenna element (step S205) and then transmitted from each antenna element (step S206). During amplification in the PA 160, nonlinear distortion occurs in the signal of each antenna element, but since the distortion compensation unit 123 executes distortion compensation of the transmission baseband signal, the nonlinear distortion of the radio signal radiated from the antenna element is suppressed, and the ACLR is sufficiently suppressed. In addition, in the distortion compensation, an integrated inverse function for distortion compensation obtained by linearly combining the individual inverse functions for each PA 160 is used, so that the nonlinear distortion in the output of each PA 160 can be uniformly suppressed, and the ACLR of the entire PA 160 can be reduced to the maximum.

[0056] Next, a specific example of updating the coefficients in the learning mode will be described. Fig. 5 is a diagram showing signals input to and output from each processing block in the learning mode. Here, an example in which the number of antenna elements (the number of PAs 160) is two will be described, but the same applies when the number of antenna elements (the number of PAs 160) is three or more.

[0057] As shown in Fig. 5, when a transmission baseband signal x(n) at a discrete time n is input to the distortion compensation unit 123, the transmission baseband signal x(n) is distortion compensated and a distortion-compensated signal z(n) is output. The distortion-compensated signal z(n) is branched into signals for two antenna elements, amplified by the PA 160 provided for each antenna element, and fed back to the PA switching unit 170. That is, the first PA 160 outputs a feedback signal y 1 (n) is fed back to the PA switching unit 170, and the feedback signal y 2 (n) is fed back to the PA switching unit 170.

[0058] In the learning mode, first, the inverse function control unit 121 calculates an individual inverse function F 1 -1 (x) is set in the distortion compensation unit 123. In addition, the feedback signal y 1 The coefficient update unit 122 is switched so that (n) is output to the coefficient update unit 122.

[0059] In the individual inverse function learning mode for the first PA160, the distortion compensation signal z 1 (n) is expressed by the following equation (5). z 1 (n)=F 1 -1 (x(n)) (5)

[0060] Equation (5) is equivalent to the above equation (1) where i is 1. This distortion compensation signal z 1 (n) is given a nonlinear distortion when amplified by each PA 160, and is fed back to the PA switching unit 170 as a feedback signal y 1 (n), y 2 (n). Here, the feedback signal y 1 (n) is output to the coefficient update unit 122.

[0061] In the coefficient update unit 122, the transmission baseband signal x(n) and the feedback signal y 1 Error ε of (n) 1 (n) is calculated. ε 1 (n)=x(n)-y 1 (n) ...(6)

[0062] Then, the individual inverse function F 1 -1 Coefficient a of (x) 1,kq will be updated. a 1,kq (j+1)=a 1,kq (j)+μ 1 ε 1 (n)·x * (n) ...(7)

[0063] Coefficient a 1,kq By updating, the individual inverse function F 1 -1 Since (x) is also updated, the inverse function control unit 121 updates the individual inverse function F 1 -1 (x) is set in the distortion compensation unit 123. Therefore, the transmission baseband signal x(n) is converted into the updated individual inverse function F 1 -1 The distortion is compensated for by (x).

[0064] The individual inverse function F for the first PA160 as described above 1 -1 Coefficient a of (x) 1,kq After the update is repeated for a predetermined time, the inverse function control unit 121 then updates the individual inverse function F 2 -1 (x) is set in the distortion compensation unit 123. In addition, the feedback signal y 2 The coefficient update unit 122 is switched so that (n) is output to the coefficient update unit 122.

[0065] In the learning mode of the individual inverse function for the second PA160, the distortion compensation signal z2 (n) is expressed by the following equation (8). z 2 (n)=F 2 -1 (x(n)) (8)

[0066] Equation (8) is equivalent to the above equation (1) where i is 2. This distortion compensation signal z 2 (n) is given a nonlinear distortion when amplified by each PA 160, and is fed back to the PA switching unit 170 as a feedback signal y 1 (n), y 2 (n). And here, the feedback signal y 2 (n) is output to the coefficient update unit 122.

[0067] In the coefficient update unit 122, the transmission baseband signal x(n) and the feedback signal y 2 Error ε of (n) 2 (n) is calculated. ε 2 (n)=x(n)-y 2 (n) ...(9)

[0068] Then, the individual inverse function F 2 -1 Coefficient a of (x) 2,kq will be updated. a 2,kq (j+1)=a 2,kq (j)+μ 2 ε 2 (n)·x * (n) ...(10)

[0069] Coefficient a 2,kq By updating, the individual inverse function F 2 -1 Since (x) is also updated, the inverse function control unit 121 updates the individual inverse function F 2 -1 (x) is set in the distortion compensation unit 123. Therefore, the transmission baseband signal x(n) is converted into the updated individual inverse function F 2-1 The distortion is compensated for by (x).

[0070] The individual inverse function F for the second PA160 as described above 2 -1 Coefficient a of (x) 2,kq When the update is repeated for a predetermined time, the update of the coefficients of the individual inverse functions for all the PAs 160 is completed, and the learning mode ends.

[0071] In this way, in the learning mode, the PA 160 that is the output source of the feedback signal is fixed, and the coefficients of the individual inverse functions related to this PA 160 are updated. Therefore, problems due to phase differences, delay time differences, and gain differences of the feedback signals for each antenna element do not occur, and the individual inverse functions of each PA 160 can be calculated with high accuracy.

[0072] In the linear combination mode, an integrated inverse function is derived by weighting and adding the individual inverse functions for each PA 160. This prevents biased distortion compensation from being performed for any of the PAs 160, and makes it possible to uniformly suppress nonlinear distortion occurring in each PA 160.

[0073] As described above, according to this embodiment, the coefficients of the individual inverse functions for each PA are updated in the learning mode, and the update of the coefficients is stopped in the linear combination mode, while the individual inverse functions for each PA are linearly combined to derive an integrated inverse function for distortion compensation, and distortion compensation of the transmission signal is performed using the integrated inverse function. Therefore, the nonlinear distortion occurring in each of the multiple power amplifiers can be uniformly suppressed, and the ACLR in all the power amplifiers can be sufficiently suppressed. In other words, the nonlinear distortion occurring in the power amplifiers provided for each of the multiple antenna elements can be sufficiently compensated.

[0074] In the above embodiment, the case where the RU 100 performs analog beamforming has been described, but the distortion compensation according to the above embodiment can also be applied to the case where the RU 100 performs hybrid beamforming. That is, when digital predistortion is performed collectively on each of the subarrays having a plurality of antenna elements, the same distortion compensation as in the above embodiment may be performed for each subarray.

[0075] In the above embodiment, the mode may be changed to the learning mode again after a predetermined time has elapsed after the linear combination mode has been switched to, thereby making it possible to periodically update the coefficients of the individual inverse functions so as to compensate for nonlinear distortion that varies depending on the environment, such as temperature. [Explanation of symbols]

[0076] 110 Communication I / F section 120 processors 121 Inverse function control section 122 Coefficient update unit 123 Distortion compensation section 130 Memory 140 D / A conversion section 150 BF Processing Section 160 Power Amplifier 170 PA Switching Unit 180 A / D conversion section

Claims

1. a plurality of power amplifiers provided corresponding to the plurality of antenna elements, each of which amplifies a signal; a distortion compensation unit that performs distortion compensation on a transmission signal using an inverse function corresponding to nonlinear distortion generated in the plurality of power amplifiers; a control unit that operates by switching between a first mode in which an individual inverse function for each power amplifier is set in the distortion compensation unit and a coefficient of the individual inverse function is updated, and a second mode in which updating of the coefficient of the individual inverse function is stopped and an integrated inverse function obtained by weighting and adding the individual inverse functions is set in the distortion compensation unit; A wireless communication device comprising:

2. a feedback section that feeds back to the control section a signal output from any one of the plurality of power amplifiers when the control section operates in the first mode; 2. The wireless communication device according to claim 1, further comprising:

3. The control unit is In the first mode, a coefficient of an individual inverse function is updated for each of the plurality of power amplifiers for a predetermined time.

2. The wireless communication device according to claim 1.

4. The control unit is In the first mode, when updating of coefficients of individual inverse functions for all of the plurality of power amplifiers is completed, the mode is shifted to the second mode.

2. The wireless communication device according to claim 1.

5. The control unit is In the second mode, a weight for each of the plurality of power amplifiers is calculated according to a gain of each of the plurality of power amplifiers, and the integrated inverse function is derived by performing weighted addition of the individual inverse functions of the respective power amplifiers using the calculated weight.

2. The wireless communication device according to claim 1.

6. The control unit is Calculate the weight of each power amplifier so that the sum of the weights of the multiple power amplifiers is 1.

6. The wireless communication device according to claim 5.

7. A distortion compensation method performed by a wireless communication device having a plurality of power amplifiers provided corresponding to a plurality of antenna elements, each of which amplifies a signal, and a distortion compensation unit that performs distortion compensation on a transmission signal using an inverse function corresponding to nonlinear distortion generated in the plurality of power amplifiers, the method comprising: an individual inverse function for each power amplifier is set in the distortion compensator, and the distortion compensator operates in a first mode in which the coefficient of the individual inverse function is updated; The second mode is operated in which updating of the coefficients of the individual inverse functions is stopped, and an integrated inverse function obtained by weighting and adding the individual inverse functions is set to the distortion compensation unit.

23. A distortion compensation method comprising the steps of:

Images