Method and apparatus for interference mitigation combining with reduced complexity

Abstract

This document discloses a scheme for interference-repressed combining. According to one aspect, a method includes obtaining an input stream set, the input stream set being associated with a spatial layer configured for a terminal device; estimating a channel vector h based on a reference signal, the channel vector h representing a radio channel response associated with the spatial layer; computing an interference covariance matrix R, the interference covariance matrix R representing a power of interference from at least one other spatial layer in the input stream set and a correlation of the interference within the input stream set of the spatial layer; performing per-layer interference-repressed combining equalization on the input stream set, including: a) estimating x = R ‑1 h as a combination of linear equations, where a number of the linear equations is defined by an input parameter to be equal to or less than a dimension of the interference covariance matrix; and b) computing an estimate of transmitted symbols based on the channel vector h and the estimated x.

Description

TECHNICAL FIELD

[0001] Various embodiments described herein relate to the field of wireless communications, and more particularly to performing interference rejection combining in a multiple-input multiple-output (MIMO) receiver. BACKGROUND

[0002] Interference rejection combining (IRC) techniques are widely used to mitigate co-channel interference. In particular, IRC reception methods are employed in cellular communication systems. IRC can be applied to multi-beam reception techniques, such as MIMO communications. The benefit of an IRC receiver is that it does not require detailed information about the interfering signals, such as radio channel propagation characteristics. Thus, IRC receivers are simple compared to other receiver architectures. As is known in the art, a characteristic of an IRC receiver is the computation of a covariance matrix, which represents the covariance between the desired signal and the interfering signals. Inversion of the covariance matrix is also a characteristic of an IRC receiver, and this inversion operation is computationally complex.

[0003] XIAO, C, et al. Low-complexity soft-output detection for massive MIMO using SCBiCG and Lanczos methods. China Communications, IEEE [online], December 2015, vol.12, pp.9-17, discloses a minimum mean square error (MMSE) detection scheme utilizing symmetric complex bi-conjugate gradient (SCBiCG) and Lanczos methods for uplink massive MIMO systems.

[0004] QU, H, et al. Efficient Channel Equalization and Symbol Detection for MIMO OTFS Systems. IEEE Transactions on Wireless Communications, IEEE [online], February 2022, vol.21, no.8, pp.6672-6686, discloses a time-space domain channel equalizer for modulation of MIMO on orthogonal time frequency space (OTFS) which relies on a mathematical least squares minimum residual algorithm to remove channel distortion on data symbols. SUMMARY

[0005] Aspects of the application are defined by the independent claims.

[0006] Some embodiments of the invention are defined in the dependent claims.

[0007] Embodiments and features described in this specification, which are not within the scope of the independent claims, are to be interpreted as examples helpful for understanding the various embodiments of the invention. Some aspects of the disclosure are defined by the independent claims.

[0008] According to one aspect, there is provided an apparatus comprising means for performing the following operations: obtaining an input stream set, the input stream set being associated with a spatial layer configured for a terminal device; estimating a channel vector , the channel vector representing a radio channel response associated with the spatial layer; computing an interference covariance matrix , the interference covariance matrix representing a power of interference from at least one other spatial layer in the input stream set and a correlation of the interference within the input stream set of the spatial layer; performing a per-layer interference suppressed combined equalization on the input stream set, comprising:

[0009] a) estimating as a combination of linear equations, wherein a number of linear equations is defined by an input parameter as equal to or smaller than a dimension of the interference covariance matrix; and

[0010] b) computing an estimate of a transmitted symbol based on the channel vector and the estimated

[0011] In an embodiment, the input parameter depends on at least one of: a number of input streams in the input stream set, a modulation and coding scheme of the spatial layer, and a total number of spatial layers configured for all terminal devices scheduled to a same time-frequency resource as the terminal device.

[0012] In an embodiment, the input parameter depends on the number of input streams, the modulation and coding scheme of the spatial layer, and the total number of spatial layers, and wherein a value of the input parameter is proportional to an order of the modulation and coding scheme, proportional to the number of spatial layers, and inversely proportional to the number of input streams.

[0013] In an embodiment, the system of linear equations defines a Krylov subspace representation comprising orthonormal basis vectors defining a Krylov subspace, wherein the input parameter defines a number of the orthonormal basis vectors.

[0014] In embodiments, the Krylov subspace representation further comprises elements of a real-valued tridiagonal matrix, and wherein step b) comprises inverting the real-valued tridiagonal matrix.

[0015] In embodiments, the component is configured to perform the inversion of the real-valued tridiagonal matrix when solving the weights of the orthonormal basis vectors by using the Thomas algorithm, to combine the weights with the corresponding orthonormal basis vectors and the initial estimate, and to compute the estimate of the transmitted symbol based on said combination.

[0016] In embodiments, the component is configured to perform the inversion of the real-valued tridiagonal matrix when solving the weights of the orthonormal basis vectors by computing an eigenvalue decomposition of the tridiagonal matrix, to compute the weights by taking the inverse of the eigenvalues of the eigenvalue decomposition, to combine the weights with the corresponding orthonormal basis vectors and the initial estimate, and to compute the estimate of the transmitted symbol based on said combination.

[0017] In embodiments, the component is configured to use the initial estimate in step b), wherein. / denotes an element-wise division operation.

[0018] In embodiments, the component is configured to generate the Krylov subspace representation by using the Lanczos algorithm.

[0019] In embodiments, the component is configured to interpolate the estimated or derived from parameters to time-frequency resources that do not carry a reference signal.

[0020] In embodiments, the component is configured to average a determined number of channel estimates, wherein the determined number is a function of a modulation and coding scheme associated with the spatial layer.

[0021] In embodiments, the determined number is smaller for a first modulation scheme than for a second modulation scheme, wherein the first modulation scheme maps a larger number of bits per symbol than the second modulation scheme.

[0022] In embodiments, the component comprises at least one processor and at least one memory including computer program code, the at least one memory and the computer program code configured to, with the at least one processor, cause the performance of the apparatus.

[0023] According to an aspect, there is provided a method comprising: obtaining a set of input streams, the set of input streams being associated with a spatial layer configured for a terminal device; estimating a channel vector based on a reference signal representing a radio channel response associated with the spatial layer; computing an interference covariance matrix , the interference covariance matrix representing a power of interference from at least one other spatial layer in the input stream group and a correlation of the interference within the input stream group of the spatial layer; performing a per-layer interference suppressed combining equalization on the input stream group, comprising:

[0024] a) estimating the as a combination of linear equations, wherein a number of linear equations is defined by an input parameter as equal to or smaller than a dimension of the interference covariance matrix; and

[0025] b) computing an estimate of the transmitted symbols based on the channel vector and the estimated in step a).

[0026] In embodiments, the input parameter depends on at least one of a number of input streams in the input stream group, a modulation and coding scheme of the spatial layer, and a total number of spatial layers configured for all terminal devices scheduled to the same time-frequency resources as the terminal device.

[0027] In embodiments, the input parameter depends on the number of input streams, the modulation and coding scheme of the spatial layer, and the total number of spatial layers, and wherein a value of the input parameter is proportional to an order of the modulation and coding scheme, proportional to the number of spatial layers, and inversely proportional to the number of input streams.

[0028] In embodiments, the system of linear equations defines a Krylov subspace representation comprising orthonormal basis vectors defining a Krylov subspace, wherein the input parameter defines a number of the orthonormal basis vectors.

[0029] In embodiments, the Krylov subspace representation further comprises elements of a real-valued tri-diagonal matrix, and wherein step b) comprises an inversion of the real-valued tri-diagonal matrix.

[0030] In embodiments, the method comprises, when solving the weights of the orthonormal basis vectors by using the Thomas algorithm, performing the inversion of the real-valued tri-diagonal matrix, combining the weights with the corresponding orthonormal basis vectors and the initial estimate, and computing the estimate of the transmitted symbols based on the combination.

[0031] In embodiments, the method comprises, when solving the weights of the orthonormal basis vectors by computing an eigenvalue decomposition of the tri-diagonal matrix, performing the inversion of the real-valued tri-diagonal matrix, computing the weights by inverting eigenvalues of the eigenvalue decomposition, combining the weights with the corresponding orthonormal basis vectors and the initial estimate, and computing the estimate of the transmitted symbols based on the combination.

[0032] In embodiments, the method comprises using the initial estimate in step b) where. denotes element-wise division.

[0033] In an embodiment, the method comprises generating the Kronecker subspace representation by using the Lanzos algorithm.

[0034] In an embodiment, the method comprises interpolating the estimated or derived from parameters to time-frequency resources not carrying reference signals.

[0035] In an embodiment, the method comprises averaging a determined number of channel estimates, wherein the determined number is a function of a modulation and coding scheme associated with the spatial layer.

[0036] In an embodiment, the determined number is smaller for a first modulation scheme than for a second modulation scheme, wherein the first modulation scheme maps a larger number of bits per symbol than the second modulation scheme.

[0037] According to an aspect, there is provided a computer program product embodied on a computer readable medium and comprising computer readable program code which, when executed by a computer, configures the computer to perform a computer process comprising: obtaining an input stream set, the input stream set being associated with a spatial layer configured for a terminal device; estimating a channel vector based on a reference signal, the channel vector representing a radio channel response associated with the spatial layer;

[0038] computing an interference covariance matrix , the interference covariance matrix representing a power of interference from at least one other spatial layer in the input stream set and a correlation of the interference within the input stream set of the spatial layer; performing per-layer interference suppressed combined equalization on the input stream set, comprising:

[0039] a) estimating as a combination of linear equations, wherein a number of linear equations is defined by an input parameter to be equal to or smaller than a dimension of the interference covariance matrix; and

[0040] b) computing an estimate of a transmitted symbol based on the channel vector and the estimated BRIEF DESCRIPTION OF DRAWINGS

[0041] By way of example only, example embodiments of the present application will be described herein below with reference to the accompanying drawings, in which:

[0042] Figure 1 a wireless communication scenario to which some embodiments of the present application can be applied is illustrated;

[0043] Figure 2 The illustration shows a process for calculating a combined solution for interference suppression on a received signal according to an embodiment;

[0044] Figure 3 The diagram illustrates the architecture of a receiver according to some embodiments;

[0045] Figure 4 The diagram illustrates the detailed steps of computing the interference suppression combinatorial solution using Krylov subspace representation; and

[0046] Figure 5 The figure shows a structural block diagram of the device according to an embodiment. Detailed Implementation

[0047] The following embodiments serve as examples. Although the specification may refer to "an," "one," or "some" embodiments in some places, this does not necessarily mean that every such reference refers to the same embodiment, or that the feature applies only to a single embodiment. Individual features of different embodiments may also be combined to provide other embodiments. Furthermore, the words "comprising" and "including" should be understood to mean that the embodiments are not limited to consisting only of those mentioned features, and such embodiments may also include features / structures not specifically mentioned.

[0048] In the following description, radio access architectures based on Advanced Long Term Evolution (LTE-A, LTE-A) or New Radio (NR, 5G) will be used as examples of access architectures to which embodiments can be applied to describe different exemplary embodiments; however, the embodiments are not limited to this architecture. Those skilled in the art will recognize that, by appropriately adjusting parameters and procedures, the embodiments can also be applied to other types of communication networks with suitable components. Some examples of other options for suitable systems include: Universal Mobile Telecommunications System (UMTS) Radio Access Network (UTRAN or E-UTRAN), Long Term Evolution (LTE, the same as E-UTRA), Wireless Local Area Network (WLAN or WiFi), Global Microwave Interconnection Access (WiMAX), Bluetooth®, Personal Communication Services (PCS), ZigBee®, Wideband Code Division Multiple Access (WCDMA), systems using Ultra Wideband (UWB) technology, sensor networks, Mobile Ad Hoc Networks (MANETs), and Internet Protocol Multimedia Subsystem (IMS), or any combination thereof.

[0049] Figure 1 An example of a simplified system architecture is depicted, showing only some components and functional entities, which are logical units whose implementations may differ from those shown. Figure 1The shown connections are logical connections; the actual physical connections can be different. It will be apparent to a person skilled in the art that the system typically also comprises Figure 1 other functions and structures than those shown.

[0050] However, the embodiments are not limited to the system given as an example, but a person skilled in the art can apply the solution to other communication systems provided with the necessary characteristics.

[0051] Figure 1 An example shows a part of an example radio access network.

[0052] Figure 1 Terminal devices or user equipments 100 and 102 are shown, which are configured to wirelessly connect with an access node, such as an (e / g)NodeB 104, providing a cell on one or more communication channels in the cell. The (e / g)NodeB refers to an eNodeB or a gNodeB as defined in the 3GPP specifications. The physical link from the user equipment to the (e / g)NodeB is called uplink or reverse link, and the physical link from the (e / g)NodeB to the user equipment is called downlink or forward link. It should be understood that the (e / g)NodeBs or their functionalities can be implemented by using any node, host, server or access point etc. suitable for such a use.

[0053] The communication system typically includes more than one (e / g)NodeB, in which case the (e / g)NodeBs can also be configured to communicate with one another, e.g., by means of a (wired or wireless) link designed for the purpose. These links can be used for signalling purposes or for routing data from one (e / g)NodeB to another (e / g)NodeB. The (e / g)NodeB is a computing device configured to control the radio resources of a communication system it is coupled to. The (e / g)NodeB can also be referred to as a base station, an access point, an access node, or any other type of interfacing device including a relay station capable of operating in a wireless environment. The (e / g)NodeB includes or is coupled to a transceiver. From the transceiver of the (e / g)NodeB, a connection is provided to an antenna unit that establishes bi-directional radio links to the user equipments. The antenna unit can include multiple antennas or antenna elements. The (e / g)NodeB is also connected to a core network 110 (CN or Next Generation Core NGC). Depending on the system, the counterpart on the CN side can be a serving gateway (S-GW, routing and forwarding user data packets), a packet data network gateway (P-GW) for providing a connection to a packet data network, or a mobile management entity (MME), etc.

[0054] A user equipment (also known as UE, user device, user terminal, terminal device, etc.) illustrates one type of apparatus for which resources on the air interface are allocated and assigned, and thus any features described herein with a user equipment can be implemented with a corresponding apparatus, such as a relay node. An example of such a relay node is a layer 3 relay towards a base station (self-backhauled relay). The 5G specification supports at least the following relay operation modes: out-band relaying, where different carriers and / or RATs (radio access technologies) can be defined for the access link and the backhaul link; and in-band relaying, where the same carrier frequency or radio resources are used for the access link and the backhaul link. In-band relaying can be considered the baseline relay scenario. The relay node is referred to as an integrated access and backhaul (IAB) node. It also has built-in support for multiple relay hops. The IAB operation assumes a so-called split architecture with a CU and multiple DUs. The IAB node contains two independent functions: the DU (distributed unit) part of the IAB node facilitates the gNB (access node) functionality in the relay cell, i.e. as an access link; and the mobile terminal (MT) part of the IAB node, which facilitates the backhaul connection. The donor node (DU part) communicates with the MT part of the IAB node, and it has a wired connection to the CU, which in turn has a connection to the core network. In a multi-hop scenario, the MT part (child IAB node) communicates with the DU part of the parent IAB node.

[0055] A user equipment typically refers to a portable computing device, such as a wireless mobile communication device operating with or without a subscriber identification module (SIM), including, but not limited to, the following types of devices: a mobile station (mobile phone), a smartphone, a personal digital assistant (PDA), a handset, a device using a wireless modem (alarm or measurement device, etc.), a laptop and / or touch screen computer, a tablet, a game console, a notebook, and a multimedia device. It should be appreciated that a user equipment can also be a nearly exclusive uplink only device, an example of which is a camera or video camera that loads images or video clips to a network. A user equipment can also be a device that has the capability to operate in the Internet of Things (IoT) network, which is a scenario for objects to provide capabilities of transferring data over a network without requiring human-to-human or human-to-computer interaction. A user equipment can also utilize the cloud. In some applications, a user equipment can include a small portable device with a radio part (such as a watch, earphones or glasses) and the computing is performed in the cloud. A user equipment (or in some embodiments a layer 3 relay node) is configured to perform one or more user equipment functions. A user equipment can also be known as a subscriber unit, a remote terminal, an access terminal, a user terminal, or user device (UE), to mention just a few names or apparatuses.

[0056] The various techniques described herein can also be applied to cyber-physical systems (CPS: systems that cooperatively interact with the physical world through the use of computing devices that control physical entities). CPS can enable the implementation and exploitation of a large number of interconnected ICT devices (sensors, actuators, processors microcontrollers, etc.) embedded in physical objects in different locations. Mobile information-physical systems are a subcategory of information-physical systems in which the physical systems involved have intrinsic mobility. Examples of mobile physical systems include mobile robots and electronic devices transported by people or animals.

[0057] In addition, although the apparatus is depicted as a single entity, different units, processors and / or memory units (not all shown in the drawings) can be implemented. Figure 1

[0058] 5G enables the use of multiple input-multiple output (MIMO) antennas, many more base stations or nodes than LTE (the so-called small cell concept), including macro stations operating in co-operation with smaller stations and the use of various radio technologies according to the service requirements, use cases and / or available frequency spectrum. 5G mobile communications supports a wide range of use cases and related applications, including video streaming, augmented reality, different ways of sharing digital content, and various forms of machine type applications (such as (massive) machine type communications (mMTC), including vehicle safety, different sensors, and real-time control). 5G is expected to have multiple radio interfaces, i.e. below 6 GHz, cmWave, and mmWave, and also to be able to integrate with the existing legacy radio access technologies, such as LTE. At least in the early phase, it is possible to implement the system as an integration with LTE, where macro coverage is provided by LTE and 5G radio interface access comes from small cells by aggregation to LTE. In other words, 5G is planned to support inter-RAT operability (e.g. LTE-5G) and inter-RI operability (inter-radio interface operability, such as below 6 GHz-cmWave, below or at 6 GHz-cmWave-mmWave). One of the concepts considered to be used in 5G networks is network slicing, in which multiple independent and dedicated virtual sub-networks (network instances) can be created within the same infrastructure to run services that have different requirements on latency, reliability, throughput, and mobility.

[0059] ​The current architecture in LTE networks is fully distributed in the radio and fully centralized in the core network. The low latency applications and services requirements in 5G drive the need to bring the content closer to the radio, which leads to local break out and multi-access edge computing (MEC). 5G enables analytics and knowledge generation to occur at the source of the data. This approach requires balancing resources that can be disconnected from the network, such as laptops, smartphones, tablets, and sensors. MEC provides a distributed computing environment for application and service hosting. It also has the ability to store and process content close to cellular subscribers for faster response time. Edge computing encompasses a wide range of technologies, such as wireless sensor networks, mobile data acquisition, mobile signature analysis, cooperative distributed peer-to-peer ad hoc networking and processing, and can also be classified as local cloud / fog computing and grid / mesh computing, dew computing, mobile edge computing, cloudlet, distributed data storage and retrieval, self-healing networks, remote cloud services, augmented and virtual reality, data caching, Internet of Things (massive connectivity and / or latency critical), critical communications (autonomous vehicles, traffic safety, real-time analytics, time-critical control, healthcare applications).

[0060] The communication system is also able to communicate with other networks 112, such as a public switched telephone network or the Internet, or utilize services provided by them. The communication system can also be able to support the usage of cloud services, wherein e.g. at least part of the core network operations can be carried out as a cloud service (this is depicted in Figure 1 by "cloud" 114). The communication system can further include a central control entity, etc., providing facilities for networks of different operators to cooperate, e.g. in spectrum sharing.

[0061] By utilizing network function virtualization (NFV) and software defined networking (SDN), edge cloud can be brought into the radio access network (RAN). Using edge cloud can mean that node operations will be executed, at least partly, in a server, host or node that is operationally coupled to a remote radio head or a base station that comprises a radio part. It is also possible that node operations will be distributed among a plurality of servers, nodes or hosts. The application of cloud RAN architecture enables RAN real-time functions to be executed in the RAN side (in the distributed unit, DU 105) and non-real-time functions to be executed in a centralized manner (in the centralized unit, CU 108).

[0062] It should also be understood that the distribution of functions between core network operations and base station operations can differ from that of LTE, or even be non-existent. Some other technological advancements that can be used are big data and all-IP, which can change the way networks are constructed and managed. 5G (or New Radio, NR) networks are designed to support multiple hierarchies, where MEC servers can be placed between the core and the base station or NodeB (gNB). It should be understood that MEC can also be applied to 4G networks.

[0063] 5G can also utilize satellite communication to enhance or complement the coverage of 5G services, for example by providing backhauling. Possible use cases are providing service continuity for machine-to-machine (M2M) or Internet of Things (IoT) devices or for passengers on board of vehicles, or ensuring service availability for critical communications, as well as for future railway, maritime and / or aeronautical communications. Satellite communication can utilize geosynchronous earth orbit (GEO) satellite systems, but also low earth orbit (LEO) satellite systems, particularly mega-constellations (systems in which hundreds of (nano)satellites are deployed). Each satellite 109 in a mega-constellation can cover several network entities supporting the satellite, which create ground cells. Ground cells can be created through ground relay nodes or through gNBs located on the ground or in a satellite.

[0064] It will be apparent to those skilled in the art that the described system is merely an example of a part of a radio access system, and in practice the system can comprise a plurality of (e / g)NodeBs, the user equipment can have access to a plurality of radio cells, and the system can also comprise other apparatuses, such as physical layer relay nodes or other network elements. At least one of the (e / g)NodeBs can be a home (e / g)NodeB. In addition, within the geographical area of the radio communication system, a plurality of radio cells of different kinds can be provided as well as a plurality of radio cells. The radio cells can be macro cells (or umbrella cells), which are large size cells, typically with a diameter of up to several tens of kilometers, or the radio cells can also be smaller in size, such as micro cells, femto cells, or pico cells. Figure 1 The (e / g)NodeBs can provide any kind of these cells. The cellular radio system can be implemented as a multi-layer network comprising a plurality of cells. Typically, in a multi-layer network one access node provides one or more cells of the same kind, and thus a plurality of (e / g)NodeBs is needed to provide such a network structure.

[0065] To meet the need to improve the deployment and performance of communication systems, the concept of "plug and play" (e / g)NodeB has been introduced. Typically, a network capable of using "plug and play" (e / g)NodeB includes, in addition to home (e / g)NodeB (H(e / g)NodeB), a home NodeB gateway, or HNB-GW (not shown in the figure). The HNB gateway (HNB-GW), typically installed within the operator's network, can aggregate traffic from a large number of HNBs back to the core network. Figure 1

[0066] As described in the background, it would be beneficial to reduce the computational complexity of an Interference Rejection Combining (IRC) equalization receiver. IRC receivers are conventionally applied to receivers, such as terminal device 100 and access node 104, having multiple antennas and multiple spatial layers configured between the transmitter and the receiver. The receiver receives multiple beams or input streams from the transmitter, one stream per receiver antenna element. The multiple beams or input streams can together form one or more spatial layers configured between the transmitter and the receiver. As is known in the art, spatial layers are generated at the transmitter and receiver via MIMO processes and antenna arrays. To effectively utilize beamforming and achieve spatial layers, multiple antenna elements and corresponding beams can be configured per spatial layer in the MIMO processing. Each spatial layer and its corresponding input stream can then be processed at a time by treating the signals from the other spatial layers as interference. The MIMO receiver process can include receive beamforming, where an input stream set for each spatial layer is extracted from the signals received by the antenna elements. An interference covariance matrix, mentioned in the background, is then computed to represent the interference power in the stream set and the correlation of the interference in the input stream set for the spatial layer for which IRC is performed, and the receiver computes an IRC scheme aimed at reducing the interference. The IRC scheme is typically aimed at whitening the interference, which involves computing the inverse of the covariance matrix. This computation is very complex, and it would be beneficial to reduce the complexity of the IRC receiver. A conventional IRC equalization scheme is defined as:

[0067] (1)

[0068] where H defines a channel matrix containing channel estimates for all antennas or input streams (and spatial layers), N defines the number of antennas or, equivalently, the number of beams or input streams, Ns defines the number of spatial layers configured for the terminal device in case of single-user MIMO (SU-MIMO), and in case of multi-user MIMO (MU-MIMO) Ns,denotes the total number of spatial layers over all scheduled terminal devices, R defines the interference covariance matrix, and​ The received data samples are defined. The interference covariance matrix can contain noise elements such as additive white Gaussian noise and can also be referred to as interference plus noise covariance matrix. The output of the IRC equalization scheme is a vector of estimates of the transmitted symbols .

[0069] In the embodiments described below, the IRC scheme for multiple spatial layers and multiple input streams is split into a single layer IRC (SL-IRC) equalization scheme. The idea is to compute the IRC equalization per spatial layer, such that the dimension and computational complexity of the IRC equalization is reduced. This simplifies the representation of equation (1) to the following form:

[0070] (2)

[0071] where defines the channel vector containing the estimates of all antennas or beams or input streams for a specific spatial layer, and defines the interference (plus noise) covariance matrix of the spatial layer under study. Accordingly, the channel matrix is simplified to the channel vector and the output of the SL-IRC scheme is an estimate of the transmitted symbols for the specific layer instead of an estimate of the symbol vector. If layer-specific beamforming has been applied, the received sample vector can also be specific to the spatial layer. The interference covariance matrix is also simplified to a matrix representing the interference power and interference correlation between the input streams of the spatial layer under SL-IRC processing. In a regular IRC scheme, the computationally heavy part is the estimation of This requires techniques to first make the covariance matrix estimate invertible, and then to invert the matrix. The whitening-based IRC approach allows to avoid the explicit inversion of by using a Cholesky decomposition as whitening matrix. Although this approach can significantly reduce the process complexity, it is still too complex to enable a power-efficient, low-latency SL-IRC equalization in the receiver.

[0072] The key to note in equation (2) is that we are interested in solving from the linear system of equations Because is Hermitian, we also have and we obtain the second term of the SL-IRC shown in equation (2) above. Since it is a single layer process, and are simplified to scalars, and we can use approximations rather than exact solutions (see below for details on approximations). This problem formulation guides us to solve for the IRC equalizer solution through the Krylov subspace, which helps us avoid Figure 2 FIGURE 1 illustrates an embodiment of a process for estimating an IRC equalizer solution using a system of linear equations. The process can be performed in a receiver for a terminal device 100 or for an access node or for another wireless receiver. The process can be used for apparatus execution, such as a chipset or at least one processor with at least one memory, of such a receiver.

[0073] Referring to FIGURE 1 Figure 2 The process includes obtaining (block 200) a set of input streams associated with a spatial layer configured for the terminal device, estimating (block 202) a channel vector representing a radio channel response associated with the spatial layer based on a reference signal, computing (block 202) an interference covariance matrix representing an interference power from at least one other spatial layer in the set of input streams and a correlation of intra-set interference for the spatial layer, and performing per-layer interference-repressed combined equalization for the set of input streams, including:

[0074] The estimation is a combination of linear equations, where the number of linear equations is defined by an input parameter to be equal to or less than a dimension of the interference covariance matrix, and

[0075] The estimate of the transmitted symbols is computed based on the channel vector and the estimated

[0076] After block 206, a determination can be made as to whether there is another spatial layer for which an IRC solution has not yet been estimated. For example, while a set of input streams for one spatial layer is executed from block 200 to 206, a second set of input streams for another spatial layer that is co-scheduled to the same time-frequency resources as the first spatial layer can be executed from block 200 to 206.

[0077] In an embodiment, the system of linear equations defines a Krylov subspace representation that includes an orthonormal basis vector defining the Krylov subspace, where the input parameter defines a number of the orthonormal basis vectors.

[0078] ​Using the Krylov subspace representation or another representation based on linear equations in the estimation of the transmitted symbols, the inversion of the covariance matrix can be avoided. This greatly reduces the computational complexity of the IRC equalization. For example, this reduction depends on the total number of input streams per spatial layer, the total number of jointly scheduled spatial layers, and the modulation and coding scheme, but can achieve 93 percent (%) less complex multiplications than the conventional scheme based on equation (1), and 89% less complex multiplications than the scheme using whitening-based IRC. With the equal dimension of the linear equations and the dimension of the interference covariance matrix, the advantage in computational complexity can be obtained. When the dimension of the linear equations (for the Krylov subspace, the parameter m described below) is less than the dimension of the interference covariance matrix, further reduction in computational complexity can be obtained.

[0079] As is well known in the art, a plurality of wireless antenna streams or beams is typically associated with each spatial layer, e.g., a first spatial layer and a second spatial layer. Thus, each spatial layer can include a plurality of input streams, which define the dimension of the interference covariance matrix and the length of the channel vector .

[0080] The following description focuses on the Krylov subspace for defining the linear equations. However, there are other subspaces or algorithms for solving the same problem, e.g., the biconjugate transpose method.

[0081] Let us now examine the use of the Krylov subspace representation in the IRC scheme estimation more closely. The embodiments described below make use of the Minimum Residual Method (MINRES), the Lanczos algorithm, and the Thomas algorithm as taught in the literature. The following description provides disclosure for adapting these well-known algorithms to IRC equalization. The MINRES method is a method of approximative solution by vectors in the Krylov subspace with low residual. The starting point is that we want to solve from which is represented in the IRC scheme as . In other words, in the following symbol = and = . To this end, the m-th order Krylov subspace is defined as

[0082]

[0083] where is the initial error given by our initial guess . The span function follows its regular mathematical definition. Because the vector are linearly dependent, so for example the Lanzos algorithm can be used to find an orthonormal basis . The algorithm described herein uses a vector to approach the exact solution, which vector provides an estimation error that is small enough. One aspect of reducing computational complexity is to choose the number of basis vectors to be small enough while achieving a small enough estimation error . One aspect of defining a small enough estimation error is that the bit error rate or block error rate performance of the wireless link between the transmitter and receiver is not compromised. Another aspect is that we can trade off link performance and power consumption, for example, in the UE receiver we can use a small value to reduce power consumption in low power scenarios. We can rewrite the vector where is the initial guess, is the approximation in , and is a vector of linear combination weights of the orthonormal basis vectors.

[0084] The initial guess can be set to zero, but in another embodiment, each element of the channel vector is partitioned into the corresponding diagonal element. It has been found that this initialization can provide improved performance, particularly in the case of SL-IRC. As described above, the number of orthonormal basis vectors defining the Krylov subspace representation is a parameter that affects computational complexity. As described above in connection with Figure 2 , the value of m is chosen to define the dimension of the Krylov subspace to be less than the dimension of the interference covariance matrix . In practice, this means that m < Nr, where Nr is the number of input streams in the spatial layers for which SL-IRC estimation is being performed. Embodiments for choosing m are described below. The same principles apply to other equivalent methods (e.g., the double-conjugate transpose method) for solving from .

[0085] The Krylov representation or the equivalent orthonormal Krylov subspace can be created by using the Lanzos algorithm described in more detail in the literature. The matrix containing the orthonormal basis vectors of the Krylov subspace is obtained as output. In addition, the Lanzos algorithm outputs the real-valued vectors and corresponding to the symmetric tridiagonal matrix The main diagonal ( ) and the first sub-diagonal ( () elements. The first subdiagonal is the same, as is known in conjunction with the Lanzos algorithm. The Lanzos algorithm belongs to a class of power methods used to find eigenvalues ​​and orthogonal eigenvectors, and other embodiments may employ another power method. The Arnoldi iterative method is another possible algorithm.

[0086] Following the MINRES method, the next task is to solve... But first we must solve it.

[0087] ,

[0088] in Vector norm, and It is a unit vector.

[0089] Through attention For tridiagonal structures, the Thomas algorithm can be used to directly solve them. On this point, we can make an approximation: . It can be configured to have only real values ​​without affecting link performance. This is due to the use of the Thomas algorithm. For finite-precision arithmetic, complex values ​​may arise, but these values ​​may be forced to be represented only using real values. This makes it possible to apply the Thomas algorithm only to real values, thereby reducing computational complexity.

[0090] Then, calculate as well as Then, the IRC scheme is solved directly using the following equation (2). Using... The received signal and channel vector Based on this knowledge, we can calculate the estimate of the transmitted symbols. It then outputs the signal to further processes in the receiver. These further processes may include, for example, demodulation and decoding.

[0091] Figure 3 The diagram illustrates the receiver process chain or equivalent architecture for calculating the IRC scheme as described above. (Reference) Figure 3 Channel vector The covariance matrix can be stored in buffer 300. The received data symbol vector can be stored in buffer 320. The channel vectors can be averaged in block 302, and the number of channel estimates to be averaged can be adaptive and a function of determined parameters, e.g. the modulation and coding scheme (MCS) applied to the data symbols. In other embodiments, such averaging is omitted.

[0092] The Kronecker subspace basis vectors are generated in block 304, and the number of basis vectors is defined by the input parameter m. As mentioned above, m can be less than the dimension of the covariance matrix, e.g. the number of columns of the covariance matrix In embodiments, m depends on at least one of the number of input streams Nr in the input stream group of spatial layers for which IRC processing is performed, the MCS of the spatial layers for which interference suppression combining equalization is performed, and the total number of spatial layers configured for the terminal device and optionally for other terminal devices in the same time-frequency resource. An access node can schedule the same time-frequency resource but different spatial layers to multiple terminal devices. In embodiments of the first or third aspect, m is a function of a number of these parameters, or even of all of these parameters as described below. Figure 2

[0093] In embodiments, depends on the number of input streams in the input stream group, the modulation and coding scheme of the spatial layers for which interference suppression combining equalization is performed, and the number of spatial layers configured for the terminal device, and wherein the value of the input parameter is proportional to the order of the modulation and coding scheme, proportional to the number of spatial layers, and inversely proportional to the number of input streams. The number of spatial layers can include all spatial layers associated with the same time-frequency resource, e.g. they can include the spatial layers configured to the terminal device, but they can also include one or more spatial layers scheduled to one or more other terminal devices. The logic is that the smaller the number of spatial layers, the less the interference, and thus the smaller the Kronecker subspace that is able to provide sufficient performance. The same principle applies to other subspaces or other linear equation systems, e.g. the double conjugate transpose method. The following table illustrates some embodiments of the dependence of m on each of these parameters.

[0094]

[0095]

[0096]

[0097] ​With respect to the modulation and coding scheme, QPSK means quadrature phase shift keying, and QAM means quadrature amplitude modulation. The number in front of QAM means the number of symbols in the symbol constellation, as is well known in the art. In case the value of m is selected using at least one of the number of spatial layers, the number of input streams, and the modulation and coding scheme, the number of spatial layers can be used to define an initial value of m, which can then be adapted according to the modulation and coding scheme and / or the number of input streams, using the above logic. The smaller the number of input streams, the larger the initial value can be, and the larger the number of input streams, the smaller the initial value can be. For lower order modulation and coding schemes, such as QPSK or 16-QAM, the initial value can be smaller, while for higher order modulation schemes, such as 64-QAM and 256-QAM, the initial value can be kept or even increased. Since the number of spatial layers is common to all spatial layers that are subject to IRC processing, this number is the same for all spatial layers. However, if the spatial layers are configured with different numbers of input streams Nr and different modulation and coding schemes, different variations of the value of m can be introduced for the spatial layers.

[0098] The dependence of m on the MCS can be extended to a dependence of m on the code rate of the spatial layer. The same logic as for the MCS can be applied, which means that a smaller m can be selected if the code rate is small (below a threshold), and a larger m can be selected for larger code rates (above a threshold). In case more than two values of m can be assumed, multiple thresholds can be used, but the logic is maintained that the value of m is proportional to the value of the code rate.

[0099] As mentioned above, the Krivonosov subspace representation output from block 304 comprises elements of the real-valued tri-diagonal matrix and where step b) comprises the inversion of the real-valued tri-diagonal matrix

[0100]

[0101] As mentioned above, in order to solve for the weights of the orthogonal basis vectors the inversion of the matrix is required. Due to being real-valued and having a reduced dimension, the complexity is much lower than the inversion of a complex-valued interference covariance matrix. In an embodiment, the inversion of the real-valued tri-diagonal matrix is performed when solving for the weights of the orthogonal basis vectors by using the Thomas algorithm. An alternative is to use an eigenvalue decomposition on e.g. the matrix and represent the inverse of by using a specific number of eigenvectors and the inverse of the eigenvalues to construct the estimate on . As is well known in the art, the Thomas algorithm is derived from H ​​m z = d d As described above. In the case where is solved as After solving the weights of the orthogonal basis vectors defining the Krylov subspace representation, we can solve as described above (block 306). From equation (2) and Figure 3 The Hermitian transpose can be assumed to be real valued to reduce memory consumption and to reduce the computational complexity in the following interpolation steps. Then, in blocks 310 and 312 time and frequency domain interpolation is performed, respectively. The purpose of the interpolation is to estimate the values of (or equivalently ) and for all subcarriers and all time domain signals (or samples) carrying data symbols As is well known in the art, the demodulation reference symbols (DMRS) are only transmitted on some (not all) subcarriers and / or time domain symbols. Therefore, the estimation of is only applied to those subcarriers and time domain symbols and can be interpolated to the other subcarriers and time domain symbols in blocks 310 and 312. In addition to the above, one advantage of this architecture is that the computation of the Lanzos algorithm, the (tri-diagonal) matrix inversion and is performed before the interpolation blocks. As a result, the number of subspace processes calls and the number of samples on which the inversion is subsequently performed, and the number of vector products needed to solve are much smaller than in the case where the inversion is performed after the interpolation. As a result, a low computational complexity can be achieved.

[0102] Again from equation (2), the Hermitian transpose of the estimate of is subsequently multiplied with the received signal samples in block 314. Thereafter, the remaining operations of equation (2) are performed in block 316, resulting in the IRC estimate of the transmitted symbols.

[0103] Figure 4 A detailed flow chart of this procedure is illustrated in Figure 3 and let us refer to Figure 4 ​​​​Some further embodiments are disclosed. After blocks 200 and 202, or in parallel with them, at least some parameters of the algorithm are initialized in block 400. The parameters initialized in block 400 can include at least one of the following parameters: the dimension m of the Krein subspace, the initial estimate , and the average parameter input into block 302.

[0104] As mentioned above, the initial estimate may be initialized to zero or where denotes the element-wise division operation.

[0105] In embodiments, the average parameter is a function of the modulation and coding scheme, similar to m. The dependence of the average parameter on the modulation and coding scheme can follow the logic where the parameter value for a first modulation scheme is smaller than the parameter value for a second modulation scheme, where the first modulation scheme maps a larger number of bits per symbol than the second modulation scheme. In other words, the average parameter for higher order modulation schemes, such as 64-QAM or 256-QAM, is smaller than the average parameter for lower order modulation schemes, such as QPSK or 16-QAM. The following table illustrates an embodiment of this dependence. The same principle applies to other subspace or other linear equation systems (e.g., the bi-conjugate transpose method).

[0106]

[0107] By including layer-specific code rate information into the MCS table, a more precise control of the value of the average parameter can be achieved. For example, if the code rate for 64-QAM is greater than 0.8, the average parameter value is set to .

[0108] The averaging can be performed over channel estimates per physical resource block (PRB), i.e., the averaging can be performed over channel estimates in the frequency domain within a PRB. As is well known in the art, a PRB can comprise a certain number of frequency resource elements (e.g., subcarriers). As shown in the table, when the modulation and coding scheme is a low order modulation and coding scheme, such as QPSK, channel estimates can be averaged. In embodiments, all channel estimates for a PRB are averaged in case of QPSK. For medium order modulation and coding schemes, such as 16-QAM and 64-QAM, channel estimates can be averaged, and for high order modulation and coding schemes, such as 256-QAM, the averaging can be omitted. Accordingly, the computational complexity can be reduced while maintaining an acceptable performance. As mentioned above and as shown in Figure 3 , the averaging can be performed before the calculation of .

[0109] The parameters to be initialized can depend on the embodiment, e.g. whether the above-mentioned averaging of channel estimates is performed or not. In embodiments not supporting averaging of channel estimates, the respective initialization of the averaging parameter can naturally be omitted. Similarly, if the initial estimate is fixed, e.g. then the initialization can not require a calculation and simplifies to a memory retrieval.

[0110] In block 402, Krylov subspace orthonormal basis vectors are generated for the spatial layer under processing based on the channel vector and the interference covariance matrix, e.g. by using the Lanczos method or another power method (cf. block 304). Then, weights for the basis vectors can be calculated, e.g. by using the Thomas algorithm. Thereafter, the (cf. block 306) can be solved in block 406, and block 406 can comprise combining the orthonormal basis vectors with the respective weights. Then, a final SL-IRC estimate can be calculated in block 408, wherein the weights are combined with the respective orthonormal basis vectors and the initial estimate. Block 408 can be repeated (via block 409) for further symbols of the spatial layer. As mentioned above, the interpolation extends the estimate of or the parameters (e.g. the ) derived from to further time-frequency resources (subcarriers and / or time-domain symbols). As a result, an IRC estimate of the value of a data symbol transmitted on the same spatial layer and on the respective time-frequency resources can be calculated by repeating block 408. In block 410, it is determined whether there are still spatial layers to be processed. If so, the process returns to block 402, wherein the next spatial layer and the respective input stream are taken into account for processing. If all spatial layers have been processed, the process can end.

[0111] It is to be understood that the order of processing the symbols within a spatial layer and processing the spatial layers (the arrangement of blocks 409 and 410) can be different. For example, different spatial layers can even be processed in parallel by different processing circuits.

[0112] Figure 5 Fig. illustrates an apparatus according to the present application, comprising processing circuitry 50, such as at least one processor, and at least one memory 60 including computer program code (software) 64, wherein the at least one memory and the computer program code (software) are configured to, with the at least one processor, cause the apparatus to perform Figure 2The apparatus can be used in a terminal device 100 or for an access node 104, e.g. for a DU or a CU. The apparatus can be a circuitry or electronics implementing some embodiments of the application in a terminal device or an access node. Thus, the apparatus performing the above described functions can be comprised in such a device, e.g. the apparatus can comprise a circuitry (such as a chip, a chipset, a processor, a microcontroller) for a terminal device or an access node or a combination of such circuitries. In other embodiments, the apparatus is generally for a radio device, e.g. a radio device or a circuitry in a radio device or a circuitry designed to operate in a radio device.

[0113] The memory 60 can be implemented using any suitable data storage technology, e.g. semiconductor-based memory devices, flash memory, magnetic memory devices and systems, optical memory devices and systems, fixed memory and removable memory.

[0114] The processing circuitry 50 can comprise a SL-IRC processing circuitry 52 configured to perform IRC estimation on a received input stream according to any one of the above described embodiments. The SL-IRC processing circuitry 52 can comprise a linear equation generating circuitry 54 configured to generate basis vectors and corresponding weights of basis vectors or linear equations to approximate from the linear equations received from the circuitry 54. The SL-IRC circuitry can further comprise SL-IRC scheme estimation circuitry configured to compute a SL-IRC scheme by using the subspace representation or the linear equations received from the circuitry 54. The SL-IRC circuitry can further comprise at least some of the other components or functions of the architecture shown in Figure 3 and described in the above embodiments.

[0115] In embodiments, the apparatus further comprises a wireless transceiver 62 having a plurality of antenna elements for receiving an input stream for IRC processing. The wireless transceiver 62 can further comprise other conventional radio receiving components, such as filters, amplifiers, frequency converters and baseband signal processing components and functions.

[0116] As used in this application, the term "circuitry" refers to all of the following: (a) hardware-only circuitry such as comprises only analog and / or digital circuitry; (b) a combination of circuits and software and / or firmware, such as (as applicable): (i) a combination of processor(s) or processor cores; or (ii) portions of processing- software and at least one memory that work together to cause an apparatus to perform various functions described herein; and (c) circuits, such as a microprocessor(s) or a portion of a microprocessor, that require software or firmware for operation, even if the software or firmware is not physically present.

[0117] This definition of "circuitry" applies to all uses of this term in this application. As a further example, as used in this application, the term "circuitry" would also cover an implementation in which the Figure 3 Or the processes and methods described in any one embodiment can also be at least partially processed by one or more computers or processors. For example, the processes and methods can be at least partially implemented using hardware, software, firmware, "wetware" (e.g., human beings), or combinations of them. As another example, the processes and methods can be at least partially implemented using one or more computers or processors that are distributed across one or more networks, such as over a satellite network, over the Internet, or using a combination of both. As a further example, if the processes and methods described here can be at least partially implemented using one or more computers or processors, then the one or more computers or processors can each also be configured with individual elements of the different features to perform the processes and / or methods. These and other examples of the processes and methods described herein can be performed using one or more computers or processors that are distributed across one or more networks.

[0118] Figures 2 to 4The processes or methods described herein, or any embodiments thereof, may also be performed as one or more computer processes defined by one or more computer programs. A computer program may be in the form of source code, object code, or some intermediate form, and it may be stored in some kind of carrier, which may be any entity or device capable of carrying the program. Such carriers include transient and / or non-transient computer media, such as recording media, computer memory, read-only memory, electrical carrier signals, telecommunication signals, and software distribution packages. Depending on the required processing power, the computer program may be executed in a single electronic digital processing unit or may be distributed across multiple processing units. References to computer-readable program code, computer programs, computer instructions, computer code, etc., should be understood to refer to software for a programmable processor, such as programmable content stored in a hardware device as instructions for a processor, or as a configured or configurable setup for a fixed-function device, gate array, or programmable logic device.

[0119] The embodiments described herein apply to the wireless networks defined above, but also to other wireless networks. The protocols used, the specifications of wireless networks and their elements are evolving rapidly. Such evolution may necessitate additional modifications to the described embodiments. Therefore, all words and expressions should be interpreted broadly, and they are intended to illustrate rather than limit the embodiments. It will be apparent to those skilled in the art that the concepts of the present invention can be implemented in various ways as technology advances. Embodiments are not limited to the examples described above, but may vary within the scope of the claims.

Claims

1. A computer-implemented apparatus for a radio receiver, comprising means for performing the following operations: obtaining an input stream set, the input stream set being associated with spatial layers configured for a terminal device; estimating a channel vector based on a reference signal , the channel vector is indicative of a radio channel response associated with the spatial layer Computing an interference covariance matrix , the interference covariance matrix representing a power of interference from at least one other spatial layer in the input stream group and a correlation of the interference within the input stream group of the spatial layer; performing per-layer interference suppressed combining equalization on the input stream set, comprising: a) calculating are estimated as a combination of linear equations, wherein the number of linear equations is defined by the input parameters to be equal to or smaller than the dimension of the interference covariance matrix; and b) computing an estimate of the transmitted symbols based on said channel vector and said , computing an estimate of the transmitted symbols.

2. The apparatus according to claim 1, wherein the input parameter depends on at least one of a number of input streams in the input stream set, modulation and coding schemes of the spatial layers, and a total number of spatial layers configured for all terminal devices scheduled to the same time-frequency resources as the terminal device.

3. The apparatus according to claim 2, wherein the input parameter depends on the number of input streams, the modulation and coding schemes of the spatial layers, and the total number of spatial layers, and wherein a value of the input parameter is proportional to an order of the modulation and coding schemes, proportional to the number of spatial layers, and inversely proportional to the number of input streams.

4. The apparatus according to any preceding claim, wherein the system of linear equations defines a Krylov subspace representation comprising orthonormal basis vectors, the orthonormal basis vectors defining the Krylov subspace, wherein the input parameter defines the number of the orthonormal basis vectors.

5. The apparatus according to claim 4, wherein the Krylov subspace representation further comprises elements of a real-valued tri-diagonal matrix, and wherein step b) comprises inversion of the real-valued tri-diagonal matrix.

6. The apparatus according to claim 5, wherein the means are configured to perform the inversion of the real-valued tri-diagonal matrix when solving weights of the orthonormal basis vectors by using a Thomas algorithm, to combine the weights with respective the orthonormal basis vectors and an initial estimate, and to compute the estimate of the transmitted symbols based on the combination.

7. The apparatus according to claim 5, wherein the means are configured to perform the inversion of the real-valued tri-diagonal matrix when solving weights of the orthonormal basis vectors by computing an eigenvalue decomposition of the tri-diagonal matrix, to compute the weights by inverting eigenvalues of the eigenvalue decomposition, to combine the weights with respective the orthonormal basis vectors and an initial estimate, and to compute the estimate of the transmitted symbols based on the combination.

8. The apparatus of claim 6, wherein the means are configured to use the initial estimate in step b) where / denotes element-wise division operation.

9. The apparatus according to claim 4, wherein the means are configured to generate the Krylov subspace representation by using a Lanczos algorithm.

10. The apparatus of any of claims 1 to 3, wherein the means are configured to interpolate the estimated parameters of the or derived from to time-frequency resources that do not carry the reference signal.

11. The apparatus according to any of claims 1 to 3, wherein the means are configured to average a determined number of channel estimates, wherein the determined number is a function of a modulation and coding scheme associated with the spatial layers.

12. The apparatus according to claim 11, wherein the determined number is smaller for a first modulation scheme than for a second modulation scheme, wherein the first modulation scheme maps a larger number of bits per symbol than the second modulation scheme.

13. The apparatus of any one of claims 1-3, wherein the component comprises: at least one memory including computer program code; 14. A computer-implemented method for a radio receiver, comprising: obtaining an input stream set, the input stream set being associated with spatial layers configured for a terminal device; estimating a channel vector based on a reference signal , the channel vector is indicative of a radio channel response associated with the spatial layer Computing an interference covariance matrix , the interference covariance matrix representing a power of interference from at least one other spatial layer in the input stream group and a correlation of the interference within the input stream group of the spatial layer; performing per-layer interference suppressed combining equalization on the input stream set, comprising: a) calculating are estimated as a combination of linear equations, where the number of linear equations is defined by the input parameters to be equal to or smaller than the dimension of the interference covariance matrix; and b) computing an estimate of the transmitted symbols based on said channel vector and said , computing an estimate of the transmitted symbols.

15. A computer program product, embodied on a computer readable medium, and comprising computer-readable computer program code, wherein the computer program code configures the computer to perform a computer process, comprising: obtaining an input stream set, the input stream set being associated with spatial layers configured for a terminal device; estimating a channel vector based on a reference signal , the channel vector is indicative of a radio channel response associated with the spatial layer Computing an interference covariance matrix , the interference covariance matrix representing a power of interference from at least one other spatial layer in the input stream group and a correlation of the interference within the input stream group of the spatial layer; performing per-layer interference suppressed combining equalization on the input stream set, comprising: a) calculating are estimated as a combination of linear equations, where the number of linear equations is defined by the input parameters to be equal to or smaller than the dimension of the interference covariance matrix; and 16. A computer program product, embodied on a computer readable medium, and comprising computer-readable computer program code, wherein the computer program code configures the computer to perform a computer process, comprising: obtaining an input stream set, the input stream set being associated with spatial layers configured for a terminal device; performing per-layer interference suppressed combining equalization on the input stream set, comprising: and 17. A computer program product, embodied on a computer readable medium, and comprising computer-readable computer program code, wherein the computer program code configures the computer to perform a computer process, comprising: obtaining an input stream set, the input stream set being associated with spatial layers configured for a terminal device; performing per-layer interference suppressed combining equalization on the input stream set, comprising: and 18. A computer program product, embodied on a computer readable medium, and comprising computer-readable computer program code, wherein the computer program code configures the computer to perform a computer process, comprising: obtaining an input stream set, the input stream set being associated with spatial layers configured for a terminal device; performing per-layer interference suppressed combining equalization on the input stream set, comprising: and 19. A computer program product, embodied on a computer readable medium, and comprising computer-readable computer b) computing an estimate of the transmitted symbols based on said channel vector and said , computing an estimate of the transmitted symbols based on said channel vector and said

Images