Channel estimation for configurable surfaces

Abstract

The present disclosure relates to channel estimation at a receiving device of a communication system using a (re)configurable surface. The channel estimation comprises a beamforming search to obtain a trained reflection coefficient of the configurable surface and an angle of arrival AoA of a signal at the receiving device. Then, based on the configurable surface and the AoA obtained at the receiving device, the reflection coefficient of the configurable surface is derived for an ideal channel portion between a transmitting device and the configurable surface. Based on the relationship between the trained reflection coefficient and the estimated reflection coefficient, an estimation of the channel characteristics between the transmitting device and the configurable surface is performed. Such channel estimation can be used in user mobility tracking.

Description

Technical Field

[0001] The present disclosure relates to channel estimation at communication devices that communicate with other communication devices via reflections on configurable surfaces. Background Art

[0002] Millimeter-wave (mmWave) communication has become one of the key technologies of 5G communication systems. Although mmWave can achieve high data rates and high spectral efficiency due to its wider signal bandwidth, it usually suffers from severe path loss and line of sight obstruction between communication devices.

[0003] Reconfigurable intelligent surface (RIS) has been discussed as a technology to improve wireless communication performance. Elements of RIS can reflect, refract, absorb or focus incident waves in any desired direction. This functionality can help overcome the mmWave propagation condition issues mentioned above, including path attenuation and blocking.

[0004] However, due to the passive nature of RIS elements and the estimation overhead incurred by multiple-input multiple-output (MIMO) systems, channel estimation in RIS-assisted communications remains a major concern. Summary of the invention

[0005] Methods and techniques are described for estimating a channel in a communication system including a receiving device and a transmitting device communicating with each other via a configurable surface.

[0006] The invention is defined by the scope of the independent claims. Some further exemplary embodiments are subject matter of the dependent claims.

[0007] According to an embodiment, a method is provided for estimating characteristics of a channel (BS-UE) between a transmitting device and a receiving device at a receiving device, the channel including reflections on a configurable surface, the method including performing a beamforming search for signal transmission through the channel (BS-UE), thereby obtaining: a trained reflection coefficient of the configurable surface, and an angle of arrival AoA of the signal at the receiving device; estimating an estimated reflection coefficient of the configurable surface for an ideal channel (BS-RIS) between the transmitting device and the configurable surface based on the trained reflection coefficient of the configurable surface and the AoA obtained at the receiving device; and estimating characteristics of the channel (BS-RIS) between the transmitting device and the configurable surface based on a relationship between the trained reflection coefficient and the estimated reflection coefficient.

[0008] These and other features and characteristics of the presently disclosed subject matter, as well as the functions of the related elements of operation and structure, together with the combination of parts and economy of manufacture, will become more apparent when considering the following description and the appended claims with reference to the accompanying drawings, all of which constitute a part of this specification. However, it should be clearly understood that the drawings are for illustration and description purposes only and are not intended as a definition of the limitations of the disclosed subject matter. As used in the specification and claims, the singular forms of "a", "an", and "the" include plural references unless the context clearly dictates otherwise. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] An understanding of the nature and advantages of the various embodiments may be realized by referring to the following drawings.

[0010] Figure 1 is a schematic diagram of a system for implementing some embodiments of the present disclosure.

[0011] Figure 2 is a schematic diagram illustrating the process of a hierarchical beam search algorithm for channel estimation.

[0012] 3 is a block diagram illustrating an exemplary embodiment of an apparatus for channel estimation.

[0013] Figure 4 is a flow chart illustrating a method for channel estimation and tracking.

[0014] Figure 5 is a schematic diagram illustrating the main beam patterns in the azimuth domain (left figure) and the elevation domain (right figure) for an ideal channel G when G is a geometric model with 5 paths.

[0015] Figure 6 is a schematic diagram of a communication device illustrating a receiver and a transmitter of a channel including a RIS.

[0016] According to some example embodiments, like reference numerals and symbols in the various drawings represent like elements. DETAILED DESCRIPTION

[0017] For the purposes of the description below, the terms "end," "upper," "lower," "right," "left," "vertical," "horizontal," "top," "bottom," "lateral," "longitudinal," and their derivatives will relate to the disclosed subject matter as it is oriented in the accompanying drawings. However, it should be understood that the disclosed subject matter can assume various alternative variations and step sequences unless expressly specified to the contrary. It should also be understood that the specific devices and processes illustrated in the drawings and described in the following specification are merely exemplary embodiments or aspects of the disclosed subject matter. Accordingly, unless otherwise indicated, specific dimensions and other physical characteristics related to the embodiments or aspects disclosed herein should not be considered limiting.

[0018] Unless clearly described in this way, the aspects, components, elements, structures, actions, steps, functions, instructions and / or the like used herein should not be interpreted as critical or necessary. In addition, as used herein, the articles "one" and "an" are intended to include one or more projects, and can be used interchangeably with "one or more" and "at least one". In addition, as used herein, the term "set" is intended to include one or more projects (for example, related projects, unrelated projects, a combination of related and unrelated projects, and / or the like), and can be used interchangeably with "one or more" or "at least one". In the case of being intended to only one project, the term "one" or similar language is used. In addition, as used herein, the term "has", "have", "having" or the like is intended to be an open term. In addition, unless otherwise expressly stated, the phrase "based on" is intended to mean "based at least in part on".

[0019] Reconfigurable smart surfaces (RIS) are also referred to as intelligent reconfigurable surfaces (IRS). Herein, they are also more simply referred to as reconfigurable surfaces or configurable surfaces. In other words, a configurable surface is a surface comprising a plurality of elements having (re)configurable reflective properties.

[0020] Typically, a RIS is a uniform planar array with a large number of reflective elements. Each reflective element is adapted to passively reflect the incident signal and introduce a phase shift to it. In contrast to a mirror, a RIS is able to adjust the reflection angle and the electric field strength. RIS technology can support MIMO systems to control and improve wireless channels. The adjustable passive elements can individually direct the incident electromagnetic (EM) wave to any specific direction simply by changing its phase and gain. Adjusting these elements can allow the multipaths of the reflected signals to be aligned so that they are constructively added at the receiver. With the appropriate RIS size and reflection coefficient, the reflected signal is such a beam, where the width of the beam is inversely proportional to the size of the RIS. RIS elements passively reflect signals, they are easy to implement, have low-cost deployment, and most importantly do not cause noise amplification.

[0021] On the other hand, RIS also brings challenges such as channel estimation. Since RIS consists of a large number of passive elements, RIS-assisted communication networks face difficulties in reliably estimating the channel. To overcome these difficulties, some channel estimation techniques have been proposed, most of which follow some simplifying assumptions.

[0022] In mmWave, channel estimation becomes even more critical and few people have paid attention to these high frequency bands. In some works, a priori knowledge about the channel between the base station (BS) and the RIS is used. However, it is not practical to consider that the channel BS-RIS is known and time-invariant because the mmWave channel is sensitive to any small changes that may be caused by any scatterers between the BS and the RIS. In addition, some two-stage cascaded channel estimation protocols have been proposed by exploiting the sparsity of the mmWave MIMO channel between a single BS, RIS and user equipment (UE). In the receiving stage, a beam search method with high angular domain information is introduced; then in the transmitting stage, an adaptive grid matching pursuit algorithm is proposed to estimate the high-resolution cascade channel.

[0023] This channel estimation technique depends on the cascaded channel concept or some restrictive assumptions for estimating the channel BS-RIS-UE. Since the RIS reflects the signal and concentrates the energy to a specific direction, the position of the UE should be considered in the estimation process. The UE position may change over time so that user tracking may be a desired feature. In addition, it has been shown that the path loss is a function of the reflection coefficients of the RIS, which are usually ignored in the channel estimation process when the phase is optimized for channel estimation.

[0024] The present disclosure addresses channel estimation, beamforming, and user tracking in mmWave RIS-MIMO systems. By providing the relationship of the RIS design to the MIMO system, a three-stage framework can then be used, which is illustrated in one of the embodiments. Specifically, in one embodiment, the channel between the BS and the RIS is estimated using a hierarchical beam search. In addition, in other embodiments, the channel between the RIS and the user is estimated, for example, using an iterative resolution algorithm. In other embodiments, a tracking algorithm is used to track the channel parameters between the RIS and the user.

[0025] Reconfigurable (or configurable) intelligent surface (RIS) model

[0026] Figure 1An exemplary communication system 100 is illustrated, which may implement some embodiments of the present disclosure. Specifically, the communication system 100 includes a base station (BS) 110, a RIS 120, and a user equipment (UE, also referred to as a user or terminal) 135. The communication system is exemplary only. In general, a communication system for implementing the present invention may include a receiving device (mobile or fixed), a RIS, and a transmitting device (mobile or fixed).

[0027] Figure 1 The BS 110 in is a base station, such as an eNB or gNB of a 3GPP system of long term evolution (LTE) or new radio (NR) or another generation system. In other words, the BS can be an access network node of a cellular wireless system. However, the BS can also be an access point (AP) of a local area network (LAN), such as, for example, a wireless LAN according to the IEEE 802.11 standard. The embodiments presented herein can be easily used in systems such as 3GPP NR or IEEE 802.11be or other emerging systems. Another possibility is to use the present disclosure for any communication between two devices in device-to-device communication in the absence of or connection to a certain network infrastructure.

[0028] RIS 120 is a planar array of passive reflective elements with configurable phase shift and possibly / optionally configurable attenuation. Figure 1 A square RIS with 3×3 elements is shown schematically only, resulting in an element number M RIS =9. However, this is only a schematic representation. In practice, the size of the RIS may be much larger. For example, M RIS =16, 64 or 256, or any other square size. The RIS is not necessarily square. It can be rectangular or have any other shape. Several different shapes have been studied, such as square, rectangular, hexagonal, circular, etc.

[0029] Figure 1 UE 135 in is illustrated as being at time t k A user device at a specific location, which may be at a specific time t k+1 A mobile device UE 130 moving towards different locations. However, the present disclosure is not limited thereto, and, in general, UE 135 may be any communication device, such as a machine-type device for Internet of Things (IoT) applications, a base station, an access point, a repeater, or any other wireless infrastructure entity. In general, RIS 120 may simultaneously reflect beams directed to different receiving devices.

[0030] The channel between BS 110 and RIS 120 is denoted as G. The channel between RIS 120 and UE 135 is denoted as H. Specifically, G and H refer to channel matrices that describe the characteristics of the respective channels. The channel matrix specifies how the channel modifies the input signal. For example, an input signal x transmitted from multiple antennas (and therefore represented as a vector) is multiplied by the channel matrix G, and a noise signal n (also a vector) is added, resulting in a signal y=G·x+n (the size of the vector corresponds to the number of receive antennas). Therefore, in this example, and They represent the channels between BS-RIS and RIS-UE (represented by channel matrices). Symbol M BS and M UE denote the number of antennas (antenna array elements) at BS 110 and UE 135, respectively.

[0031] In the communication system 100, not only the RIS 120, but also the BS 110 and the UE 135 may be equipped with an equidistant uniform planar array (UPA) as an antenna structure with a half-wavelength distance between antenna elements. The uplink and downlink transmissions may use a time-division duplex (TDD) protocol that exploits channel reciprocity for obtaining channel state information (CSI) at the RIS 120 in both link directions (such as uplink and downlink, or generally in both directions between communicating entities). It is typically assumed that the BS 110 has M RF RF chains, where the number of these chains is much smaller than the number of antenna array elements and greater than the number k of UEs 135, i.e., k≤M RF < <M BS UE 135 is considered to have one RF chain. RIS 120 can be placed close to the UE side and away from the BS side to reduce (minimize) the path loss effect. In order to fully utilize the function of RIS 120, the channel path between BS 110 and UE 135 is Figure 1 1 is shown as being blocked by obstacle 140. Therefore, there is no line of sight (LoS) between BS 110 and UE 135. The distance between BS and RIS is denoted as d g , and the distance between RIS and UE is represented by d h .

[0032] RIS elements can be placed in uniform rectangular shapes. For example, suppose the size is N RIS ×N RIS The reflection coefficients of these elements are MRIS =N RIS 2 It can be represented by the following reflection coefficient matrix:

[0033]

[0034] in, is the reflection coefficient of the (n, m)th RIS element, where α n,m ∈[0,2π) represents the phase shift caused by the (n, m)th element, γ n,m ∈[0,1] represents the reflection gain. Usually, assuming unity reflection gain is a good approximation: The indices n and m address the vertical and horizontal (row and column) coordinates of the corresponding antenna element within the UPA of RIS 120 .

[0035] In terms of computational convenience, another convenient representation of the reflection coefficient can be given as follows:

[0036] Θ = diag{vec(φ)}, (Equation #2)

[0037] This corresponds to a diagonal matrix, and the diagonal corresponds to the vectorized matrix φ. The reflection coefficient φ of RIS 120 n,m is configurable and may be configured (set) by BS 110 in some embodiments.

[0038] Considering Figure 1 In the system model shown, the reflected signals from each element of RIS 120 can be configured to be phase aligned to enhance the received signal power. The free space path loss can be calculated as shown in W. Tang, M. Z. Chen, X. Chen, J. Y. Dai, Y. Han, M. Di, Y. Zeng, S. Jin, Q. Cheng, and T. J. Cui, "Wireless communications with reconfigurable intelligent surface: Path loss modeling and experimental measurement", arXiv preprint arXiv:1911.05326, 2019, which is incorporated herein by reference.

[0039] When it is assumed that the signals will be reflected towards the same location and G is an ideal channel, i.e., it has unity gain, the channel with the minimum path loss can be considered as the channel description. However, when it is assumed that G is a sparse channel, where the LoS path between BS 110 and RIS 120 is the dominant path, it is observed that the beam is shifted towards a direction different from the location of UE 135. Sparse channels have a smaller number of scattering, resulting in a small number of correlated multipaths, which gives rise to a channel representation with a small number of non-zero elements. On the other hand, when the scattering of channel G is very rich, UE 135 will receive very low power from RIS 120, so that RIS may reflect worse than ordinary reflectors (metal surfaces, walls, etc.). Therefore, for successful reflections, G can be estimated separately and then equalized at RIS 120 by simply inverting its effect.

[0040] After reflecting the beam at RIS 120, UE 135 estimates and equalizes H to successfully complete the communication. In other words, RIS 120 alone performs the following two different operations:

[0041] 1) Accumulation, which collects all the energy received by each element (resulting in a gain of proportional to each other), and then align them by canceling the effect of channel G.

[0042] 2) Beamforming / steering, the RIS 120 behaves like a virtual BS (usually a virtual transmitter) and focuses or directs the incident electromagnetic waves to the location of the UE.

[0043] As mentioned above, most known methods assume that the UE 135 is stationary and that the BS 110 always has LoS with the RIS 120. However, these assumptions are unrealistic and tend to limit the use of the RIS 120. They are based on H cascaded =G·H gives the result of the cascade channel model. This representation makes channel tracking almost impossible because H cascaded Any change in may be due to changes in G or H or both. Here, since it is G rather than H that affects the phase of RIS, G can be estimated alone, and it becomes feasible to estimate H, which can enable tracking of UE 135.

[0044] BS-RIS channels and channel matrix G

[0045] To estimate the channel, training (reference) symbols s are transmitted from BS 110 to UE 135. Specifically, the training symbols s may be transmitted via an orthogonal precoded beam for user 135 so that there is no inter-user interference to other users in the system. Under this assumption, the analysis may be limited to one representative UE 135 without loss of generality. The sparsity of the channel is exploited using geometric channel modeling under the assumptions of flat fading and perfect timing and frequency synchronization. Given the channel matrix The model represented can be given as follows:

[0046]

[0047]

[0048] Among them, L g is the number of channel paths (i.e., paths between BS 110 and RIS 120) received at RIS 120, are the elevation and azimuth angles of arrival (AoA) at the RIS of the lth BS-RIS-path, are the elevation and azimuth angles of departure (AoD) from the base station of the l-th BS-RIS-path, and z g,l is the complex channel coefficient between BS-RIS at the lth path. In addition, diag(z g ) is a vector z on its diagonal g The elements z g,l The diagonal matrix of represents the array response vector of the UPA of the RIS 120, and the symbol represents the array response vector of the UPA of BS 110. As can be seen from the above notation, the superscript B refers to the base station 110 (usually a transmitting device), R refers to the RIS 120, and U refers to the UE 135 (usually a receiving device).

[0049] Typically, the array response vector of UPA is expressed as follows:

[0050]

[0051] in and For i∈{RIS,BS}. Notation represents the Kronecker product, λ is the wavelength of the signal, Nx and Ny represent the number of elements of UPA in the vertical and horizontal dimensions, respectively, and du and dv represent the distances between adjacent elements of UPA in the vertical and horizontal directions, respectively. and is the matrix given by:

[0052] For i∈{RIS,BS} and j∈{R,B}.

[0053] Note that for sparse channels, the path L g The number of is much lower than the number of antenna elements at the RIS. For example, for about 10 antenna elements, 1 or 2 paths can be considered.

[0054] RIS-UE channel and channel matrix H

[0055] Similarly, It can be expressed as:

[0056]

[0057] in, are the elevation and azimuth angles AoA of the lth RIS-UE-path, are the elevation angle and azimuth angle AoD of the lth RIS-UE-path, respectively. In addition, symbol represents the array response vector of the UPA of UE 135, L h is the number of paths between RIS 120 and UE 135, And diag(z h ) is a vector z on its diagonal h The elements z h,l The diagonal matrix

[0058] Accordingly, and is the matrix given by:

[0059] For i∈{UE,RIS} and j∈{U,R}

[0060] BS-RIS-UE channel and channel matrix H eff

[0061] Total channels between BS-RIS-UE Given by:

[0062]

[0063] where β(dg, dh, θdes, φdes) is the total path loss calculated as in the above-mentioned W. Tang et al. paper. That is, the path loss is usually the distance dg between the base station and the RIS, the distance dh between the RIS and the UE (see Figure 1 ) and the departure angle of the signal from the RIS.

[0064] Channel Estimation Framework

[0065] To estimate the channel, the effective channel in (Equation #6) can be obtained after the modified channel matrix and The aspect is rewritten as follows:

[0066]

[0067] in,

[0068] as well as

[0069]

[0070] Here, Among them, arg(z g,l ) represents the coefficient Since G is directly responsible for changing the RIS phase, only in z g,l =arg(z g,l ) aspect, it may make more sense and replace the channel gain |z g | is included in H. z contains the channel coefficient z of channel H h,l and the channel gain of channel G |z g,l |.

[0071] Typically, the modified channel matrix A matrix corresponding to a channel matrix G obtained from a channel between a transmitting device (110) and a configurable surface (120) by setting the absolute value of each element of the channel matrix G to 1, and further comprising the steps of: based on the modified channel matrix Determine the modified channel matrix In other words, the modified channel matrix Only the phase change caused by the BS-RIS channel is considered. The power change of G becomes the modified matrix part of.

[0072] Writing the channel in this form allows for separate estimation of H and G. A hierarchical beam search algorithm can then be used to estimate the BS-RIS channel G. In addition, for example, the RIS-UE channel H can be estimated using angles derived from a beam search algorithm to estimate only the channel path coefficients by using an iterative reweighting algorithm.

[0073] BS-RIS channel estimation

[0074] Since mmWave channels are sparse and the new BS-RIS = ... ),estimate The problem becomes equivalent to each path In other words, This causes a shift in the reflected beam, so estimating this shift leads to estimating itself. This is performed in the following three steps, which are described in more detail below:

[0075] 1) estimating the AoA and AoD for the signal reflected by the RIS 120, and

[0076] 2) By changing the angle and Substituting into the equation (Equation 16), we obtain The reflection coefficient of RIS under the influence of is then obtained from the matrix Θ in equation (Equation 2).

[0077] 3) Then, the reflection coefficient thus obtained is compared with the last coefficient set by the BS, and their phases are subtracted from each other to obtain

[0078] Estimated Angle of Arrival (AoA) and Angle of Departure (AoD)

[0079] Generally, by performing a beamforming search (or beamforming training) for signal transmission over a BS-UE channel, a trained reflection coefficient of the configurable surface 120 and an angle of arrival (AoA) of the signal at the receiving device 135 may be obtained.

[0080] More specifically, and This can be determined by an exhaustive beam search algorithm. In the exhaustive beam search algorithm, possible angles are tested to find an optimal AoA and AoD. However, this method may take a lot of time due to its complexity.

[0081] For example, as further explained below and as Figure 2 As shown in FIG. 1 , the beamforming search can be performed basically only between the RIS and the UE. More specifically, the and RIS The AoD and AoD of the RIS channel may be kept fixed during the beamforming search. For example, the locations of the base station and the RIS and thus the optimal AoD and AoD of the BS-RIS channel may be known (e.g., predetermined). The beamforming search may then be performed by adjusting the phase of the RIS (i.e., by adjusting the reflection coefficient). In other words, in the beamforming search, the AoD at the RIS is can be changed, and the RIS may remain fixed. Thus, an optimal AoD from the RIS to the UE corresponding to an optimal beam direction may be determined, wherein the term "optimal" refers to, for example, reception quality and / or signal strength at the UE. Thus, the trained reflection coefficient corresponds to the optimal AoD from the RIS found during the beamforming search. However, the present invention is not limited thereto. In general, the trained reflection coefficient may be obtained by a beamforming search using different beam positions at the RIS. and / or differences at RIS For example, the base station The reflection coefficient at the RIS can be varied to find an optimal pair of beam direction at the base station and beam direction of the reflected beam at the RIS.

[0082] Accordingly, in the following, a two-stage beam training method is used, including a primary beam search and an auxiliary beam search, as described in detail in W. Wu, D. Liu, Z. Li, X. Hou, and M. Liu, "Two-stage 3D codebook design and beam training for millimeter-wave massive MIMO systems," IEEE 85th Vehicular Technology Conference (VTC Spring), Sydney, New South Wales, Australia, 2017, pp. 1-7, which is incorporated herein by reference. However, it is noted that the present invention is not limited to this beam training method, and generally any other method can be used, including the exhaustive search mentioned above or its modification.

[0083] In the following, the hierarchical two-stage beam training method is briefly summarized. For simplicity, the training process is only described in azimuth. By analogy, the process for elevation can be derived. Figure 2 A hierarchical beam search algorithm process for channel estimation is described. The primary codebook generates basic directional beams. The number of phase shifts for the primary codebook may be very limited to reduce complexity. Small-sized auxiliary codebooks provide finer beams focused on each primary beam. Here, the term "codebook" refers to a predefined set of beams in the absence of channel state information.

[0084] In an exemplary embodiment, the main search uses a hierarchical search to reduce the search time. Here, a bidirectional (binary) tree search is used at each level. Let represents the codeword of the nth beam vector at the lth layer. Correspondingly, the codeword is about the beam vector ( i is "r" on the RIS side or "u" on the UE side, and l is a scalar of layers. The beam vector specifies the antenna factors that form the beam. In each layer l, there are only 2 l The antennas are activated (in each direction, the transmission direction is as follows Figure 2 The receiving direction is shown as Figure 2 In general, there will be possible beams, and Each pro code word With two sub-code words and The purpose is to obtain the AoA angle through multiple steps corresponding to the main beam search including the V layer and the auxiliary beam search including the V' layer. and AoD angle like Figure 2 As shown, Layer 1 (i.e., first layer) operation begins by testing four possible pairs of wide beams in four consecutive time slots, with RIS 120 operating in reflection mode using (determine AOD), UE 135 uses in receive mode (Determine AoA). In layer 2, more antenna elements are used to generate four narrower beams, two in reflection mode In receive mode, two and so on until the Vth layer of the main search is terminated. For example, two beams (the same applies to reflection and reception modes) are obtained by screening all possible beam patterns of the current lth layer to select the two beams with the greatest similarity to the best beam of the previous (l-1)th layer. The similarity can be measured by any similarity measure (metric), for example, as shown in W. Wu et al. above, for example, by comparing the array factors of the beams. However, the present disclosure is not limited to comparing array factors or any particular similarity measure.

[0085] The signal obtained from layer l can be written as:

[0086]

[0087] in, s=[s1,s2,…,s Z ] T is the Zx1 vector of transmitted training symbols (known at the BS and UE), and n is a vector with zero mean and variance The Zx1 complex Gaussian noise vector.

[0088] At each level l, we search for the beam vector pair that satisfies the following maximum received SNR:

[0089]

[0090] After V layers of primary beam search, the optimal (best) beam vector pair is obtained in the main reception phase.

[0091] The primary codebook may be selected in a variety of different ways, and the present disclosure is not limited to any particular method. By way of example only, the primary codebook matrix τ in azimuth for the K beam patterns is the number of possible shifts. That is, the discrete shift is 2pi / τ, and N RIS The elements are given in the above-mentioned literature by W.Wu et al. as follows:

[0092]

[0093] Where n = 0, 1...N RIS -1 and k = 0, 1...K-1. The parameter K may depend on the design (characteristics) of the antenna. The codebook may be designed so that it has possible states (i.e., possible beams) and completely spans the azimuth range. This design is achieved by choosing the desired values ​​of N, K, and τ. Similarly, the main beam codebook matrix in elevation is given by:

[0094]

[0095] The transmit phase begins after acquiring the primary codebook, where a secondary beam search is performed by rotating the primary beam to create a higher resolution secondary beam.

[0096] Specifically, a predefined number (such as one, two or more) of auxiliary beam patterns are evenly and symmetrically distributed on both sides of the known best primary beam. These beams define the auxiliary codebook. Finally, when the above equation #9 is satisfied, a pair of beam vectors is considered to be the best pair among the pairs based on the auxiliary codebook. Figure 2As shown, for example, nine beam pairs including the best main beam and two adjacent beams (transmit, Tx) and (receive, Rx) on either side of the RIS and US are evaluated to find the beam pair with the maximum receive SNR. The beam may be selected, and the search may be further continued by selecting a beam direction with a higher SNR potential and performing a beam scan to find a pair with the maximum SNR. If the SNR starts to decrease, the beam search may be stopped. As will be clear to one skilled in the art, variations of the auxiliary search are possible. The above method is intended to illustrate only one possibility. In general, as Figure 2 As shown, auxiliary searches may be performed in the V' layer.

[0097] Since the optimal transmission beam is obtained by weighting vector = ...

[0098]

[0099] In other words, AoA and AoD correspond to the directions of the best receiver (UE) and transmitter (RIS) beams found in beamforming training.

[0100] In this model, the RIS 120 is considered to be located near the UE 135. Therefore, it can be assumed that the antenna array of the UE is always parallel to the RIS, so Applicable. In other words, the AoD of the signal at the configurable surface may be estimated to be equal to the AoA at the receiving device. In general, the AoD of the signal at the configurable surface may be estimated from or based on the AoA at the receiving device. This may be based on some a priori knowledge or estimate of the geometry of the RIS and UE, such as mutual position.

[0101] Once the modified matrix According to some embodiments it is estimated that the modified channel matrix is determined by an iterative reweighting algorithm. Note that the present disclosure is not limited to the iterative reweighting algorithm. The separation of the two matrices for the BS-RIS channel and the RIS-UE channel makes it possible to use any other estimation / determination algorithm for the matrix of determination.

[0102] The results of the estimation can be used in many different ways. For example, the channel estimation can be used for channel equalization at the UE. Alternatively or additionally, the UE can report the channel estimation or some estimated channel parameters to the base station, and the base station can take appropriate actions. For example, the base station can configure its own transmission characteristics, such as departure angle, and / or transmission power, and / or beamforming pattern accordingly.

[0103] In some embodiments, the trained reflection coefficients are configured by the transmitting device. Note that the estimated channel characteristics can be used to determine a new set of reflection coefficients for the RIS, and configure the RIS accordingly. In some embodiments, the configuration of the RIS can be performed by the BS (transmitting device) based on the channel estimates reported from the UE. However, it is possible that the RIS is configured by the UE. Alternatively, as mentioned above, in general, the present disclosure is applicable to communications between any devices, such as direct communications between two devices in which no layers such as AP-STA or gNB-UE are involved. Therefore, in principle, the receiving device or the transmitting device or both may be allowed to configure the RIS based on the channel estimate by resetting the reflection coefficients of the RIS.

[0104] As described above, in some embodiments, the beamforming search includes a hierarchical beamforming search, which includes a first stage of performing a beamforming search in multiple (V) layers, wherein for each current layer after the first layer of the multiple layers: (1) a predefined number of beams are searched to find the best beam in the current layer, wherein the searched predefined number of beams are selected based on the best beam of the previous layer of the current layer; and (2) the number of antennas that contribute to beamforming is increased in the current layer compared to the previous layer of the current layer. Typically, the search can end with a main search, and the best beam found therein can be adopted to estimate AoA and AoD, channel and / or reflection coefficient. Alternatively, some other searches can be performed based on the best beam found in the main search. The present disclosure is not limited to any particular other beam search.

[0105] Nonetheless, in some embodiments, as mentioned above, the hierarchical beamforming search also includes a second stage of searching multiple beams near the beam found after the first stage. The secondary search (second stage search) can search a predetermined number b of beams near the best beam. This search can be performed iteratively more than once, i.e., in V' layers. For example, in each layer of such secondary search, the best beam of the previous layer is adopted, and b beams near the best beam are searched. This can be performed a predetermined and fixed number of times (e.g., V'), and / or the search can be terminated if the received quality (e.g., SNR or SINR or other channel quality metric) decreases.

[0106] Estimation of the "ideal" reflectance

[0107] In general, the ideal reflection coefficient of the configurable surface 120 can be estimated based on the trained reflection coefficient of the configurable surface 120 and the AoA obtained at the receiving device 135. Here, it should be noted that the term "ideal reflection coefficient" refers to the reflection coefficient of the configurable surface in the case of an ideal channel BS-RIS. The term "ideal channel" here refers to a channel with unity gain (e.g., all coefficients of the corresponding channel matrix have an absolute value of 1).

[0108] The ideal reflection coefficient can be estimated based on the following formula:

[0109]

[0110] as well as

[0111]

[0112] in,

[0113]

[0114] as well as

[0115] Λ x (.) and Λ y (.) can be viewed as steering vectors in azimuth and elevation directions, respectively. Note that the indices x and y refer to the position of the RIS element (in the plane of the planar RIS). Furthermore, They are the elevation and azimuth destination angles (corresponding to ),and are the elevation and azimuth angles incident on the RIS (corresponding to ). In addition, λ is an imaginary unit of a signal wavelength, dx is a spacing between elements of the configurable surface in an x-direction (e.g., a vertical direction), and dy is a spacing between elements of the configurable surface in a y-direction (e.g., a horizontal direction). Specifically, the y-direction and the x-direction may be perpendicular to each other.

[0116] Typically, the ideal reflectance can be estimated based on the AoA at the configurable surface.

[0117] For example, the reflection coefficient can be estimated according to the following formula:

[0118] as well as

[0119] (Equation 16)

[0120] ·φn,m is the reflection coefficient of the (n, m)th element of the configurable surface,

[0121] γ n,m is the reflection gain of the (n, m)th element of the configurable surface, γ n,m It is scheduled.

[0122] ·j is the imaginary unit (j 2 =-1),

[0123] λ is the wavelength of the signal,

[0124] dx is the spacing between elements of the configurable surface in the x direction,

[0125] dy is the spacing between elements of the configurable surface in the y direction,

[0126] · is the predetermined elevation angle AoA at the configurable surface,

[0127] · is the predetermined azimuth angle AoA at the configurable surface,

[0128] · is the height AoD of the estimated AoD at the configurable surface, and

[0129] · is the azimuth angle AoD of the estimated AoD at the configurable surface.

[0130] here, is the elevation angle AoA at the configurable surface, is the azimuth angle AoA at the configurable surface. As mentioned above, is the elevation angle AoD at the configurable surface, are the azimuth angles AoD at the configurable surface, and these can be obtained through beamforming training. Then, and / or It can be obtained by approximating and / or taking into account prior knowledge about the relative positions of the BS and the RIS. For example, in some infrastructure-based wireless networks, the location of the BS may be fixed and known to the BS. In addition, the location of the RIS may be known to the BS. The BS may control and know the direction of both its own UPA and the UPA of the RIS. Therefore, based on the geometry of the BS and the RIS (approximate, measured or known), the relative position of the RIS may be obtained. and from the base station In a simplified model, for example, the AoA can be considered to be 0 (at least in elevation). However, this assumption is not required, and any other geometry can be considered. Equivalent to (Equation #16), the reflection coefficient can be estimated according to the following formula:

[0131]

[0132] Further details on some exemplary possible channel modeling can be found in the above-mentioned reference by W. Tang et al.

[0133] estimate

[0134] Generally, the channel BS-RIS between the transmitting device 110 and the configurable surface 120 may be characterized by a method based on a relationship (or a relationship related to a relationship) between a trained reflection coefficient and an estimated reflection coefficient.

[0135] More specifically, if the RIS phase is set to direct the beam of the reflected signal toward the location of the UE The beam will be distorted and due to the channel Due to the influence of , the radiation is shifted in different directions. Mathematically, this can be expressed as:

[0136]

[0137] Among them, G opt is the channel matrix corresponding to the best (or ideal) channel between the base station and the RIS, which is a single-path channel without any fading. According to (Equation #3), G opt Can be written as

[0138]

[0139] in, and are the elevation and azimuth angles AoD at the base station for an ideal channel, and are the elevation and azimuth AOA at the configurable surface for an ideal channel, respectively, and is the array response vector, as defined, for example, in (Equation #4). Angle and It can be known from the geometry of the deployment of the RIS and the base station. Specifically, the angle can be obtained based on the location of the base station, the location of the RIS, the orientation of the UPA of the base station and / or the orientation of the UPA of the RIS. It should also be noted that, in general, z g,optcan be used to arbitrarily (e.g., when appropriate) normalize the channel matrix. For example, as already assumed above, the channel matrix can be normalized to have unity gain. Accordingly, z g,opt Can depend on a MRIS For example, z g,opt Can be set to 1 (z g,opt =1) or set to

[0140] Thus, an ideal channel can be a free space channel or a channel with only one non-zero component. For example, the impulse response of an ideal channel may be an impulse, meaning that the signal is delayed, attenuated, or amplified, but not distorted. In some cases, the channel matrix of an ideal channel can also be considered a unitary matrix.

[0141] In addition, θ V′ is the phase set finally configured by the BS in the V'th phase of the beam search process. is the AoA at the RIS, expressed as the above in elevation and azimuth In other words, when there are multiple paths between the BS and the RIS, the angle may be different for different paths. is the AoD at the RIS and is expressed as above In the case where there are multiple paths between the RIS and the UE, the AoD may be different for different paths.

[0142] By using the angle obtained from (Equation 12) and substituting it into the following (Equation 13), can be directly estimated as:

[0143]

[0144] Or, equivalently, without considering the possible multiple paths, using simplified notation:

[0145]

[0146] in, and They are the elevation angle and azimuth angle AoD (also expressed here as and ); and They are the elevation and azimuth angles AoA (also expressed here as and );as well as and are the elevation angle and azimuth angle AoD (also expressed here as and ).

[0147] Note that for a known (e.g., predefined, estimated, or given by known geometry) angle (Equation #18) becomes

[0148] By adopting this design, the effect of channel G is known and can be determined by RIS simply by setting θ = θ V′ to control, which results in:

[0149]

[0150] in, can be set for any desired position θ des , And the channel estimation problem is reduced to only estimating

[0151] In other words, according to some embodiments, the estimated characteristic for a non-ideal channel (BS-RIS) is the modified channel matrix obtained according to the relationship (Equation #18) in:

[0152] - is a diagonal matrix containing the estimated reflection coefficients on its diagonal,

[0153]

[0154] -Θ V′ is a diagonal matrix that includes the trained reflection coefficients on its diagonal; and

[0155] -(Θ V′ ) -1 is a diagonal matrix Θ V′ The inverse matrix of .

[0156] In some embodiments, as mentioned above, The effect of is known and its effect can be eliminated at the RIS. For example, the RIS can be configured (e.g., by the BS and / or by the UE) with the following reflection coefficient θ:

[0157]

[0158] This is achieved by considering the following: The result of the effective channel should be as given in equation #19, which is equivalent to:

[0159]

[0160] Then, the following can be found:

[0161]

[0162]

[0163] Typically, the reflectance of the configurable surface (120) can also be calculated based on To set, this can be derived from the above equation in combination with (Equation #18).

[0164] To control RIS in this way, only G needs to be compensated (e.g., by setting the reflection coefficient as shown above). H is left to the user (e.g., UE) to estimate. In other words, in some embodiments, the BS can control the RIS reflection coefficient to compensate for the estimated channel Therefore, the phase modification of the channel between the BS and the RIS can be compensated by adjusting the reflection coefficient of the RIS. The remaining influence of the channel between the RIS and the UE and the power on the channel between the BS and the RIS can be compensated by the UE based on the estimated modified channel matrix.

[0165] In particular, the UE (usually a receiving device) can be configured to signal some information about the channel estimation to the BS (usually a transmitting device). For example, such information can be, for example, the best beam found, and / or the best AoA at the UE found in the beam search, and / or the estimated reflection coefficient, etc. Correspondingly, the BS (usually a transmitting device) can be configured to receive the information and configure the reflection coefficient of the configurable surface accordingly.

[0166] RIS-UE channel estimation

[0167] Without loss of generality, assuming that one RF chain is activated on the BS side and Z symbols are transmitted, the channel estimation model given in "Super-resolution channel estimation for mmwave massive MIMO with hybrid precoding" by C. Hu, L. Dai, T. Mir, Z. Gao, and J. Fang, 2018 IEEE Journal of Vehicular Technology, Vol. 67, No. 9, pp. 8954-8958 (incorporated herein by reference) is adopted here to estimate the path gains of all paths. Note that this estimation is only exemplary and other methods can be applied to estimate the RIS-UE channel with knowledge of the estimated phase of the BS-RIS channel.

[0168] The system model is given as follows:

[0169] y=Q H H eff Fs+Q H n, (Equation #20)

[0170] in, is the received signal at UE 135; and are the hybrid combination and precoder matrix respectively. The received signal at the UE can be clearly expressed as:

[0171]

[0172] Assumptions where each element x i is the i-th transmitted symbol. For channel estimation, known symbols at known indices are transmitted. Each received signal corresponding to a pilot symbol transmitted at time slot u is given by:

[0173]

[0174] In U time slots, U is sent in each time slot. p different pilot sequences, and

[0175]

[0176] y p =[y p,1 ,y p,2 ,...,y p,U ] T And Q=[q1,q2,...,q U ] T By setting and We get:

[0177]

[0178] Using the fact that mmWave channels are sparse, the channel The estimate of becomes equivalent to z, Ψ U and R An estimate of , and the problem is formulated as:

[0179]

[0180] in, Represents the number of non-zero elements, i.e., sparse channels The sparsest solution of is used for The estimated channel matrix is ​​, and ∈ is the estimation error tolerance. ||.|| F represents the Frobenius norm.

[0181] Since the logarithmic and penalty encourages sparsity more, the logarithmic norm can be used here instead of See, for example, J. Fang, F. Wang, Y. Shen, H. Li, and RS Blum, “Super-resolution compressed sensing for line spectral estimation: An iterative reweighted approach”, IEEE Transactions on Signal Processing, Vol. 64, No. 18, pp. 4649-4662, 2016. In addition, Ψ U , R Both have been obtained in the previous section using the beam search algorithm, so the optimization is performed only in terms of z, and problem P1 is given as:

[0182]

[0183] Among them, δ ensures that the logarithmic function is always within its domain. In addition to minimizing the number of paths, it is also necessary to minimize the channel estimation error. Therefore, the regularization parameter ζ>0 is added, and P2 is reshaped into the following optimization problem:

[0184]

[0185] As a result, minimization of P3 is equivalent to minimization of an iterative surrogate function, as can be seen in the above-mentioned paper by J. Fang et al.:

[0186]

[0187] Among them, D (i) Expressed as:

[0188]

[0189] as well as is the estimate of z at iteration i. Then, the optimization of (Equation #27) becomes as follows:

[0190]

[0191] in,

[0192]

[0193]

[0194] For the optimization (Equation #30), the next step is to obtain:

[0195]

[0196] Therefore, corresponding to the i-th iteration The optimal estimate of Given by:

[0197]

[0198] In this iterative approach, ζ is designed to be adaptive to accommodate sparse estimation and fast search.

[0199]

[0200] in, is the scaling factor, ζ {max} Make the problem well-conditioned and r (i) is the squared residual of the previous iteration. ζ is studied in detail in other works such as the above-mentioned J. Fang et al.

[0201] Channel Tracking

[0202] After estimating the channel parameters (i.e., channel coefficients, AoA, and AoD), and since the UE 135 can be mobile, a channel tracking method can be used to avoid frequent channel estimation. Channel tracking is typically much faster than a full channel estimation. They use the results from previous channel estimations.

[0203] There are many possibilities in the prior art on how to perform channel tracking. Any of them can be applied in the context of the present disclosure. There are remarkably fast, reliable and robust channel tracking algorithms that can allow efficient data transmission between transmitters and receivers in mmWave communication systems.

[0204] An exemplary channel tracking in mmWave systems is described in "Tracking angles of departure and arrival in a mobile millimeter wave channel" by C. Zhang, D. Guo, and P. Fan, IEEE International Conference on Communications (ICC), Kuala Lumpur, Malaysia, 2016, pp. 1-6, which is incorporated herein by reference. This channel tracking applies an Extended Kalman Filtering (EKF) to track AoA / AoD while the channel coefficients remain unchanged. This method can provide good results for more stable environments with low mobility, however, some difficulties may arise when the task is to track in a rapidly changing channel environment because it requires a prerequisite of a full scan, which can cause long measurements. In order to reduce the measurement time and provide an improved tracking algorithm, authors V.Va, H.Vikalo and RWHeath proposed an alternative solution in the 2016 IEEE Global Conference on Signal and Information Processing (Global SIP) in Washington, USA, pages 743-747, which only requires a single measurement estimated by the EKF and a beam switching design. In addition, Y.Yaplcl and I. The Least Mean Square (LMS) and BiLMS (Bidirectional LMS) algorithms are discussed in "Low-complexity adaptive beam and channel tracking for mobile mmWave communications" at the 52nd Asilomar Signals, Systems, and Computers IEEE Conference, Pacific Grove, California, USA, 2018, pp. 572-576, where both algorithms have advantages over the EKF algorithm under imperfect CSI conditions and have faster convergence characteristics as the SNR increases. Accordingly, after the above-mentioned channel estimation, tracking can be performed by using the EKF or LMS tracking algorithms because they can provide low complexity and good performance.

[0205] The tracking algorithm starts by setting a pair of transmit and receive beams based on the estimated azimuth and elevation angles AoA / AoD from the channel estimator. While tracking, the predicted channel parameters should remain close to the actual values ​​so that the UE 135 remains within half the beam width. Otherwise, if the tracking is no longer reliable or the beam path no longer exists, the channel parameters can be re-estimated. Specifically, the discrete time model for the received signal symbol period on the UE side is given in (Equation #21).

[0206] Assume that each vector in F is composed of To start the tracking process, the measurement function should be known. From (Equation #21), the measurement function is used to track the observed signal and can be given as:

[0207]

[0208] Among them, g measure Depending on the channel parameters, including the path coefficients, azimuth and elevation AoD / AoA angles from the two channels BS-RIS and RIS-UE, the LMS or EKF algorithms mentioned above can be used to track these parameters.

[0209] In summary, a method for tracking the location of a receiving device is provided, the method comprising: a method for estimating the characteristics of a channel (BS-UE) between a transmitting device 110 and a receiving device 135 at a receiving device 135 according to any one of the above-mentioned embodiments and examples, and based on the estimated characteristics of the channel (BS-UE) between the transmitting device 110 and the receiving device 135 and / or the modified channel matrix To track the location of communication devices.

[0210] Summary of the three-stage approach

[0211] Figure 4 An exemplary embodiment is summarized, which utilizes the above-described separate estimation of the channels between the BS and the RIS and between the RIS and the UE. In this general three-stage framework for RIS-assisted communication networks, all practical issues can be considered in realistic scenarios. Even if all RIS elements are passive, the scheme can estimate both the BS-RIS and RIS-UE channels separately. It starts with estimating the BS-RIS channel G using a hierarchical beam search algorithm, and then estimating the RIS-UE channel H by employing an iterative weighting algorithm to estimate only the channel path coefficients, applying the synthetic angles from the beam search algorithm. The proposed scheme then enables RIS-assisted communication to track mobile users. For example, well-known algorithms such as the Extended Kalman Filter (EKF) and the Least Mean Square (LMS) algorithm are used to track the parameters of the channel H.

[0212] Figure 4 An exemplary flow chart illustrating a method of a three-stage RIS channel estimation framework is shown.

[0213] The first phase is the estimation of the BS-RIS channel G. It starts with finding the AoA / AoD between the RIS and the UE. This phase includes the primary beam search (with layers 1 to V) and the secondary beam search (with layers 1 to V'). At step 410, the beam search is initiated by layer 1 of the primary search and continues 420 over the V layer of the primary search and the V' layer of the secondary search until the secondary search is completed 430.

[0214] Figure 5 The main beam patterns for an ideal channel G and in the azimuth domain (left) and elevation domain (right) are illustrated when G is a geometric model with 5 paths. RIS =8, τ=5, K=10.

[0215] The output of this beam search is Θ V′ as well as Based on these outputs, a modified channel matrix is ​​estimated as the channel characteristics between BS and RIS As shown in equation #18.

[0216] The first stage may also include setting the reflection coefficient 450 of the RIS according to the following equation:

[0217]

[0218] The second stage is to estimate the modified channel matrix 460 between the RIS and the UE. The input to this stage is the received signal Y, the pilot signal X, the combination matrix W, Ψ U , R , pruning threshold z th and the termination threshold ∈. The output of this stage is the path gain of all paths. The exemplary algorithm is summarized below as a high-level pseudo-code based on the above detailed embodiments, for example, referring to equations #32 and #33:

[0219]

[0220] The third stage of the framework is to track the channel parameters 470 of channel H. The input to channel tracking 470 is z opt , θ and These parameters are updated using, for example, an EKF algorithm. For example, the observation signal is tracked using equation #34. The updating and tracking are repeated until there is too much mobility 480, in which case the estimation 410-460 is repeated. Here, the term "mobility" can include UE movement, as well as large changes in the channel caused by other factors.

[0221] Note that step 450 of setting the RIS phase may be performed, but does not have to be performed. This step may be performed at any stage, for example, after estimation or at any time. Updating or setting the reflection coefficient may be performed according to the following method.

[0222] A method for setting a reflection coefficient of a configurable surface may include: performing a beamforming search for signal transmission through a channel (BS-UE) between a transmitting device and a receiving device, the channel including reflections on a configurable surface, thereby obtaining a trained reflection coefficient of the configurable surface and an angle of arrival AoA of the signal at the receiving device. The method may also include: estimating an angle of departure AoD of the signal at the configurable surface from the AoA at the receiving device; estimating a reflection coefficient of the configurable surface for an ideal channel (BS-RIS) between the transmitting device and the configurable surface based on the configurable surface and the AoA obtained at the receiving device; estimating a characteristic of a channel (BS-RIS) between the transmitting device and the configurable surface based on a relationship between the trained reflection coefficient and the estimated reflection coefficient; and setting the reflection coefficient of the configurable surface according to the following formula:

[0223] or

[0224]

[0225] Among them, Θ V′ is a diagonal matrix including the training reflection coefficients on its diagonal; θ des is the target elevation angle AoD at the configurable surface, is the target azimuth AoD at the configurable surface, is the predetermined elevation angle AoA at the configurable surface, is the predetermined azimuth angle AoA at the configurable surface, is the elevation angle AoD of the estimated AoD at the configurable surface; and is the azimuth angle AoD of the estimated AoD at the configurable surface.

[0226] In summary, estimating G will eliminate all channel effects between the BS and the RIS, making it easy to control the reflected beam direction by changing the phase of the RIS. In addition, by estimating H, it is possible to accurately recover the information sent from the BS to the UE. In other words, the estimated H or Heff can be used for channel equalization.

[0227] Implementation in hardware and software

[0228] Figure 6 An exemplary apparatus in which some embodiments may be implemented is described. Specifically, a receiving device 600_Rx is shown, which may include an apparatus for channel estimation for estimating, at the receiving device, characteristics of a channel (BS-UE) between a transmitting device and a receiving device, the channel including reflections on a configurable surface 650. The channel estimation apparatus may include a processing circuit 690 configured to control the transceivers 670-680 of the receiving device to perform a beamforming search for signal transmission through the channel (BS-UE), thereby obtaining: a trained reflection coefficient of the configurable surface 650, and an angle of arrival AoA of the signal at the receiving device 600_Rx. The processing circuit may also estimate the reflection coefficient of the configurable surface of an ideal channel (BS-RIS) between the transmitting device and the configurable surface based on the configurable surface and the AoA obtained at the receiving device 600_Rx. In addition, the processing circuit may also estimate the characteristics of the channel (BS-RIS) between the transmitting device 600_Tx and the configurable surface based on the relationship between the trained reflection coefficient and the estimated reflection coefficient.

[0229] like Figure 6 As shown, in the receiving device 600_Rx, the processing circuit 690 implements the function of baseband signal processing. The processing circuit can be a combination of one or more software and / or hardware. The method described in the above exemplary embodiments and embodiments can be implemented by the baseband signal processing part. In addition, the receiving device 600_Rx may include a transceiver, which may further include an analog radio frequency (RF) combiner Q 670 and one or more RF chains 680. The combiner 670 provides the received analog signal to one or more RF chains.

[0230] Note that the receiving device 600_Rx may include other parts known to those skilled in the art, such as parts performing coding and modulation, etc.

[0231] also, Figure 6A transmitting device 600_Tx is provided, which may be capable of setting a reflection coefficient of a configurable surface 650. Such a transmitting device 600_Tx may include a processing circuit 610. The processing circuit may perform the functions of baseband digital processing. The baseband processing may include (and the processing circuit 610 may be configured to): perform a beamforming search for signal transmission through a channel (BS-UE) between the transmitting device 600_Tx and the receiving device 600_Rx, the channel including reflections on the configurable surface, thereby obtaining: a trained reflection coefficient of the configurable surface, and an angle of arrival AoA of the signal at the receiving device.

[0232] The processing circuit (baseband digital processing) may also implement: estimating the departure angle AoD of the signal at the configurable surface relative to the AoA at the receiving device; estimating the reflection coefficient of the configurable surface for an ideal channel (BS-RIS) between the transmitting device and the configurable surface based on the configurable surface and the AoA obtained at the receiving device; and estimating the characteristics of the channel (BS-RIS) between the transmitting device and the configurable surface based on the relationship between the trained reflection coefficient and the estimated reflection coefficient.

[0233] Based on the above estimated value, the baseband digital signal processing may also include: Set the reflectance of the configurable surface (120), where θ V′ is a diagonal matrix that includes the trained reflection coefficients on its diagonal; θ des is the target elevation angle AoD at the configurable surface; is the target azimuth AoD at the configurable surface; is the predetermined elevation angle AoA at the configurable surface, is the predetermined azimuth angle AoA at the configurable surface; is the elevation angle AoD of the estimated AoD at the configurable surface; and is the azimuth angle AoD of the estimated AoD at the configurable surface.

[0234] from Figure 6 As can be seen in FIG, the transmitting device 600_Tx may also include a transceiver. The transceiver may include one or more RF chains 620 and an analog beamforming module 630, which receives a signal from the RF chain 620 and generates an analog beamformed signal, which is then transmitted from the UPA of the transmitting device 600_Tx toward the configurable surface 650 of the receiving device 600_Rx.

[0235] Note that even though some of the above examples are described with the transmitting device being a base station and the receiving device being a user equipment, this may also be reversed. Furthermore, in some embodiments, each BS and UE may implement both the receiving device and the transmitting device described above.

[0236] Figure 3A An exemplary hardware structure of a transmitting and / or receiving device is shown. Figure 3A A device 350 according to some exemplary embodiments is illustrated. The device 350 includes a memory 310, a processing circuit 320, and a wireless transceiver 330, which are able to communicate with each other via a bus 301. The device 350 may also include a user interface 340. However, for some applications, the user interface 340 is not necessary (e.g., some devices for machine-to-machine communication or the like). The device 350 may be, for example, a wireless module (such as 5G or Wi-Fi) or as part of a computer (laptop or tablet), it may be part of a mobile phone, a smartphone or other portable / personal device, etc.

[0237] The memory 310 may store a plurality of firmware or software modules implementing some embodiments of the present disclosure. The memory 310 may be read by the processing circuit 320. Thus, the processing circuit may be configured to execute the firmware / software implementing the embodiments. The processing circuit 320 may include one or more processors that perform the steps of any of the above methods in operation. This corresponds to a transmitting device (apparatus) including corresponding functional modules (units).

[0238] Figure 3B A schematic functional block diagram of a memory 310 and a functional code portion stored therein is shown. When executed on a processor 320, the functional code portion performs the following corresponding functions. Application code 360 ​​implements beam search (e.g., primary beam and secondary beam search). Application code 370 implements channel estimation between a BS and a RIS. Application code 380 implements channel estimation between a RIS and a UE. Application code 390 can implement channel tracking. These functions of application codes 360 to 390 have been described above. Depending on the application, the methods described herein (on the transmitter side and the receiver side) can be implemented by various means. For example, these methods can be implemented in hardware, operating system, firmware, software, or any combination of both or all of them. For hardware implementation, any processing circuit can be used, which may include one or more processors. Specifically, the hardware may include one or more of the following: an application specific integrated circuit (ASIC), a digital signal processor (DSP), a digital signal processing device (DSPD), a programmable logic device (PLD), a field programmable gate array (FPGA), a processor, a controller, any electronic device or other electronic circuit unit or element designed to perform the above functions.

[0239] If implemented as program code, the functions performed by the transmitting device (equipment) may be stored as one or more instructions or codes on a non-transitory computer-readable storage medium, such as memory 310 or any other type of storage device. Computer-readable media include physical computer storage media, which may be any available media that can be accessed by a computer or, in general, by processing circuit 320. Such computer-readable media may include RAM, ROM, EEPROM, optical disk storage, magnetic disk storage, semiconductor memory, or other storage devices. Some specific and non-limiting examples include compact disks (CDs), CD-ROMs, laser disks, optical disks, digital versatile disks (DVDs), Blu-ray (BD) disks, and the like. Combinations of different storage media are also possible, in other words, distributed and heterogeneous storage devices may be used.

[0240] The embodiments and exemplary implementations mentioned above illustrate some non-limiting examples. It should be understood that various modifications may be made without departing from the claimed subject matter. Specifically, modifications may be made to adapt the examples to new systems and scenarios without departing from the central concepts described herein.

[0241] Main contents of the embodiment

[0242] According to a first aspect, a method for estimating channel (BS-UE) characteristics at a receiving device (135) is provided. The channel is between a transmitting device (110) and a receiving device (135) and includes reflections on a configurable surface (120). The method comprises: performing a beamforming search for signal transmission through the channel (BS-UE) to obtain: i) a trained reflection coefficient of the configurable surface (120), and ii) an angle of arrival AoA of the signal at the receiving device (135). In addition, the method comprises: estimating an estimated reflection coefficient of the configurable surface (120) for an ideal channel (BS-RIS) between the transmitting device (110) and the configurable surface (120) based on the trained reflection coefficient of the configurable surface (120) and the AoA obtained at the receiving device (135). In addition, the method comprises: estimating a characteristic of the channel (BS-RIS) between the transmitting device (110) and the configurable surface (120) based on a relationship between the trained reflection coefficient and the estimated reflection coefficient.

[0243] According to a second aspect provided in addition to the first aspect, estimating the estimated reflection coefficient comprises the step of estimating a departure angle AoD of the signal at the configurable surface from the AoA at the receiving device (135). Furthermore, according to the second aspect, estimating the estimated reflection coefficient is based on the estimated AoD at the configurable surface.

[0244] According to a third aspect provided in addition to the second aspect, the reflection coefficient is estimated according to the following formula:

[0245]

[0246]

[0247] as well as

[0248] in

[0249] ·φ n,m is the reflection coefficient of the (n, m)th element of the configurable surface,

[0250] γ n,m is the reflection gain of the (n, m)th element of the configurable surface, γ n,m It is scheduled.

[0251] j is the imaginary unit,

[0252] λ is the wavelength of the signal,

[0253] dx is the spacing between elements of the configurable surface in the x direction,

[0254] dy is the spacing between elements of the configurable surface in the y direction,

[0255] · is the predetermined elevation angle AoA at the configurable surface,

[0256] · is the predetermined azimuth angle AoA at the configurable surface,

[0257] · is the height AoD of the estimated AoD at the configurable surface, and

[0258] · is the azimuth angle AoD of the estimated AoD at the configurable surface.

[0259] According to a fourth aspect in addition to the third aspect, the estimated characteristic of the non-ideal channel (BS-RIS) is based on the relationship The modified channel matrix obtained is in:

[0260] · is a diagonal matrix, whose diagonal includes the estimated reflection coefficients, Θ = diag{vec(φ n,m )};

[0261] · is the channel matrix of the ideal channel (BS-RIS) between the transmitting device (110) and the configurable surface (120), where and are the elevation and azimuth angles AoD of the ideal channel at the base station, respectively, and and are the elevation and azimuth angles AoA of the ideal channel at the configurable surface, respectively;

[0262] ·Θ V′ is a diagonal matrix that includes the trained reflection coefficients on its diagonal; and

[0263] ·(Θ V′ ) -1 is a diagonal matrix Θ V′ The inverse matrix of .

[0264] According to a fifth aspect in addition to the fourth aspect, the modified channel matrix A matrix corresponding to a channel matrix G obtained from a channel between a transmitting device (110) and a configurable surface (120) by setting the absolute value of each element of the channel matrix G to 1. In addition, according to the fifth aspect, the method comprises the following steps: based on the modified channel matrix Determine the modified channel matrix

[0265] According to a sixth aspect in addition to the fifth aspect, the modified channel matrix Determined by an iterative reweighting algorithm.

[0266] According to a seventh aspect in addition to the sixth aspect, the trained reflection coefficient is configured by the transmitting device.

[0267] According to an eighth aspect other than any one of the first to seventh aspects, the beamforming search includes a hierarchical beamforming search, and the hierarchical beamforming search includes a first stage of performing the beamforming search in a plurality of V layers. Specifically, for each current layer after the first layer of the plurality of layers: i) searching a predetermined number of beams to find the best beam in the current layer, wherein a beam among the searched predetermined number of beams is selected based on the best beam of a previous layer of the current layer; and increasing the number of antennas contributing to beamforming in the current layer compared to the previous layer of the current layer.

[0268] According to a ninth aspect in addition to the eighth aspect, the hierarchical beamforming search further includes a second stage of searching for a plurality of beams near the beam found after the first stage.

[0269] According to a tenth aspect, a method for tracking the location of a receiving device is provided. The method comprises: estimating the characteristics of a channel (BS-UE) between a transmitting device (110) and a receiving device (135) at a receiving device (135) according to any one of the first to ninth aspects. In addition, according to the tenth aspect, the method comprises: based on the characteristics of the channel (BS-UE) between the transmitting device (110) and the receiving device (135) and / or a modified channel matrix Track the location of communication devices.

[0270] According to an eleventh aspect, there is provided an aspect according to the tenth aspect for tracking the position of a receiving device using extended Kalman filtering or least mean squares LMS based tracking.

[0271] According to a twelfth aspect, there is provided a computer program product comprising a non-transitory computer readable medium containing code instructions, which when executed on one or more processors cause the one or more processors to perform the method according to any one of the first to eleventh aspects.

[0272] According to a thirteenth aspect, there is provided an apparatus for estimating characteristics of a channel (BS-UE) at a receiving device (135). The channel is between a transmitting device (110) and a receiving device (135) and includes reflections on a configurable surface (120). The apparatus comprises a processing circuit. The processing circuit is configured to: control a transceiver of the receiving device (135) to perform a beamforming search for signal transmission through the channel (BS-UE), thereby obtaining: i) a trained reflection coefficient of the configurable surface (120), and ii) an angle of arrival AoA of the signal at the receiving device (135). In addition, the processing circuit is configured to: estimate the estimated reflection coefficient of the configurable surface (120) for an ideal channel (BS-RIS) between the transmitting device (110) and the configurable surface (120) based on the configurable surface (120) and the AoA obtained at the receiving device (135). Furthermore, the processing circuit is configured to estimate the characteristics of a channel (BS-RIS) between the transmitting device (110) and the configurable surface (120) based on the relationship between the trained reflection coefficient and the estimated reflection coefficient.

[0273] According to a fourteenth aspect, a communication device is provided, comprising: an apparatus for channel estimation according to the thirteenth aspect; and a transceiver.

[0274] According to a fifteenth aspect, a method is provided for setting a reflection coefficient of a configurable surface (120) by a transmitting device (110). The method comprises: performing a beamforming search for signal transmission through a channel (BS-UE) between the transmitting device (110) and a receiving device (135), wherein the channel includes reflections on the configurable surface (120). By performing the beamforming search, i) a trained reflection coefficient of the configurable surface (120), and ii) an angle of arrival AoA of the signal at the receiving device (135) are obtained. In addition, the method comprises: estimating an angle of departure (AoD) of the signal at the configurable surface from the AoA at the receiving device (135). In addition, the method comprises: estimating an estimated reflection coefficient of the configurable surface (120) for an ideal channel (BS-RIS) between the transmitting device (110) and the configurable surface (120) based on the configurable surface (120) and the AoA obtained at the receiving device (135). Furthermore, the method comprises: estimating the characteristics of the channel (BS-RIS) between the transmitting device (110) and the configurable surface (120) based on the relationship between the trained reflection coefficient and the estimated reflection coefficient. ... The reflectance of the configurable surface (120) is set, where:

[0275] ·Θ V′ is a diagonal matrix that includes the trained reflection coefficients on its diagonal,

[0276] ·θ des is the target elevation angle AoD at the configurable surface (120),

[0277] · is the target azimuth AoD at the configurable surface (120),

[0278] · is the predetermined elevation angle AoA at the configurable surface,

[0279] · is the predetermined azimuth angle AoA at the configurable surface,

[0280] · is the elevation angle AoD of the estimated AoD at the configurable surface, and

[0281] · is the range AoD of the estimated AoD at the configurable surface.

[0282] Furthermore, a corresponding method is provided, comprising steps performed by any one of the above-mentioned processing circuit embodiments.

[0283] Furthermore, there is provided a computer program stored on a non-transitory medium and comprising code instructions which, when executed by a computer or by a processing circuit, perform the steps of any one of the above-mentioned methods.

[0284] According to an embodiment, the processing circuit and / or the transceiver are implemented in an integrated circuit IC.

[0285] Although the disclosed subject matter has been described in detail for purposes of illustration based on what are presently considered to be the most practical and preferred embodiments, it should be understood that such detail is solely for that purpose and that the disclosed subject matter is not limited to the disclosed embodiments, but rather is intended to cover modifications and equivalent arrangements within the spirit and scope of the appended claims. For example, it should be understood that the presently disclosed subject matter contemplates that, to the extent possible, one or more features of any embodiment may be combined with one or more features of any other embodiment.

Claims

1. A channel estimation method for estimating, at a receiving device (135), characteristics of a channel between a transmitting device (110) and a receiving device (135), the channel between the transmitting device (110) and the receiving device (135) comprising reflections on a configurable surface (120), the method comprising: A beamforming search is performed for signal transmission over the channel between a transmitting device (110) and a receiving device (135) to obtain: - a trained reflection coefficient of the configurable surface (120), wherein the trained reflection coefficient is a reflection coefficient after a beamforming search, and - Angle of arrival AoA of the signal at the receiving device (135); estimating an estimated reflection coefficient of the configurable surface (120) for an ideal channel between the transmitting device (110) and the configurable surface (120) based on the AoA obtained at the receiving device (135); as well as The phase characteristic of the channel between the transmitting device (110) and the configurable surface (120) is estimated based on the relationship between the trained reflection coefficient and the estimated reflection coefficient.

2. The channel estimation method according to claim 1, wherein: Estimate the estimated reflection coefficient: - Includes the following steps: estimating a departure angle AoD of the signal at the configurable surface from the AoA at the receiving device (135); as well as - Based on the estimated AoD at a configurable surface.

3. The channel estimation method according to claim 2, wherein: The reflection coefficient is estimated according to the following formula, as well as in -φ n,m is the reflection coefficient of the (n, m)th element of the configurable surface, -γ n,m is the reflection gain of the (n, m)th element of the configurable surface, γ n,m is predetermined, -j is an imaginary unit, -λ is the wavelength of the signal, -dx is the spacing between elements of the configurable surface in the x direction, -dy is the spacing between elements of the configurable surface in the y direction, - is the predetermined elevation angle AoA at the configurable surface, - is the predetermined azimuth angle AoA at the configurable surface, - is the height AoD of the estimated AoD at the configurable surface, and - is the azimuth angle AoD of the estimated AoD at the configurable surface.

4. The channel estimation method according to claim 3, wherein The estimated phase characteristic is the modified channel matrix obtained according to the following relationship in: - is a diagonal matrix containing the estimated reflection coefficients on its diagonal, - is the channel matrix of the ideal channel between the transmitting device (110) and the configurable surface (120), where and are the elevation and azimuth angles AoD of the ideal channel at the base station, respectively, and and are the elevation and azimuth angles AoA of the ideal channel at the configurable surface, respectively; -Θ V′ is a diagonal matrix that includes the trained reflection coefficients on its diagonal; and -(Θ V′ ) -1 is a diagonal matrix Θ V′ The inverse matrix of .

5. The channel estimation method according to claim 4, wherein The modified channel matrix a matrix corresponding to a channel matrix G obtained from a channel between a transmitting device (110) and a configurable surface (120) by setting the absolute value of each element of the channel matrix G to 1, and The method further comprises the following steps: Based on the modified channel matrix A modified channel matrix that determines the channel between the configurable surface and the receiving device The power change of the channel matrix G becomes the modified channel matrix part of.

6. The channel estimation method according to claim 5, wherein: Modified channel matrix Determined by an iterative reweighting algorithm.

7. The channel estimation method according to any one of claims 1 to 6, wherein: The trained reflection coefficients are configured by the transmitting device.

8. The channel estimation method according to claim 1, wherein: The beamforming search comprises a hierarchical beamforming search comprising a first stage of performing a beamforming search in a plurality of V layers, wherein for each current layer after a first layer of the plurality of layers: - searching a predetermined number of beams to find the best beam in the current layer, wherein a beam among the searched predetermined number of beams is selected based on the best beam of a previous layer of the current layer; and - Increasing the number of antennas contributing to beamforming in the current layer compared to the previous layer of the current layer.

9. The channel estimation method according to claim 8, wherein: The hierarchical beamforming search also includes a second stage of searching for multiple beams in the vicinity of the beam found after the first stage.

10. A channel estimation method for tracking a position of a receiving device, the method comprising: The channel estimation method according to claim 5 or 6, used for estimating the characteristics of the channel between the transmitting device (110) and the receiving device (135) at the receiving device (135), and Based on the characteristics of the channel between the transmitting device (110) and the receiving device (135) and / or the modified channel matrix The location of the receiving device is tracked.

11. The channel estimation method according to claim 10, used to track the position of the receiving device using extended Kalman filter or least mean square LMS based tracking.

12. A computer program product comprising a non-transitory computer readable medium containing code instructions which, when executed on one or more processors, cause the one or more processors to perform the channel estimation method according to any one of claims 1 to 11.

13. A channel estimation apparatus for estimating, at a receiving device (135), characteristics of a channel between a transmitting device (110) and a receiving device (135), the channel between the transmitting device (110) and the receiving device (135) comprising reflections on a configurable surface (120), the apparatus comprising: The processing circuit is configured to: Controlling a transceiver of the receiving device (135) to perform a beamforming search for signal transmission through the channel between the transmitting device (110) and the receiving device (135), thereby obtaining: - a trained reflection coefficient of the configurable surface (120), wherein the trained reflection coefficient is a reflection coefficient after a beamforming search, and - Angle of arrival AoA of the signal at the receiving device (135); estimating an estimated reflection coefficient of the configurable surface (120) for an ideal channel between the transmitting device (110) and the configurable surface (120) based on the AoA obtained at the receiving device (135); as well as The phase characteristic of the channel between the transmitting device (110) and the configurable surface (120) is estimated based on the relationship between the trained reflection coefficient and the estimated reflection coefficient.

14. A communication device comprising: Apparatus for channel estimation according to claim 13; as well as Transceiver.

15. A method for setting a reflection coefficient, for setting a reflection coefficient of a configurable surface (120) by a transmitting device (110), the method comprising: A beamforming search is performed for signal transmission over a channel between a transmitting device (110) and a receiving device (135), the channel between the transmitting device (110) and the receiving device (135) including reflections on a configurable surface (120), thereby obtaining: - a trained reflection coefficient of the configurable surface (120), wherein the trained reflection coefficient is a reflection coefficient after a beamforming search, and - Angle of arrival AoA of the signal at the receiving device (135); estimating a departure angle AoD of the signal at the configurable surface from the AoA at the receiving device (135); estimating an estimated reflection coefficient of the configurable surface (120) for an ideal channel between the transmitting device (110) and the configurable surface (120) based on the AoA obtained at the receiving device (135); A modified channel matrix for estimating a channel between a transmitting device (110) and a configurable surface (120) based on a relationship between the trained reflection coefficient and the estimated reflection coefficient It only takes into account phase changes caused by the channel between the transmitting device (110) and the configurable surface (120); and The reflectance of the configurable surface (120) is set according to in, - is the channel matrix of an ideal channel between the transmitting device (110) and the configurable surface (120), - and are the elevation and azimuth angles AoD for the ideal channel at the base station, - is a diagonal matrix with the diagonals corresponding to the angles θ des and The reflection coefficient, -θ des is the target elevation angle AoD at the configurable surface (120), - is the target azimuth AoD at the configurable surface (120), - is the predetermined elevation angle AoA at the configurable surface, and - is the predetermined azimuth angle AoA at the configurable surface.

Images