Novel channel estimation method for enhancing beam domain resolution

Abstract

The invention discloses a novel channel estimation method with enhanced beam domain resolution, and belongs to the technical field of channel estimation in the communication field. The method comprises the following implementation steps: step 1, carrying out beam domain oversampling; step 2, estimating a beam domain channel based on spatial resolution enhancement; step 3, carrying out beam domain oversampling reconstruction; according to the invention, the precision and performance of channel estimation can be effectively improved; and the channel clustering effect can be enhanced, so that the performance of various processing technologies based on channel clustering is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of channel estimation in the communication field, and in particular to a novel channel estimation method with enhanced beam domain resolution. Background Art

[0002] The effective operation of wireless communication systems, such as signal detection, precoding, and wireless resource allocation, all rely on the acquisition of channel information. Therefore, the acquisition of channel state information (CSI) is an important part of the 3GPP standard protocol. From the 4G LTE standard (R8) released in 2009 to the 5G NR standard (R15) first released in 2017, 3GPP has been continuously improving and enhancing the ability to obtain channel information. The quality of channel estimation and channel information acquisition directly determines the overall performance of the wireless communication system and is the key to the success or failure of the wireless communication system.

[0003] Large-scale antenna array technology can significantly improve data throughput and spectrum efficiency by equipping more antennas, so it is widely used in various wireless communication systems.

[0004] For a multiple input multiple output MIMO (Multiple Input and Multiple Output) wireless communication system equipped with multiple antenna units, assuming that the number of antenna units equipped on the base station side is M, the number of antenna units equipped on the user equipment side is N, and the number of transmission streams is S (S≤N), then the uplink channel matrix can be expressed as H∈C MXS , where C represents the complex domain. In the training phase, the pre-designed reference signal x t ∈C SX1 , sent by the user at time t and received by the base station t Expressed as

[0005] y t =H t x t +n (1)

[0006] where n∈C MX1 represents noise. Since the mobile communication device at the user end is powered by a battery, the power consumption and computing power it can provide are relatively limited, so in practice N = S = 1. Because the reference signal x t It is known that the least squares LS (Least Squares) channel estimation can be expressed as

[0007]

[0008] Among them, H LS is the original channel signal Ht Least squares estimation. The reference signal x t Generally satisfies orthogonality, and its conjugate transpose is also often used to replace it. Since the least squares channel estimation is simple to implement, it has been widely used in the industrial community. However, its disadvantages are also obvious. Since the noise term is not considered, when the channel condition is poor, the noise will be greatly amplified, resulting in a significant decline in the channel estimation performance. Therefore, the linear minimum mean square error (LMMSE) channel estimation is often used to improve the least squares (LS) channel estimation;

[0009]

[0010] where R HH = E{HH H} is the autocorrelation matrix of the channel signal, β is a constant determined by the transmitted signal constellation points (for example, for 16-QAM, β = 17 / 9), and SNR represents the signal-to-noise ratio. The LMMSE-based method can improve the channel estimation performance, but its implementation complexity is very high. Especially, it is assumed that the channel autocorrelation matrix R HH is known, which does not hold in practice. Generally, the R HH value is approximately obtained by long-term collection and statistics of the channel after LS estimation. This makes it very difficult and inefficient to obtain the channel statistical information R HH required by LMMSE in practical applications. To reduce the computational and implementation complexity of LMMSE channel estimation, researchers often use the time-domain sparsity of wireless signals to approximately implement LMMSE channel estimation. For example, in the time domain, only the first L (1 < L < N) elements of the channel time-domain impulse response vector g (the channel response is a vector when the user has a single antenna) are considered. In its cross-correlation matrix R gg only the first L elements are retained, and the other elements are set to zero, thereby reducing the computational complexity. The time-domain sparsity of the signal is also reflected in the eigenvalue space. In [4], the autocorrelation function R hh of the channel frequency-domain impulse response vector of a single-antenna user is subjected to singular value decomposition (SVD) to obtain R hh = UΛU H , and only the eigenvalue vector with rank p is retained, so as to obtain the low-rank channel estimation as

[0011] h LowRank = UΔ p U H h LS (4)

[0012] where Δ p is a diagonal matrix, and the first p diagonal elements δ k = λk / (λ k +β / SNR)(λ k For R hh The eigenvalue of , and the elements at other positions are zero. Low-rank channel estimation can approach the theoretical optimal performance, but it still requires a high-complexity singular value decomposition operation. On the other hand, LMMSE and low-rank channel estimation are both based on channel statistics R hh In practice, the channel statistics R hh It is very difficult and inefficient to obtain R, and it is necessary to use long-term channel statistics to obtain sufficiently accurate R hh The approximation makes the algorithm implementation complexity and cost of LMMSE and its low-rank channel estimation type still very high.

[0013] With the help of further exploration of signal sparsity, orthogonal matching pursuit OMP (Orthogonal Matching Pursuit) and basis pursuit BP (Basis Pursuit) get rid of the dependence on channel statistical information. As a result, the traditional channel estimation problem is transformed into a sparse recovery and reconstruction problem. Traditional time-domain channel estimation only collects the energy delay characteristics PDP (Power Delay Profile) of the channel, and estimates the delay corresponding to the multipath peak point from the obtained PDP, and then reconstructs the channel according to the position corresponding to the main energy delay. With the increase in the number of base station antennas, array signal processing comes into play. Taking a linear antenna array ULA (Uniform Linear Array) equipped with M antenna units as an example, the channel model between it and a single-antenna user can be expressed as

[0014]

[0015] Where a0u(θ0) is the line-of-sight (LOS) component, a0 is the path gain, the component corresponding to 1≤p≤P is the non-line-of-sight (NLOS) path component, P is the total number of multipaths, and θ p is the angle of arrival AoA (Angle of Arrive) of path p, u(θ p ) is usually called the steering vector of the antenna array, and its elements can be expressed as m∈[q-(M-1) / 2,q=0,1,…,M-1],λ is the wavelength, d is the antenna spacing, and d=λ / 2 is usually satisfied. In a typical 3GPP environment, the angular spread is relatively limited, generally between 2°-5°. Similar to the sparsity in the time domain, the wireless signal also exhibits spatial sparsity. As a result, the channel estimation problem is transformed into the spatial orientation estimation of the main path. The estimation of the arrival angle is a classic problem in array signal processing, such as the MUSIC and ESPRIT algorithms. These classic algorithms for array signal processing were originally designed for radar perception. They are based on the assumption of blind estimation and do not have pre-designed training sequences like commercial wireless communication systems. Moreover, these algorithms are highly complex and difficult to apply to commercial wireless communication systems. The existing beam-domain low-rank channel estimation samples the spatial domain of the channel signal through the fast Fourier transform FFT algorithm. After the channel signal is transformed into the beam domain through the FFT fast Fourier transform, beam screening is performed. Generally, the beams are arranged from large to small according to energy, and the beams whose sum of beam energy is greater than a certain threshold, such as 85%, are selected. Next, the other beams are set to zero, and the channel is reconstructed in reverse through the selected beam points. These beam-domain low-rank channel estimates use the spatial sparsity of wireless signals to reduce the dimension of the originally large channel, reducing the demand for RF resources in large-scale antenna array systems, and effectively alleviating the data pressure of the fronthaul between the edge antenna RF end and the digital end.

[0016] like Figure 1 As shown in the figure, a typical antenna array A consists of a single dual-polarized antenna B. The antenna array can be divided into multiple sub-arrays C, where each sub-array D is connected to two wireless RF links, one for each polarization direction. For 3GPP base stations, the commonly used antenna arrays are configured in the vertical, horizontal, and polarization dimensions of 2×8×2, 4×8×2, and 8×8×2. It can be seen that the number of antennas in these commonly used antenna arrays in the horizontal or vertical single dimension is not large (2, 4 or 8), which makes its spatial resolution in a single dimension not high, resulting in the problem of missing the main direction of the wave, resulting in the error in the selection of the main energy beam point position of the beam domain channel estimation, greatly reducing the accuracy of the channel estimation. Increasing the number of antennas can certainly solve this problem, but its cost is too expensive and often has no practical application value. Summary of the invention

[0017] To achieve the above objectives, the present invention is implemented through the following technical solutions:

[0018] A new channel estimation method with beam domain resolution enhancement is divided into the following steps;

[0019] Step 1, beam domain oversampling: For the antenna dimension that needs spatial resolution enhancement, perform Fourier transform based on oversampling to obtain the beam domain channel H with enhanced spatial resolutionEhdBeam ; Specific oversampling coefficient ρ os , depending on the antenna configuration and system design requirements;

[0020] Step 2: Beam-domain channel estimation based on spatial resolution enhancement: The beam-domain channel H after spatial resolution enhancement obtained in step 1 is EhdBeam , perform corresponding beam-domain channel estimation processing;

[0021] Step 3, beam domain oversampling reconstruction: retain the beam position selected in step 2, set the beams at other positions to zero, and perform inverse Fourier transform on the screened beam domain channel to obtain the time-frequency domain channel signal estimation result with enhanced spatial resolution.

[0022] Furthermore, the beam domain transform based on Fourier oversampling in step 1 can act on any one or a combination of the horizontal and vertical dimensions of the antenna array corresponding to the received channel signal, thereby enhancing the spatial resolution of the antenna array in the corresponding dimension.

[0023] Furthermore, in step 2, when performing subsequent beam domain channel estimation on the obtained spatial resolution enhanced channel signal, the following methods are included but not limited to: channel estimation based on threshold screening, channel estimation based on noise energy screening, and channel estimation based on beam domain peak iterative elimination.

[0024] Furthermore, for the oversampling Fourier transform and inverse transform performed in step 1 and step 3, the transformed channel energy is scaled and restored according to the oversampling coefficient and the number of days, respectively.

[0025] Furthermore, the beam domain signal is clustered; the parameters referenced by clustering include but are not limited to: beam delay, beam arrival angle, beam energy; the specific parameters used for clustering may be one or more combinations of the listed reference parameters.

[0026] The beneficial effects of the present invention are as follows:

[0027] The present invention proposes a novel channel estimation method with enhanced beam domain resolution, which can effectively improve the accuracy and performance of channel estimation, and can also enhance the effect of channel clustering, thereby improving the performance of various channel clustering-based processing technologies.

[0028] According to 3GPP TR 38.901, a channel model is established for simulation. The antenna array is a uniform linear array ULA (Unitary Linear Array). A new channel estimation method with beam domain resolution enhancement proposed in this application has achieved significant performance gains. Among them, BmSelThrd represents threshold-based beam domain channel estimation, that is, the minimum number of beams required is selected that is greater than the total energy of the beam domain channel, such as 85%; BmAbvNs represents noise energy-based beam domain channel estimation, that is, all beams with beam energies greater than noise energy are selected; Os1 represents a spatial resolution enhancement factor of 2; Os2 represents a spatial resolution enhancement factor of 4; SNR is the signal-to-noise ratio; MSE is the mean square error between the obtained channel estimate and the actual channel. Even if the number of antennas is increased to 128, the technology and method can still achieve excellent performance gains. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 This is an example of an antenna array model;

[0030] Figure 2 It is the 3GPP UMi channel model, 4-antenna ULA, carrier frequency 2GHz, user speed 30KM / H;

[0031] Figure 3 It is the 3GPP UMi channel model, 64-antenna ULA, carrier frequency 5GHz, user speed 30KM / H;

[0032] Figure 4 It is the 3GPP UMi channel model, 128-antenna ULA, carrier frequency 10GHz, and user speed 30KM / H. DETAILED DESCRIPTION

[0033] The preferred embodiments of the present invention are described in detail below in conjunction with the accompanying drawings so that the advantages and features of the present invention can be more easily understood by those skilled in the art, thereby making a clearer and more definite definition of the protection scope of the present invention.

[0034] In response to the problem of insufficient spatial resolution of the above-mentioned beam-domain low-rank channel estimation in a single dimension, the present application introduces an oversampling method to increase the FFT sampling rate by 2 or 4 times, and the spatial resolution in a single dimension can be equivalently increased by 2 or 4 times. It is worth noting that oversampling will destroy the orthogonality of the original FFT sampling, resulting in spectrum leakage. In order to overcome this problem, the present application reconstructs the multipath represented by the peak point obtained by each oversampling, and then eliminates it from the remaining channel matrix or tensor, so that the next beam peak point can be accurately found, until the main beam peak points are found, the channel can be accurately reconstructed. Compared with the practice of increasing the number of antennas, the present application achieves higher-precision channel estimation and reconstruction at a lower cost.

[0035] The implementation steps of the present invention are as follows:

[0036] Step 1, beam domain oversampling;

[0037] For the antenna dimensions that require spatial resolution enhancement, vertical and / or horizontal, an oversampling-based Fourier transform is performed to obtain a beam domain channel H with enhanced spatial resolution. EhdBeam ; Specific oversampling coefficient ρ os , which needs to be determined according to the antenna configuration and system design requirements; usually, the beam domain oversampling factor can be 2 or 4;

[0038] Step 2: beam-domain channel estimation based on spatial resolution enhancement;

[0039] The traditional beam-domain channel estimation is the channel signal H obtained after the initial estimation of the least mean square LS channel estimation method. LS , by using Fourier transform, it is transformed into the beam domain to obtain the beam domain channel signal H Beam , then H Beam Further beam domain channel estimation processing is performed; Different from the traditional beam domain channel estimation, the present invention is the beam domain channel H after the spatial resolution enhancement obtained in step 1 EhdBeam , and perform corresponding beam-domain channel estimation processing.

[0040] As an embodiment, the beam domain channel estimation may be based on a threshold screening method. That is, according to the threshold set by the system design, such as 85% of the total energy, the beam domain channel H after the spatial resolution enhancement is selected. EhdBeam Perform beam screening. Screen out the minimum number of beams that meet the defined threshold, which are the main beam directions, and then retain these beam directions and set the beams in other directions to zero. The designed threshold can be fixed, or it can be set in sections or adaptively adjusted according to the signal-to-noise ratio of the channel signal.

[0041] As an embodiment, the beam domain channel estimation may be based on a noise screening method, that is, the beam screening threshold set by the system is fixed to the noise energy, beams with energy greater than the noise energy are screened out, and other beams are set to zero.

[0042] As an embodiment, the beam domain channel estimation may be an iterative elimination screening method. That is, the system first selects the spatial resolution enhanced beam domain channel H EhdBeamThe first energy peak point is then reconstructed, and the spatial impulse response of the peak point in the direction of the wave is subtracted from the beam domain channel signal of multiple aliased wave directions, and then the second peak point is selected from the beam domain channel signal after subtracting the influence of the first peak point. And so on. Until the number of peak points required by the system is selected, for example, the total energy of the selected peak point is greater than the initial spatial resolution enhancement beam domain channel H EhdBeam 85% of total energy.

[0043] Step 3, beam domain oversampling reconstruction;

[0044] Keep the beam position selected in step 2, set the beams at other positions to zero, and perform inverse Fourier transform on the screened beam domain channel to obtain the time-frequency domain channel signal estimation result after spatial resolution enhancement. It is worth noting that in order to maintain the normalization of system energy, after performing spatial domain oversampling Fourier transform on the input channel signal in step 1, the transformed signal should be divided by Elimination of oversampling factor Oversampling factor ρ os and the number of antennas N ant In step 3, the estimated channel signal in the time-frequency domain is multiplied by Make compensation.

[0045] As an important supplement, the novel channel estimation method for beam domain resolution enhancement proposed in the present invention can also be extended to enhance channel clustering.

[0046] In a wireless channel, signal propagation will form multipath components, which are caused by different propagation paths. When the arrival angle or departure angle of the multipath components are similar, they can be classified into the same cluster. This situation is particularly significant in millimeter wave communications. Due to the high path loss of high-frequency signals, millimeter wave channel signals usually show strong sparsity characteristics. Not all angles will have significant signal components, and only a few strong incoming wave directions have significant signal components. Accordingly, these signal components in the same incoming wave direction are regarded as a cluster. Through the spatial resolution enhancement technology of the present invention, the original channel signal is subjected to an oversampling-based Fourier transform in the spatial beam domain, which is equivalent to encrypting the sampling points in the spatial beam domain. In this way, the clustering effect of the channel signal will be significantly enhanced. The channel signal after clustering enhancement can be applied to various subsequent processing including but not limited to: cluster-based channel estimation, cluster-based user resource scheduling and allocation, cluster-based user grouping, cluster-based beamforming, cluster-based precoding, etc. So that the processing performance based on clustering is enhanced.

[0047] The above descriptions are merely embodiments of the present invention and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A novel channel estimation method with enhanced beam domain resolution, characterized in that: It is divided into the following steps; Step 1, beam domain oversampling: For the antenna dimension that needs spatial resolution enhancement, perform Fourier transform based on oversampling to obtain the beam domain channel H with enhanced spatial resolution EhdBeam ; Specific oversampling coefficient ρ os , depending on the antenna configuration and system design requirements; Step 2: Beam-domain channel estimation based on spatial resolution enhancement: The beam-domain channel H after spatial resolution enhancement obtained in step 1 is EhdBeam , perform corresponding beam-domain channel estimation processing; Step 3, beam domain oversampling reconstruction: retain the beam position selected in step 2, set the beams at other positions to zero, and perform inverse Fourier transform on the screened beam domain channel to obtain the time-frequency domain channel signal estimation result with enhanced spatial resolution.

2. According to claim 1, a novel channel estimation method with beam domain resolution enhancement is characterized in that: The beam domain transform based on Fourier oversampling in step 1 can act on any one or a combination of the horizontal and vertical dimensions of the antenna array corresponding to the received channel signal, thereby enhancing the spatial resolution of the antenna array in the corresponding dimension.

3. The novel channel estimation method with beam domain resolution enhancement according to claim 1, characterized in that: When performing subsequent beam-domain channel estimation on the obtained spatial resolution enhanced channel signal in step 2, the following methods are included but not limited to: channel estimation based on threshold screening, channel estimation based on noise energy screening, and channel estimation based on beam-domain peak iterative elimination.

4. The novel channel estimation method with beam domain resolution enhancement according to claim 1, characterized in that: For the oversampled Fourier transform and inverse transform performed in step 1 and step 3, the transformed channel energy is scaled and restored according to the oversampled coefficient and the number of days, respectively.

5. The novel channel estimation method with beam domain resolution enhancement according to claim 1, characterized in that: The beam domain signal is clustered; the parameters referenced by the clustering include but are not limited to: beam delay, beam arrival angle and beam energy; the parameters used for specific clustering are one or more combinations of the listed reference parameters.

Images