Method and apparatus for calibration of an antenna array
Patent Information
- Application Number
- EP2023818266
- Authority / Receiving Office
- EP · EP
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2023-10-31
- Publication Date
- 2026-09-09
AI Technical Summary
Existing Mutual Coupling Antenna Calibration (MCAC) methods for large antenna arrays suffer from high phase and gain alignment errors due to frequency selective hardware impairments, near-field environmental effects, and low signal-to-noise ratios, leading to degraded beamforming performance.
The method involves deriving Wiener filter coefficients based on power-delay domain features and signal-to-noise ratios for the antenna array, which are then used to estimate transmitter and receiver parameters for calibration.
This approach reduces the performance impact of interference and hardware degradations, decreases calibration complexity, and achieves faster convergence, resulting in improved antenna array calibration and beamforming performance.
Smart Images

Figure CN2023128282_08052025_PF_FP_ABST
Abstract
Description
METHOD AND APPARATUS FOR CALIBRATION OF AN ANTENNA ARRAY
[0001] TECHINCAL FIELD
[0002] The present disclosure generally relates to calibration of an antenna array. The present disclosure discloses a method, an apparatus, a computer program product, a non-transitory computer readable medium and a computer program.BACKGROUND
[0003] A key attribute for Fifth Generation (5G) radio systems is increased capacity in radio networks. Beamforming is a technology that is used by 5G radio systems to provide the desired increased capacity in an efficient manner. In particular, a 5G radio base station may utilize a large antenna array including tens if not hundreds of antennas, which are also referred to herein as antenna elements. Each antenna element is connected to a radio transceiver path. Applying proper scaling in the transceiver paths enables beamforming by efficient control of spatial coherent additions of desired signals and coherent subtractions of unwanted signals. Such beamforming is used both to enable high antenna gain for a desired User Equipment (UE) as well as to enable parallel communication with several UEs using the same time / frequency resource by using orthogonal spatial communication paths (i.e., by using orthogonal beams) .
[0004] Periodic calibration of antenna gains and phases for all transmitter and receiver antenna paths in an antenna array, is useful to achieve desired beamforming performances in a base station, BS. Commonly used periodic calibration methods are the Mutual Coupling Antenna Calibration, MCAC, methods.
[0005] The aim of Mutual Coupling, MC, based, Antenna Calibration, AC, is to estimate and equalize the gain and phase differences among the transmitter and receiver paths utilizing the existing mutual coupling effects. The patent US 2022149517 A1, with title Efficient Antenna Calibration For Large Antenna Arrays, describes the MCAC functionality in detail.
[0006] MC based AC methods available in the literature produce very high phase and gain alignment errors when a system is subjected to frequency selective hardware, HW, impairments, e.g. HW leakage, near field environmental impairments, e.g. reflection from nearby metal objects, HW aging related degradations along with low signal-to-noise ratio, and interference dominated operating conditions. Poor parameter estimation capabilities of the existing algorithms result in degraded antenna branch alignments, and thereby in degraded beamforming function performance support by the BS system.
[0007] Further, the complexity of existing MC based AC algorithms is high as they require many iterations to reach acceptable performance target levels when estimating transmitter and receiver phase-gain errors. Furthermore, the complexity increases linearly as the supported system bandwidth increases.SUMMARY
[0008] An object of the invention is to facilitate a calibration of an antenna array.
[0009] According to an aspect of the invention, a method for calibration of an antenna array is provided. The method comprises deriving at least one Wiener filter coefficient for the antenna array based on at least one power-delay domain feature related to the antenna array, and at least one signal-to-noise ratio related to the antenna array. The method further comprises estimating, using the at least one Wiener filter coefficient for the antenna array, at least one transmitter parameter and / or at least one receiver parameter. The method further comprises calibrating the antenna array based on the at least one transmitter parameter and / or the at least one receiver parameter.
[0010] A Wiener filter, in the field of signal processing, is a filter that may be used to produce an estimate of a random process. A Wiener filter tries to minimize the mean square error between the estimate of a random process and the random process.
[0011] A Wiener filter coefficient may be a coefficient of an equation representing a Wiener filter or a coefficient of a matrix representing a Wiener filter.
[0012] A power-delay domain feature may be a feature of a power-delay domain profile of a propagation medium. The power-delay domain profile of a propagation medium may represent the average power of a received signal in terms of the delay with respect to the first arrival path in multi-path transmission environment.
[0013] A signal-to-noise ratio is a measure that compares the level of a signal to the level of noise, in particular is defined as the ratio of signal power to noise power. Further, a signal-to-interference-plus-noise ratio is a measure that compares the level of a signal to the level of noise plus interference, in particular is defined as the ratio of signal power to the sum of noise power and interference power.
[0014] A transmitter / receiver parameter may be a transmitter / receiver amplitude parameter configured to be applied to a gain control element of a transmitter / receiver path of the antenna array. The gain control element may be an analog or digital element capable of altering the amplitude of an input signal based on an input and / or target amplitude value. A transmitter / receiver parameter may be a transmitter / receiver phase parameter configured to be applied to a phase control element of a transmitter / receiver path of the antenna array. The phase control element may be an analog or digital element capable of altering the phase of an input signal based on an input and / or target phase value. A transmitter / receiver parameter may be a transmitter / receiver frequency parameter configured to be applied to a frequency control element of a transmitter / receiver path of the antenna array. The frequency control element may be an analogic or digital element capable of altering the frequency of an input signal based on an input and / or target frequency value.
[0015] According to an example, the deriving at least one Wiener filter coefficient for the antenna array may comprise deriving at least one Wiener filter coefficient for a transmitter path of the antenna array based on at least one power-delay domain feature related to the transmitter path of the antenna array, and a signal-to-noise ratio related to the transmitter path of the antenna array.
[0016] An antenna array may include separate transmitter paths of the antenna array. Each of the transmitter paths of the antenna array may include a gain control element, a phase control element and / or a frequency control element to provide gain, phase and frequency calibration, respectively, between the transmitter paths of the antenna array.
[0017] According to an example, the estimating may comprise estimating, using at least one Wiener filter coefficient for a transmitter path of the antenna array, at least one transmitter parameter.
[0018] According to an example, the deriving at least one Wiener filter coefficient for the antenna array may comprise deriving at least one Wiener filter coefficient for a receiver path of the antenna array based on at least one power-delay domain feature related to the receiver path of the antenna array, and a signal-to-noise ratio related to the receiver path of the antenna array.
[0019] An antenna array may include separate receiver paths of the antenna array. Each of the receiver paths of the antenna array may include a gain control element, a phase control element and / or a frequency control element to provide gain, phase and / or frequency calibration, respectively, between the receiver paths of the antenna array.
[0020] According to an example, the estimating may comprise estimating, using at least one Wiener filter coefficient for a receiver path of the antenna array, at least one receiver parameter.
[0021] According to an example, the at least one power-delay domain feature may be based on at least one transmission and / or reception calibration signal, and / or at least one coupling matrix of the antenna array, at least one antenna array end-to-end coupling response measurement, and / or at least one antenna-coupling parameter.
[0022] A transmission or reception calibration signal may be a known signal or a pre-determined signal between a transmitter and a receiver. A transmission or reception calibration signal may also be a period of silence, i.e. absence of transmitted signals, between a transmitter and a receiver.
[0023] A coupling matrix of the antenna array may be defined as a square matrix that describes the coupling between different elements in a system. Elements of the coupling matrix are typically coupling coefficients, which quantify the strength of the coupling between two elements in a system. A coupling matrix is used to characterize mutual coupling between the antenna elements at the port level. Mutual coupling is the electromagnetic interaction between the antenna elements in an array. The coupling matrix helps in understanding how each antenna element interacts with the others antenna elements, and may be beneficial for designing and optimizing antenna arrays.
[0024] An antenna array end-to-end coupling response describes the energy absorbed by one antenna element of the antenna array when another nearby antenna element of the antenna array is operating. The antenna array end-to-end coupling may comprise an effect of a hardware impairment, i.e., leakage or hardware degradation due to aging, of the transmitter path of the antenna array, a hardware impairment of the receiver path of antenna array and / or a mutual coupling effect. Mutual coupling may be defined as the electromagnetic interaction between the antenna elements in an antenna array. Current which develops in each antenna element of an antenna array depends on its own excitation and also on the contributions from adjacent antenna elements. Mutual coupling is inversely proportional to the spacing between the different antenna elements in an antenna array. Mutual coupling is typically undesirable because energy that should, or could, be radiated away is absorbed by nearby antenna array elements. Hence, mutual coupling reduces the antenna array element efficiency and performance of antenna array elements when transmitting or receiving a signal.
[0025] An antenna-coupling parameter may be a parameter of a coupling matrix of the antenna array.
[0026] The effect of mutual coupling can be observed or modeled by varying the space between the antenna elements in the array. Any change in the inter-element spacings changes the mutual impedance between the antenna elements.
[0027] According to an example, the at least one power-delay domain feature may comprise at least one of a hardware power-delay domain feature, at least one of a near-field power-delay domain feature, and / or at least one of a far-field power-delay domain feature.
[0028] A hardware power-delay domain feature may be an average power of the received signal in terms of a delay with respect to a first arrival path in a multi-path transmission created by a transceiver hardware impairments present in a radio transceiver system.
[0029] A near-field power-delay domain feature may be an average power of the received signal in terms of the delay with respect to the first arrival path in multi-path transmission created by the coupling behavior of the transceiver antenna and near-field environment of a field installed radio transceiver system.
[0030] A far-field power-delay domain feature may be an average power of the received signal in terms of the delay with respect to the first arrival path in multi-path transmission created by an external calibration device, i.e. a remote antenna array in a distributed multi-antenna system, or by electromagnetic emission.
[0031] According to an example, the at least one hardware power-delay domain feature and / or the at least one near-field power-delay domain feature may be based on at least one coupling matrix of the antenna array.
[0032] According to an example, the at least one hardware power-delay domain feature and / or the at least one near-field power-delay domain feature and / or the at least one far-field power-delay domain feature may be based on at least one transmission and / or reception calibration signal.
[0033] According to an example, at least one basis function may be assigned to the at least one power-delay domain feature.
[0034] A basis function may be an element of a particular basis for a function space. Every function in the function space can be represented as a linear combination of basis functions, just as every vector in a vector space can be represented as a linear combination of basis vectors. In the field of approximation theory, basis functions can be used in the interpolation, and a mixture of basis functions can provide an interpolating function. In this application, an exponential basis function and a sinc basis function have been discussed, but any other basis function could be used, i.e., a uniform distribution basis function.
[0035] According to an example, at least one exponential basis function may be assigned to the at least one hardware domain feature and / or to the at least one near-field power-delay domain feature.
[0036] According to an example, at least one sinc basis function may be assigned to the at least one far-field power-delay domain feature and / or to the at least one near-field power-delay domain feature.
[0037] According to an example, the at least one Wiener filter coefficient for the antenna array, the transmitter path of the antenna array, and / or the receiver path of the antenna array may be scaled such that the gain of at least one Wiener filter coefficient for the antenna array, the transmitter path of the antenna array, and / or the receiver path of the antenna array is unitary.
[0038] According to an example, the at least one transmitter and / or receiver parameter may be at least one of a transmitter and / or receiver phase parameter, a transmitter and / or receiver amplitude parameter, and / or a transmitter and / or receiver frequency parameter.
[0039] According to an example, the estimating may be based on LS estimation method, an ML estimation method, an EM estimation method, an MMSE estimation method, or an LMMSE estimation method.
[0040] A least square, LS, estimation method, in the fields of statistics and signal processing, is an estimation method which minimize the sum of the squares of the errors.
[0041] A maximum likelihood, ML, estimation method, in the fields of statistics and signal processing, is an estimation method which maximize a likelihood function of the parameters to be estimated.
[0042] An expectation maximization, EM, estimation method, in the fields of statistics and signal processing, is an is an iterative estimation method to find a ML or maximum a posteriori, MAP, estimates of the parameters to be estimated.
[0043] A minimum mean square error, MMSE, estimation method, in statistics and signal processing, is an estimation method which minimizes the mean square error, MSE.
[0044] A linear MMSE, LMMSE, estimation method, in statistics and signal processing, is a linear approximation of the MMSE estimation method.
[0045] According to an example, the estimating may be based on at least one iterative estimation algorithm.
[0046] An iterative estimation algorithm may be a mathematical procedure that uses an initial value to generate a sequence of improving approximate solutions for a class of problems, in which the n-th approximation is derived from the previous ones. A specific implementation with termination criteria for a given iterative method like gradient descent, coordinate-descent, hill climbing, Newton’s method, or quasi-Newton methods like Broyden–Fletcher–Goldfarb–Shanno, BFGS, is an algorithm of the iterative method. An iterative method is called convergent if the corresponding sequence converges for given initial approximations.
[0047] According to an aspect of the invention, an apparatus, for calibration of an antenna array, is presented. The apparatus comprises a processor and a memory. The memory contains instructions executable by the processor whereby the apparatus is operative to derive at least one Wiener filter coefficient for the antenna array based on at least one power-delay domain feature related to the antenna array, and a signal-to-noise ratio related to the antenna array. The apparatus is further operative to estimate, using the at least one Wiener filter coefficient for the antenna array, at least one transmitter parameter and / or at least one receiver parameter. The apparatus is further operative to calibrate the antenna array based on the at least one transmitter parameter and / or the at least one receiver parameter.
[0048] According to an example, the apparatus may be operative to perform any of the presented examples.
[0049] According to an example, the apparatus may comprise the antenna array.
[0050] According to an aspect of the invention, an antenna array having been calibrated according to any presented aspect or example is presented.
[0051] According to an aspect of the invention, a computer program product is presented. The computer program product comprising instructions which, when executed on at least one processor, cause the at least one processor to carry out the method according to any presented aspect or example.
[0052] According to an aspect of the invention, a computer program is presented. The computer program comprising instructions which, when executed on at least one processor, cause the at least one processor to carry out the method according to any presented aspect or example.
[0053] According to an aspect of the invention, a tangible, non-transient computer-readable medium comprising instructions is presented. The tangible, non-transient computer-readable medium comprises instructions that, when executed on at least one processor, cause the at least one processor to carry out the method according to any presented aspect or example.BRIEF DESCRIPTION OF THE DRAWINGS
[0054] The inventive concept is now described, by way of example, with reference to the accompanying drawings, in which:
[0055] Figure 1 is a flow diagram of an example of a mutual coupling antenna calibration, MCAC, algorithm that employs Wiener filters.
[0056] Figure 2 is a block diagram of an example of simplified delay-domain signal analysis and feature extraction.
[0057] Figure 3 is a block diagram of an example of a feature set extraction and classification based on simplified delay-domain signal analysis, DSA.
[0058] Figure 4 is a block diagram of an example of computation of a Wiener filter coefficient.
[0059] Figure 5 is a flow diagram of an example of a method for calibration of an antenna array performed by an apparatus for calibration of an antenna array.
[0060] Figure 6 is a flow diagram of an example of an embodiment of a method for calibration of an antenna array performed by an apparatus for calibration of an antenna array.
[0061] Figure 7 is a block diagram of an apparatus for calibration of an antenna array performing a method for calibration of an antenna array.
[0062] Figure 8 is a block diagram of an embodiment of an apparatus for calibration of an antenna array performing a method for calibration of an antenna array.
[0063] Figure 9 is a block diagram of examples of a computer program product, a non-transitory computer readable medium and a computer program.DETAILED DESCRIPTION
[0064] The inventive concept will now be described more fully hereinafter with reference to the accompanying drawings, in which certain embodiments of the inventive concept are shown. This inventive concept may, however, be embodied in many different forms and should not be construed as limited to the embodiments set forth herein; rather, these embodiments are provided by way of example so that this disclosure will be thorough and complete, and will fully convey the scope of the inventive concept to those skilled in the art. Like numbers refer to like elements throughout the description. Any step or feature illustrated by dashed lines should be regarded as optional.
[0065] Note that the description given herein may refer to a 3GPP cellular communications system and, as such, 3GPP terminology or terminology similar to 3GPP terminology is used. However, the concepts disclosed herein are not limited to a 3GPP system.
[0066] Note, that for the purpose of the description given herein, the terms Wiener filter (s) , Wiener filter (s) coefficient (s) , set (s) of Wiener filters, and set (s) of Wiener filter (s) coefficients are used interchangeably and should not be used to unduly limit the scope of the application.
[0067] Note that, for the purpose of the description given herein, an antenna array may comprise one or more antenna elements, AE. An antenna array may be also referred to as transceiver antenna array to highlight the capabilities to both receive and transmit signals. An antenna array may include separate transmitter branches (also referred to herein as transmitter paths) and separate receiver branches (also referred to herein as receiver paths) for each of the one or more antenna elements. Each of the transmitter and receiver branches may include a gain control element, a phase control element and / or a frequency control element to provide gain, phase and / or frequency calibration between the transmit and the received branches, respectively.
[0068] MCAC methods use natural mutual coupling between antenna elements to perform the calibration of an antenna array. In MCAC methods, a calibration signal is sent from a transmitter path of a transmitter antenna and is received at a receiver path of a receiver antenna. The received calibration signal may be represented as an end-to-end transfer function Y expressed as follows. Y=RST+N
[0069] Where T and R are transfer functions of transmitter and receiver paths, respectively, of the antenna array. S is transmitter and receiver antenna paths mutual coupling and is considered to be known in advance and could be, for example, measured in a lab environment or tested after the installation of the antenna array in a physical site. N is an additive white Gaussian noise.
[0070] In other words, an end-to-end transfer function Y is obtained by transmitting a calibration signal from a transmitter path of an antenna array and by receiving the calibration signal at a receiver path of an antenna array. The end-to-end transfer function is affected by the hardware characteristics of the antenna array and by the channel conditions between the transmitter path of the antenna array and the receiver path of the antenna array.
[0071] Based on the end-to-end transfer function Y measurements, the transmitter path T and the receiver path R transfer functions may be estimated from the above equation. For example, an iterative Maximum Likelihood, ML, estimation technique, and iterative Expectation Maximization, EM, estimation technique or an iterative Least Square, LS, estimation technique, may be used for the estimation.
[0072] This application proposes a calibration method and an apparatus to design Wiener filter coefficients for transmitter and receiver antenna paths and then use the designed Wiener filter coefficients along with existing estimation methods to improve the performance of an antenna array calibration. Thereby the following improvements may be achieved:
[0073] · Reduced performance impact from fast changes in far-filed interference level and near-field environment conditions which impact the quality of received MCAC calibration signal and hence antenna calibration performance.
[0074] · Reduced performance impact from the slow degradations of the hardware characteristics (hardware leakage, oscillator characteristics changes etc. ) over time in the field.
[0075] · Reduced complexity compared to existing calibration methods and faster convergence compared to existing iterative calibration methods. Thus, the disclosed calibration method can be performed by an efficient apparatus that performs the procedure at low cost with minimized storage needs.
[0076] Figure 1 is a flow diagram of an example of a mutual coupling antenna calibration, MCAC, algorithm that employs Wiener filters.
[0077] An example of a mutual coupling antenna calibration algorithm based on an iterative parameter estimation method, may start from an Iterative Parameter Estimation Loop Start phase. In the Iterative Parameter Estimation Loop Start phase, a new instance of a periodic antenna calibration procedure is started.
[0078] In an MCAC process initialization phase, parameters to be estimated may be identified. For example, these parameters may be optimal amplitude, phase and / or frequency for each of antenna elements in an antenna array. Further, in an MCAC process initialization phase, other input parameters may be defined and initialized with a value, for example the initial ranges of each of the parameters to be estimated and the number K of signal delay-domain features to be selected in a Tx / Rx Wiener filter design phase.
[0079] In an MCAC End-to-End Transfer Function Sequence Extraction at Rx phase, a received end-to-end transfer function estimate is sent to a Tx / Rx Wiener filter design phase.
[0080] In the Tx / Rx Wiener filter design phase, Wiener filters for transmitter path (s) and receiver path (s) of the antenna elements of the antenna array are obtained. An example of this phase is illustrated in detail in Figures 2, 3 and 4 and in the description relating thereto. The Tx / Rx Wiener filter design phase receive, as an input, K number of signal delay-domain features to be selected, an end to end transfer function and a Scattering matrix, S-matrix, of the antenna array. The Tx / Rx Wiener filter design phase will output a transmitter Wiener filter and a receiver Wiener filter.
[0081] In an Apply Wiener Filter for Tx parameter estimation phase and an Apply Wiener Filter for Rx parameter estimation phase, the transmitter and receiver Wiener filters are applied to the estimation, that could be done, for example, with an existing LS, EM or ML method. The application of the Tx / Rx Wiener filters to a TX / Rx estimation method may result in a MMSE or LMMSE estimator of the transmitter and receiver parameters.
[0082] In a Tx and Rx Convergence Error Computation phase, an estimation error of the transmitter and receiver parameters is obtained. If the estimation error is below a threshold, then the transmitter and receiver parameters are used to calibrate the antenna array in a Calibrate the Antenna Array phase. If the estimation error is not below the threshold, a new estimation of the transmitter and receiver parameters will be performed in the in the Apply Wiener Filter for Tx parameter estimation phase and the Apply Wiener Filter for Rx parameter estimation phase based on the current estimate of transmitter and receiver parameters. The threshold may be predetermined or be determined during the loop.
[0083] Figure 2 is a block diagram of an example of simplified delay-domain signal analysis and feature extraction.
[0084] A signal in the frequency domain S (f) is transformed, using an inverse Discrete Fourier Transform, DFT, to a signal in the time domain s (n) .
[0085] A delay-domain signal power s (n) 2 is then computed based on the signal in the time domain s (n) .
[0086] A Teager-Kaiser, TK, energy operator is applied on the delay-domain signal power s (n) 2 to reveal any signal delay-domain feature present in the delay-domain signal power s (n) 2.
[0087] The Teager-Kaiser energy operator estimates the energy of a signal. The estimate of the energy of the signal is derived from the instantaneous amplitude and instantaneous frequency of the signal. Thus, the TK operator can be used to detect instantaneous changes in either amplitude or frequency in a given signal.
[0088] A signal delay-domain feature is defined as an identifiable artifact present in power-delay domain signals and is represented through tuple (relative power, delay value) .
[0089] A predefined power threshold is applied to the signal delay-domain features to select K number of signal delay-domain features that may be based on the relative power of each of the power-delay domain features. The selection of the K number of signal delay-domain features allow for the reduction of complexity of the antenna array calibration algorithm. The reduction of complexity may also come in the form of faster convergence of an iterative calibration algorithm. K number of signal delay-domain features is a parameter that can be tuned to optimize the tradeoff between complexity and accuracy of the representation of the signal in the power-delay domain.
[0090] The K number of signal delay-domain features constitute a signal delay-domain feature set Ψ.
[0091] The kth entry ψk of the signal delay-domain feature set Ψ is represented trough the tuple ψk: [ (relative power) k, (delay value) k] .
[0092] The use of simplified delay-domain signal analysis methods is illustrated in the following figure and in the description relating thereof.
[0093] Figure 3 is a block diagram of an example of a feature set extraction and classification based on simplified delay-domain signal analysis, DSA.
[0094] An example of computation of feature set extraction and classification based on simplified delay-domain signal analysis, DSA is exemplified below.
[0095] A Scattering matrix, S-matrix, is processed by a first DSA block.
[0096] An output of the first DSA block is a feature set ΨHW having k number of signal delay-domain features, wherein k≤K. These features represent hardware, HW, delay-domain features which are part of the hardware of the antenna array and of the hardware of a device comprising the antenna array. The S-matrix, also referred to as a mutual coupling matrix, is considered to be known through measurement performed on the hardware of the antenna array or on the hardware of the device comprising the antenna array, estimation or previously obtained knowledge of coupling parameters.
[0097] An end-to-end transfer function Y (f) is processed by a second DSA block. An output of the second DSA block is a feature set ΨHW+OTA having K number of signal delay-domain features. These features represent both hardware delay-domain features and over-the-air, OTA, delay-domain features. The over-the-air delay-domain features may be divided into either near-field delay-domain features and / or far-field delay-domain features.
[0098] The end-to-end transfer function Y (f) is considered to be known to be known through measurement, estimation or previously obtained as part of the antenna calibration process.
[0099] The feature sets ΨHW and ΨHW+OTA are used as inputs for a feature classification process. The feature classification process classifies each feature from the feature sets ΨHW and ΨHW+OTA as either a hardware delay-domain feature, a near-field delay-domain feature or a far-field delay-domain feature. Any feature which is not part of the feature sets ΨHW is identified as an OTA delay-domain feature and may be classified as either a near-field delay-domain feature or as a far-field delay-domain feature.
[0100] As an example, the classification may be performed by comparing each of the signal delay-domain features of the two features sets ΨHW and ΨHW+OTA. All the signal delay-domain features belonging to only the feature set ΨHW may be classified as hardware delay-domain features. All the signal delay-domain features belonging to only the feature set ΨHW+OTA may be classified as either a near-field delay-domain feature or a far-field delay-domain feature. A signal delay-domain feature may be classified as a near-field delay-domain feature if the relative power of the delay-domain feature is above a predetermined power threshold and if the relative delay value of the delay-domain feature is below a predetermined over-the-air delay threshold. The delay-domain feature may be classified as a far-field delay-domain feature if the relative power of the delay-domain feature is above the predetermined power threshold and if the relative delay value of the delay-domain feature is above the predetermined over-the-air delay threshold.
[0101] In the event of overlapping signal delay-domain features between the feature sets ΨHW and ΨHW+OTA, a threshold may be used for the classification.
[0102] As an example, a signal delay-domain feature may be classified as a hardware delay-domain feature if the relative delay value of the delay-domain feature is below a predetermined hardware delay threshold and if the relative power value of the delay-domain feature is above a predetermined power threshold. The hardware delay threshold and / or power threshold may be related to the S-parameter set. The delay-domain feature may be classified as a near-field delay-domain feature if the relative power of the delay-domain feature is above a predetermined power threshold and if the relative delay value of the delay-domain feature is below a predetermined over-the-air delay threshold and above the predetermined hardware delay threshold. The delay-domain feature may be classified as a far-field delay-domain feature if the relative power of the delay-domain feature is above the predetermined power threshold and if the relative delay value of the delay-domain feature is above the predetermined over-the-air delay threshold and above the predetermined hardware delay threshold.
[0103] If the relative power of a delay-domain features is below a predetermined power threshold, the delay-domain feature may be classified as noise and not used in the estimation of the Wiener filters coefficients.
[0104] After the feature classification, each signal delay-domain feature will be either a hardware delay-domain feature, a near-field delay-domain feature or a far-field delay-domain feature and will be referred to as a classified signal delay-domain feature.
[0105] In the basis function association, each classified signal delay-domain feature is associated with a respective basis function.
[0106] As an example, each of the hardware delay-domain features are associated with an exponential function in the frequency domain. Each of the near-field delay-domain features is associated with a sinc function in the frequency domain. Each of the far-field delay-domain features is associated with a sinc function in the frequency domain. Other basis functions may be used to exploit the different properties and characteristic of the different basis functions.
[0107] After performing the feature classification and the basis function association, a feature set Ψin is obtained. The feature set Ψin is a set composed by each of the classified signal delay-domain feature having respective associated basis functions. The feature set Ψin is used for Wiener filter design.
[0108] As discussed above, the DSA method applied to both the S-Matrix and to the end-to-end transfer function Y (f) may facilitate the reduction of computational complexity in the feature extraction and classification process. Further, the DSA method applied to both the S-Matrix and to the end-to-end transfer function Y (f) may facilitate the feature set construction and appropriate basis function association and may facilitate the convergence of an iterative parameter estimation method.
[0109] The K number of signal-delay domain features are identified following the use of the DSA method on both the S-Matrix and on the end-to-end transfer function Y (f) . K1 number of signal delay-domain features are identified as hardware delay-domain feature and K2number of signal delay-domain features are identified as over-the-air delay-domain features, where K1+K2= K. In other words, K number of signal-delay domain features is the total amount of selected signal-delay domain feature. K1 number of signal delay-domain features, wherein the K1 number of signal delay-domain features are identified as hardware delay-domain feature, is a subset of the K number of signal-delay domain features. K2 number of signal delay-domain features, wherein the K2 number of signal delay-domain features are identified as over-the-air delay-domain feature, is a subset of the K number of signal-delay domain features. The requirement is that K1+K2= K.
[0110] Thereby, each entry of a classified feature set Ψin will have following three associated values: { (relative poweri, delay valuei, BasisFunctioni) } i=1, …, K
[0111] Where BasisFunctioni, according to the previous example, may be a sinc function when referring to the near-field and far-field delay-domain features or an exponential function when referring to hardware delay-domain features, and is used to compute a frequency autocorrelation function.
[0112] Figure 4 is a block diagram of an example of computation of a Wiener filter coefficient.
[0113] An example of computation of a Wiener filter coefficient is exemplified below.
[0114] A classified feature set Ψin, an instantaneous SNR value of a transmitter path of an antenna array σTx2 and an instantaneous SNR value of a receiver path of an antenna array σRx2 may be used for the design of a Wiener filters coefficients. Further, a Wiener filter length NWin may defined based on a largest delay value of the classified feature set Ψin.
[0115] A frequency autocorrelation matrix of a transmitter path of an antenna array is represented as
[0116] A frequency autocorrelation matrix of a receiver path of an antenna array is represented as
[0117] and may be computed with the knowledge of the classified feature set Ψin, of the instantaneous SNR value of the transmitter path of the antenna array σTx2 of the instantaneous SNR value of the receiver path of the antenna array σRx2 and of the Wiener filter length NWin using known methods.
[0118] Wiener filter coefficients may be computed using the following equations:
[0119] Wherein WTx is a Wiener filter of the transmitter path of the antenna array and WRx is a Wiener filter of the receiver path of the antenna array.
[0120] WTx and WRx may be used within an iterative multiparameter estimation method (e.g., iterative Maximum Likelihood method or iterative Least Square method) or any other multiparameter estimation method.
[0121] Parameter convergence may be facilitated by scaling the one or more coefficients of WTx and WRx such that overall filtering gain is unitary and thus avoiding non-convergence in iterative multi-parameter estimation. Further, the convergence may be facilitated by applying the scaled WTx and WRx filters independently for the respective parameter estimation assuming that the other parameter is known.
[0122] A factor of iteration number reduction can be achieved compared to existing per frequency-based antenna parameters estimation approaches, thereby facilitating a complexity reduction of the calibration process.
[0123] As mentioned in the above, one instance of MCAC estimation-calibration event is shown in figure 1. The Wiener filter coefficients for the transmitter path of the antenna array and for the receiver path of the antenna array remains the same during the multiparameter estimation iteration. For every new instance of MCAC, Wiener filters are redesigned based on the received MCAC training sequence.
[0124] Figure 5 is a flow diagram of an example of a method 1000 for calibration of an antenna array performed by an apparatus (not shown; see e.g. figures 7 and 8 and the description relating thereto) for calibration of an antenna array.
[0125] According to an example, the method 1000 for calibration of an antenna array comprises deriving 1100 at least one Wiener filter coefficient for the antenna array based on at least one power-delay domain feature related to the antenna array and at least one signal-to-noise ratio related to the antenna array.
[0126] According to an example, the method 1000 for calibration of an antenna array comprises estimating 1200, using the at least one Wiener filter coefficient for the antenna array, at least one transmitting parameter and / or at least one receiver parameter.
[0127] According to an example, the method 1000 for calibration of an antenna array comprises calibrating 1300 the antenna array based on the at least one transmitter parameter and / or the at least one receiver parameter.
[0128] Figure 6 is a flow diagram of an example of an embodiment of a method 1000 for calibration of an antenna array performed by an apparatus (not shown; see e.g. figures 7 and 8 and the description relating thereto) for calibration of an antenna array.
[0129] According to an embodiment, the deriving 1100 at least one Wiener filter coefficient for the antenna array based on at least one power-delay domain feature related to the antenna array and at least one signal-to-noise ratio related to the antenna array, may comprise, alternatively or additionally, deriving 1110 at least one Wiener filter coefficient for a transmitter path of the antenna array based on at least one power-delay domain feature related to the transmitter path of the antenna array, and a signal-to-noise ratio related to the transmitter path of the antenna array.
[0130] According to an embodiment, the deriving 1100 at least one Wiener filter coefficient for the antenna array based on at least one power-delay domain feature related to the antenna array and at least one signal-to-noise ratio related to the antenna array, may comprise, alternatively or additionally, deriving 1120 at least one Wiener filter coefficient for a receiver path of the antenna array based on at least one power-delay domain feature related to the receiver path of the antenna array, and a signal-to-noise ratio related to the receiver path of the antenna array.
[0131] According to an embodiment, the deriving 1100 at least one Wiener filter coefficient for the antenna array based on at least one power-delay domain feature related to the antenna array and at least one signal-to-noise ratio related to the antenna array, may comprise, additionally, that the at least one power-delay domain feature is based on: at least one transmission and / or reception calibration signal, at least one coupling matrix of the antenna array, at least one antenna array end-to-end coupling response measurement, and / or at least one antenna-coupling parameter.
[0132] According to an embodiment, the deriving 1100 at least one Wiener filter coefficient for the antenna array based on at least one power-delay domain feature related to the antenna array and at least one signal-to-noise ratio related to the antenna array, may comprise, additionally, that the at least one power-delay domain feature comprises at least one of a hardware power-delay domain feature, at least one of a near-field power-delay domain feature, and / or at least one of a far-field power-delay domain feature.
[0133] According to an embodiment, the at least one power-delay domain feature comprising at least one of a hardware power-delay domain feature, at least one of a near-field power-delay domain feature, and / or at least one of a far-field power-delay domain feature, may comprise, additionally, that the at least one hardware domain feature and / or the at least one near-field power-delay domain feature is based on at least one coupling matrix of the antenna array.
[0134] According to an embodiment, the at least one power-delay domain feature comprising at least one of a hardware power-delay domain feature, at least one of a near-field power-delay domain feature, and / or at least one of a far-field power-delay domain feature, may comprise, additionally, that the at least one hardware domain feature and / or the at least one near-field power-delay domain feature and / or the at least one far-field power-delay domain feature is based on at least one transmitter path of the antenna array and / or receiver path of the antenna array calibration signal.
[0135] According to an embodiment, the deriving 1100 at least one Wiener filter coefficient for the antenna array based on at least one power-delay domain feature related to the antenna array and at least one signal-to-noise ratio related to the antenna array, may comprise, additionally, that at least one basis function is assigned to the at least one power-delay domain feature.
[0136] According to an embodiment, the at least one basis function being assigned to the at least one power-delay domain feature, may comprise, alternatively or additionally, that at least one exponential basis function is assigned to the at least one hardware domain feature and / or to the at least one near-field power-delay domain feature.
[0137] According to an embodiment, the at least one basis function being assigned to the at least one power-delay domain feature, may comprise, alternatively or additionally, that at least one sinc basis function is assigned to the at least one far-field power-delay domain feature and / or to the at least one near-field power-delay domain feature.
[0138] According to an embodiment, the deriving 1100 at least one Wiener filter coefficient for the antenna array based on at least one power-delay domain feature related to the antenna array and at least one signal-to-noise ratio related to the antenna array, may comprise, additionally, that the at least one Wiener filter coefficient for either, the antenna array, and / or the transmitter path of the antenna array, and / or the receiver path of the antenna array is scaled such that the gain of at least one Wiener filter coefficient for either, the antenna array, and / or the transmitter path of the antenna array, and / or the receiver path of the antenna array is unitary.
[0139] According to an embodiment, the estimating 1200, using the at least one Wiener filter coefficient for the antenna array, at least one transmitting parameter and / or at least one receiver parameter, may comprise, alternatively or additionally, estimating 1210, using at least one Wiener filter coefficient for a transmitter path of the antenna array, at least one transmitter parameter.
[0140] According to an embodiment, the estimating 1200, using the at least one Wiener filter coefficient for the antenna array, at least one transmitting parameter and / or at least one receiver parameter, may comprise, alternatively or additionally, estimating 1220, using at least one Wiener filter coefficient for a receiver path of the antenna array, at least one receiver parameter.
[0141] According to an embodiment, the estimating 1200, using the at least one Wiener filter coefficient for the antenna array, at least one transmitting parameter and / or at least one receiver parameter, may comprise, additionally, being based on an LS estimation method, an ML estimation method, an EM estimation method, an MMSE estimation method, or an LMMSE estimation method.
[0142] According to an embodiment, the estimating 1200, using the at least one Wiener filter coefficient for the antenna array, at least one transmitting parameter and / or at least one receiver parameter, may comprise, additionally, being based on at least one iterative estimation algorithm.
[0143] According to an embodiment, the method 1000 for calibration of an antenna array comprises calibrating 1300 the antenna array based on the at least one transmitter parameter and / or the at least one receiver parameter.
[0144] Figure 7 is a block diagram of an apparatus 2000 for calibration of an antenna array performing the method 1000 for calibration of an antenna array.
[0145] According to an example, the apparatus 2000 comprises a processor 2100 and a memory 2200.
[0146] According to an example, the memory 2200 contains instructions executable by the processor 2100 whereby the apparatus 2000 is operative to derive at least one Wiener filter coefficient for the antenna array based on at least one power-delay domain feature related to the antenna array, and a signal-to-noise ratio related to the antenna array.
[0147] According to an example, the memory 2200 contains instructions executable by the processor 2100 whereby the apparatus 200 is operative to estimate, using the at least one Wiener filter coefficient for the antenna array, at least one transmitter parameter and / or at least one receiver parameter.
[0148] According to an example, the memory 2200 contains instructions executable by the processor 2100 whereby the apparatus 2000 is operative to calibrate the antenna array based on the at least one transmitter parameter and / or the at least one receiver parameter.
[0149] According to an embodiment, the apparatus 2000 is operative to perform any of the embodiments of the method 1000.
[0150] Figure 8 is a block diagram of an embodiment of an apparatus 2000 for calibration of an antenna array performing the method 1000 for calibration of an antenna array.
[0151] The apparatus 2000 for calibration of an antenna array, may comprise, alternatively or additionally, a deriving unit 2300 for deriving 1100 at least one Wiener filter coefficient for the antenna array based on at least one power-delay domain feature related to the antenna array and at least one signal-to-noise ratio related to the antenna array.
[0152] The apparatus 2000 for calibration of an antenna array, may comprise, alternatively or additionally, an estimating unit 2400 for estimating 1200, using the at least one Wiener filter coefficient for the antenna array, at least one transmitting parameter and / or at least one receiver parameter.
[0153] The apparatus 2000 for calibration of an antenna array, may comprise, alternatively or additionally, a calibrating unit 2500 for calibrating 1300 the antenna array based on the at least one transmitter parameter and / or the at least one receiver parameter.
[0154] The apparatus 2000 is preferably a radio access node in a cellular communications network (e.g., a base station in a 3GPP 5G NR network) . However, the apparatus 2000 may alternatively be, for example, an access point in a local wireless network (e.g., an access point in a WiFi network) , a wireless communication device (e.g., a UE in a 3GPP 5G NR network) , a beamforming transceiver, or the like. The apparatus 2000 may perform beamforming via an antenna array 2600. This beamforming may be, e.g., analog beamforming, which is performed by controlling gain and phase for each antenna branch via respective gain and phase control elements. However, it should be appreciated that, in some other embodiments, the apparatus 2000 may perform, e.g., hybrid beamforming, i.e., perform beamforming partly in the digital domain and partly in the analog domain or may perform digital beamforming (i.e., beamforming fully in the digital domain) .
[0155] The apparatus 2000 for calibration of an antenna array 2600, may comprise, alternatively or additionally, a deriving unit 2310 for deriving 1110 at least one Wiener filter coefficient for a transmitter path of the antenna array based on at least one power-delay domain feature related to the transmitter path of the antenna array, and a signal-to-noise ratio related to the transmitter path of the antenna array. Further, the deriving unit 2300 may comprise the deriving unit 2310.
[0156] The apparatus 2000 for calibration of an antenna array, may comprise, alternatively or additionally, a deriving unit 2320 for deriving 1120 at least one Wiener filter coefficient for a receiver path of the antenna array based on at least one power-delay domain feature related to the receiver path of the antenna array, and a signal-to-noise ratio related to the receiver path of the antenna array. Further, the deriving unit 2300 may comprise the deriving unit 2320.
[0157] The apparatus 2000 for calibration of an antenna array, may comprise, alternatively or additionally, an estimating unit 2410 for estimating 1210, using at least one Wiener filter coefficient for a transmitter path of the antenna array, at least one transmitter parameter. Further, the estimating unit 2400 may comprise the estimating unit 2410.
[0158] The apparatus 2000 for calibration of an antenna array, may comprise, alternatively or additionally, an estimating unit 2420 for estimating 1220, using at least one Wiener filter coefficient for a receiver path of the antenna array, at least one receiver parameter. Further, the estimating unit 2400 may comprise the estimating unit 2420.
[0159] According to an embodiment, the apparatus 2000 for calibration of an antenna array 2600, may comprise, additionally, the antenna array 2600.
[0160] The antenna array 2600 may be a Phased Antenna Array Module (PAAM) , an Advanced Antenna System (AAS) or an antenna system.
[0161] The antenna array 2600 may be implemented as one or more radio ASICs, and the processor 2100 is a baseband processor implemented as, e.g., one or more processors such as, e.g., one or more CPUs, one or more baseband ASICs, one or more Field Programmable Gate Arrays (FPGAs) , or the like, or any combination thereof.
[0162] The antenna array 2600 may include many Antenna Elements (AEs) .
[0163] The antenna array 2600 may include separate transmit branches (also referred to herein as transmit paths) and separate receive branches (also referred to herein as receive paths) for each AE.
[0164] Each transmit branch may include a gain control element, a phase control element and / or a frequency control element that is controlled by the processor 2100 to provide gain and phase calibration between the transmit branches.
[0165] The gain control element, the phase control element and / or the frequency control element of the transmit branch that is controlled by the processor 2100 may provide analog beamforming for signals transmitted by the apparatus 2000 for calibration of the antenna array 2600.
[0166] Note that analog calibration and analog beamforming are shown herein as an example; however, the present disclosure is not limited thereto.
[0167] Each receive branch may include a gain control element, a phase control element and / or a frequency control element that is controlled by the processor 2100 to provide gain and phase calibration between the receive branches.
[0168] The gain control element, the phase control element and / or the frequency control element of the receive branch that is controlled by the processor 2100 may provide analog beamforming for signals received by the apparatus 2000.
[0169] The apparatus 2000, may comprise, additionally, a self-calibration unit 2700.
[0170] The self-calibration unit 2700 may be implemented in hardware or a combination of hardware and software. At least some of the functionality of the self-calibration unit 2700 described herein may be implemented in software that is executed by one or more processors (e.g., one or more CPUs, one or more ASICs, one or more FGPAs, or the like, or any combination thereof) .
[0171] The self-calibration unit 2700 may include a controller unit 2710. The controller unit 2710 generally operates to control the self-calibration subsystem 2700 and the antenna array 2600 to perform a calibration procedure as described herein.
[0172] Figure 9 is a block diagram of examples of a computer program product 3100, a non-transitory computer readable medium 3200 and a computer program 3300.
[0173] According to an example, the computer program product 3100 comprises instructions which, when executed on at least one processor, cause the at least one processor to carry out the computer program 3300.
[0174] The computer program product 3100 may comprise the non-transitory computer readable medium 3200 such as, for example, a universal serial bus (USB) memory, a plug-in card, an embedded drive, or a read-only memory.
[0175] According to an example, the non-transitory computer readable medium 3200 may comprises instructions that, when executed on at least one processor, cause the at least one processor to carry out the computer program 3300. The non-transitory computer readable medium 3200 may comprise the computer program 3300. In other words, the computer program 3300 may be stored on the non-transitory computer readable medium 3200.
[0176] According to an example, the computer program 3300 comprises instructions which, when executed on at least one processor 3100, cause the at least one processor to carry out the method 1000.
Claims
1.A method (1000) for calibration of an antenna array, the method comprising:deriving (1100) at least one Wiener filter coefficient for the antenna array based on:at least one power-delay domain feature related to the antenna array, andat least one signal-to-noise ratio related to the antenna array;estimating (1200) , using the at least one Wiener filter coefficient for the antenna array, at least one transmitter parameter and / or at least one receiver parameter;calibrating (1300) the antenna array based on the at least one transmitter parameter and / or the at least one receiver parameter.2.The method of claim 1, wherein the deriving (1100) at least one Wiener filter coefficient for the antenna array comprises deriving (1110) at least one Wiener filter coefficient for a transmitter path of the antenna array based on:at least one power-delay domain feature related to the transmitter path of the antenna array, anda signal-to-noise ratio related to the transmitter path of the antenna array.3.The method of claim 2, wherein the estimating (1200) comprises estimating (1210) , using at least one Wiener filter coefficient for a transmitter path of the antenna array, at least one transmitter parameter.4.The method of any preceding claim, wherein the deriving (1100) at least one Wiener filter coefficient for the antenna array comprises deriving (1120) at least one Wiener filter coefficient for a receiver path of the antenna array based on:at least one power-delay domain feature related to the receiver path of the antenna array, anda signal-to-noise ratio related to the receiver path of the antenna array.5.The method of claim 4, wherein the estimating (1200) comprises estimating (1220) , using at least one Wiener filter coefficient for a receiver path of the antenna array, at least one receiver parameter.6.The method of any preceding claim, wherein the at least one power-delay domain feature is based on:at least one transmission and / or reception calibration signal,at least one coupling matrix of the antenna array,at least one antenna array end-to-end coupling response measurement, and / orat least one antenna-coupling parameter.7.The method of any preceding claim, wherein the at least one power-delay domain feature comprises:at least one of a hardware power-delay domain feature,at least one of a near-field power-delay domain feature, and / orat least one of a far-field power-delay domain feature.8.The method of claim 7, wherein the at least one hardware domain feature and / or the at least one near-field power-delay domain feature is based on at least one coupling matrix of the antenna array.9.The method of any one of claims 7 to 8, wherein the at least one hardware domain feature and / or the at least one near-field power-delay domain feature and / or the at least one far-field power-delay domain feature is based on at least one transmitter path of the antenna array and / or receiver path of the antenna array calibration signal.10.The method of any preceding claim, wherein at least one basis function is assigned to the at least one power-delay domain feature.11.The method of claims 10, wherein at least one exponential basis function is assigned to the at least one hardware domain feature and / or to the at least one near-field power-delay domain feature.12.The method of any one of claims 10 to 11, wherein at least one sinc basis function is assigned to the at least one far-field power-delay domain feature and / or to the at least one near-field power-delay domain feature.13.The method of any preceding claim, wherein the at least one Wiener filter coefficient for either, the antenna array, and / or the transmitter path of the antenna array, and / or the receiver path of the antenna array is scaled such that the gain of at least one Wiener filter coefficient for either, the antenna array, and / or the transmitter path of the antenna array, and / or the receiver path of the antenna array is unitary.14.The method of any preceding claim, wherein the at least one transmitter and / or receiver parameter is at least one of:a transmitter and / or receiver phase parameter,a transmitter and / or receiver amplitude parameter, and / ora transmitter and / or receiver frequency parameter.15.The method of any preceding claim, wherein the estimating is based on an LS estimation method, an ML estimation method, an EM estimation method, an MMSE estimation method, or an LMMSE estimation method.16.The method of any preceding claim, wherein the estimating is based on at least one iterative estimation algorithm.17.An apparatus (2000) , for calibration of an antenna array, comprising a processor (2100) and a memory (2200) , the memory containing instructions executable by the processor whereby the apparatus is operative to:derive at least one Wiener filter coefficient for the antenna array based on:at least one power-delay domain feature related to the antenna array, anda signal-to-noise ratio related to the antenna array;estimate, using the at least one Wiener filter coefficient for the antenna array, at least one transmitter parameter and / or at least one receiver parameter;calibrate the antenna array based on the at least one transmitter parameter and / or the at least one receiver parameter.18.The apparatus of claim 17, wherein the apparatus (2000) is operative to perform the method of any of claims 2 to 16.19.The apparatus of claim 17, wherein the apparatus comprises the antenna array (2600) .20.An antenna array (2600) having been calibrated according to the method of any one of claims 1 to 16, or by the apparatus according to claim 17 or 19.21.A computer program (3300) , comprising instructions which, when executed on at least one processor, cause the at least one processor to carry out the method according to any one of claims 1 to 16.22.A computer program product (3100) , comprising instructions which, when executed on at least one processor, cause the at least one processor to carry out the computer program (3300) according to claim 21.23.A tangible, non-transient computer-readable medium (3200) comprising instructions that, when executed on at least one processor, cause the at least one processor to carry out the computer program (3300) according to claim 21.