This application claims priority under 35 U.S.C. §119 to European Patent Application No. 12162634.5 filed on Mar. 30, 2012, the entire content of which is hereby incorporated by reference.
Embodiments according to the invention are related to a transmission filter calculator, a method for calculating a transmission filter, a communication device, a method for operating a communication device and to computer programs.
An embodiment according to the invention is related to an adaptive unitary precoding for multicell MIMO communications.
Embodiments according to the invention are useable in the field of wireless communications. Further embodiments according to the invention are related to the field of transmission technology. Some embodiments according to the invention are related to the field of cooperative multi-point (CoMP) transmission. Some embodiments according to the invention are related to the field of orthogonal beamforming. Some embodiments according to the invention are related to a MIMO downlink.
In the following, some application scenarios which occur in current communication networks like, for example, mobile communication networks, will be described. In particular, a target configuration in which embodiments according to the invention can be advantageously applied will be described.
In many communication environments there are multiple users in a multicell/cooperative multi-point (CoMP) system.
One of the challenges in such a system is the downlink transmission in the multiuser multicell/CoMP system. In this case, there is a so-called “interference network”. For example, there is a very significant interference level due to the presence of multiple base stations (and possibly also multiple mobile stations). Also, advanced multiple-input-multiple-output transmission is often applied in such systems, which increases the complexity. Furthermore, there is often unknown inter-cell interferences, which also makes it more difficult to choose proper communication parameters.
A graphical representation of a typical communication environment, in which embodiments according to the present invention may be applied, is shown in
In view of such communication environments, it is a goal to reduce uncertainty in interference to obtain more robust downlink strategies.
Moreover, it should be noted that cell-edge users experience inter-cell interference in many modern communication environments. Inter-cell interference of mobile communication devices which are in the proximity of a cell-edge are shown in
It has been found that for MIMO transmission, the spatial signature of the interference is typically unknown. It has also been found that this causes problems for link adaptation and algorithmic impairments in higher layers (like, for example, scheduler and resource allocation, for example fractional reuse).
In view of this situation, it is a goal to have robust MIMO downlink strategies and to solve the above mentioned problems.
In the following, some alternative solutions will be described.
For example, a so-called “codebook-based MIMO” concept may be used. An example of such a concept is known as “PU2RC”. In this concept, a unitary codebook makes interference well-predictable. However, it has been found that said concept does not exploit the full capabilities a of multiuser multiple-input-multiple-output approach.
Another alternative solution is the so-called “advanced multiuser MIMO” (also briefly designated as “MU-MIMO”). This concept uses an adaptive precoding. Moreover, low complexity solutions are available like, for example, a solution called “SESAM” (see, for example, reference [B3]) and a solution called “LISA” (see, for example, reference [B4]). In this advanced multiuser MIMO, robustness is typically achieved either by a second pilot (which constitutes an additional overhead) or by conservative rate adaptation (which brings along a performance loss). For details, reference is made to reference [B1]).
However, it has been found that codebook based approaches do not exploit the full potential of MU-MIMO. Moreover, it has been found that robustness for adaptive MIMO causes additional overhead in alternative implementations. Also, it has been recognized that it is difficult to find optimal receivers, which has a huge impact on performance. Thus, it can be seen that alternative solutions comprise a number of problems and limitations.
In the following, some recent advances will be briefly summarized.
Several concepts for adaptive unitary precoding have been found. For example, references is made to documents [B6] to [B10]. The concepts use adaptive precoding and it has been found that interference is well predictable. However, an optimal solution is not available for linear precoding. Also, it has been found that it is difficult to optimize (for example, because the optimization problem is non-convex and combinatorial). Moreover, it has been found that low complexity solutions are only available for single antenna receivers.
In view of the above discussion, it is an object of the present invention to create an efficient concept for communicating in a multicell multi-input-multi-output communication environment.
An embodiment may have a transmission filter calculator for calculating at least a transmit filter for a communication from a transmit communication device including a plurality of antennas to a receive communication device out of a plurality of receive communication devices including a plurality of antennas via a multiple-input-multiple-output channel, wherein the transmission filter calculator is configured to iteratively determine transmit filters for a plurality of data streams, wherein the transmission filter calculator is configured to choose a transmit filter in a given iteration such that the transmit filter chosen in the given iteration is orthogonal to one or more transmit filters chosen in one or more previous iterations, and such that a weighted or unweighted transmission rate obtainable using the transmit filter chosen in the given iteration is maximized, and wherein the transmission filter calculator is configured to choose the transmit filter in dependence on a matrix product term which is based on a channel matrix describing MIMO channel characteristics between the transmit communication device and a currently considered one of the receive communication devices and on a distortion covariance matrix describing a noise, an inter-cell interference and an upper limit of an intra-cell interference.
Another embodiment may have a method for calculating at least a transmit filter for a communication from a transmit communication device including a plurality of antennas to a receive communication device out of a plurality of receive communication devices including a plurality of antennas via a multiple-input-multiple-output channel, wherein the method includes iteratively determining transmit filters for a plurality of data streams, wherein a transmit filter is chosen in a given iteration such that the transmit filter chosen in the given iteration is orthogonal to one or more transmit filters chosen in one or more previous iterations, and such that a weighted or unweighted transmission rate obtainable using the transmit filter chosen in the given iteration is maximized, and wherein the transmit filter is chosen in dependence on a matrix product term which is based on a channel matrix describing MIMO channel characteristics between the transmit communication device and a currently considered one of the receive communication devices and on a distortion covariance matrix describing a noise, an inter-cell interference and an upper limit of an intra-cell interference.
Another embodiment may have a communication device for communicating with a transmit communication device in the presence of one or more interferer communication devices, wherein the communication device is configured to determine a channel matrix describing MIMO channel characteristics between the communication device and the transmit communication device, and to also determine one or more channel matrices describing MIMO channel characteristics between the communication device and the one or more interferer communication devices, and wherein the communication device is configured to determine a matrix product which is based on the channel matrix describing a MIMO channel between the communication device and the transmit communication device and on a distortion covariance matrix describing a noise, an inter-cell interference and an upper limit of an intra-cell interference, to obtain a product matrix, and wherein the communication device is configured to transmit the product matrix to the transmit communication device.
Another embodiment may have a method for operating a communication device for communicating with a transmit communication device in the presence of one or more interferer communication devices, wherein the method includes determining a channel matrix describing MIMO channel characteristics between the communication device and the transmit communication device, determining one or more channel matrices describing MIMO channel characteristics between the communication device and the one or more interferer communication devices, and wherein the method includes determining a matrix product which is based on the channel matrix describing a MIMO channel between the communication device and the transmit communication device and on a distortion covariance matrix describing a noise, an inter-cell interference and an upper limit of an intra-cell interference, to obtain a product matrix, and wherein the method includes transmitting the product matrix to the transmit communication device.
Another embodiment may have a computer program for performing the inventive methods when the computer program runs on a computer.
An embodiment according to the invention creates a transmission filter calculator for calculating at least a transmit filter for a communication from a transmit communication device comprising a plurality of antennas to a receive communication device out of a plurality of receive communication devices comprising a plurality of antennas via a multiple-input-multiple-output (MIMO) channel. The transmission filter calculator is configured to iteratively determine transmit filters for a plurality of data streams, wherein the transmission filter calculator is configured to choose a transmit filter in a given iteration such that the transmit filter chosen in the given iteration is orthogonal to one or more transmit filters chosen in one or more previous iterations, and such that a weighted or unweighted transmission rate obtainable using the transmit filter chosen in the given iteration is maximized. Also, the transmission filter calculator is configured to choose the transmit filter in dependence on a matrix product term which is based on a channel matrix describing MIMO channel characteristics between the transmit communication device and a currently considered one of the receive communication devices and which is also based on a distortion covariance matrix describing a noise, an inter-cell interference and an upper limit of an intra-cell interference.
It is a key idea of the present invention that the usage of a matrix product term, which is based on the channel matrix describing MIMO channel characteristics between the transmit communication device and a currently considered one of the receive communication devices and which is based on a distortion covariance matrix describing a noise, an inter-cell interference and an upper limit of an intra-cell interference in a transmission filter calculator which ensures that a transmit filter chosen in a given iteration is orthogonal to one or more transmit filters chosen in one or more previous iterations allows for an efficient low-complexity iterative algorithm for a calculation of transmission filters, wherein the resulting transmit filters (which are described by transmit filter vectors) result in a good usage of the available resources.
It has been found that this concept allows for an extension of unitary precoding to multiple receive antennas with moderate complexity. Moreover, it has been found that no additional feedback is necessitated using the above described concept. Also, it has been found that an optimal, or at least almost optimal, rate adaptation is possible since the distortion covariance matrix describing a noise, an inter-cell interference and an upper limit of an intra-cell interference is very well-suited for this purpose and at the same time allows for a computationally efficient formulation of the algorithm for the transmit filter calculation. Moreover, the concept described above also allows for a computationally efficient provision of optimal (or almost optimal) receive filters.
To conclude, the usage of the matrix product term, as described above, in an iterative calculation of transmit filters allows for an efficient algorithm which nevertheless brings along very good results which cause a good usage of the radio resources.
In an embodiment, the transmission filter calculator is configured to choose the transmit filter in a given iteration using a projector matrix, and to update the projector matrix in response to a choice of a transmit filter, such that Eigenvectors associated with non-zero Eigenvalues of the updated projector matrix are orthogonal to transmit filters (or, more precisely, to transmit filter vectors describing transmit filters) chosen in previous iterations and in the current iteration. It has been found that the application of a projector matrix in a iterative algorithm allows for a particularly efficient choice of orthogonal transmit filters (or, more precisely, transmit filter vectors) while avoiding the efficiency restrictions implied by a usage of a predetermined set of orthogonal transmit filters. In other words, usage of the projector matrix allows to choose the best transmit filter vector (describing a transmit filter) in each iteration (under the restriction that the chosen transmit filter vector has to be orthogonal to previously chosen transmit filter vectors).
In an embodiment, the transmission filter calculator is configured to initialize the projector matrix to be an identity matrix, or a multiple thereof, prior to an iterative determination of transmit filters for a plurality of data streams. The transmission filter calculator is configured to subtract a matrix defined by an outer product of a chosen transmit filter vector and a transposed version of the chosen transmit filter vector (describing the chosen transmit filter) from the projector matrix (used in the current iteration), to obtain the updated projector matrix (to be used in the next iteration). Accordingly, a particularly efficient mechanism is created for ensuring the orthogonality of transmit filter vectors chosen in subsequent iterations.
In an embodiment, the transmission filter calculator is configured to determine the weighted transmission rate associated with a currently considered one of the receive communication devices in dependence on a matrix product term which is based on the projector matrix of a current iteration, the channel matrix associated with the currently considered one of the receive communication devices and the distortion covariance matrix associated with the currently considered one of the receive communication devices. This embodiment is based on the finding that such a matrix product term, which is based on the projector matrix of the current iteration, the channel matrix associated with the currently considered one of the receive communication devices and the distortion covariance matrix associated with the currently considered one of the receive communication devices, allows for a simple derivation of a transmission rate.
In an embodiment, the transmission filter calculator is configured to determine a largest Eigenvalue of the matrix product term which is based on the projector matrix of the current iteration, the channel matrix associated with the currently considered one of the receive communication devices and the distortion covariance matrix associated with the currently considered one of the receive communication devices, in order to determine the weighted or unweighted transmission rate obtainable for the currently considered one of the receive communication devices. Moreover, the transmission filter calculator is configured to determine a selected receive communication device on the basis of a determination of weighted or unweighted transmission rates obtainable for a plurality of receive communication devices. Moreover, the transmission filter calculator is configured to determine an Eigenvector of the matrix product term based on the projector matrix of the current iteration, the channel matrix associated with the currently considered one of the receive communication devices and the distortion covariance matrix associated with the currently considered one of the receive communication devices, in order to obtain the transmit filter (or transmit filter vector) associated with a data stream to the receive communication device selected in the current iteration. It has been found that the evaluation of both the Eigenvalues and the Eigenvectors of said matrix product term allows for a computationally efficient way of iteratively selecting orthogonal transmit filters (or orthogonal transmit filter vectors describing said transmit filters) while at the same time making good usage of the available channel resources and avoiding a large waste of resources.
In an embodiment, the transmission filter calculator is configured to compare, in an iteration, maximum weighted or unweighted transmission rates obtainable for a plurality of receive communication devices in view of the condition that the transmit filter to be chosen in the current iteration is orthogonal to all transmit filters chosen in one or more previous iterations, and to select, in an iteration, a receive communication device and a corresponding transmit filter for a data stream. A simultaneous choice of a receive communication device and a corresponding transmit filter for a data stream towards the chosen receive communication device provides for a high efficiency of the algorithm.
In an embodiment, the transmission filter calculator is configured to obtain the maximum weighted or unweighted transmission rates obtainable for the plurality of receive communication devices on the basis of respective largest Eigenvalues of matrix product terms based on a projector matrix of the current iteration, a channel matrix associated with the currently considered one of the receive communication devices and the distortion covariance matrix associated with the currently considered one of the receive communication devices. It has been found that this procedure is particularly efficient since the determination of largest Eigenvalues is possible using computationally efficient algorithms.
In an embodiment, the transmission filter calculator is configured to use, as the distortion covariance matrix associated with a given one of the receive communication devices, a distortion covariance matrix which is based on noise information describing a noise effective for the given one of the receive communication devices, based on a channel matrix describing MIMO channel characteristics between the given one of the receive communication devices and the transmit communication device, and based on one or more channel matrices describing MIMO channel characteristics between the given one of the receive communication devices and one or more interference communication devices. It has been found that such a choice of the distortion covariance matrix allows for a consideration of different types of distortion, namely a noise, an inter-cell interference and an intra-cell interference, within a single matrix. Also, it has been found that such a choice of a distortion covariance matrix allows for the provision of transmit filters (or transmit filter vectors) which are very well adapted to the communication environment.
In an embodiment, the transmission filter calculator is configured to use, as the matrix product term for choosing the transmit filter, a matrix product term received from a currently considered one of the receive communication devices by the transmit communication device (which comprises the transmission filter calculator). It has been found that the matrix product term can already be computed at the side of a receive communication device, which makes the transmission of the necessitated information from the receive communication device to the transmit communication device resource-efficient.
In an embodiment, the transmission filter calculator is configured to determine receive filters for a plurality of data streams. The transmission filter calculator is configured to choose a receive filter in a given iteration on the basis of the distortion covariance matrix associated with the receive communication device selected in the given iteration, on the basis of the channel matrix associated with the receive communication device selected in the given iteration and in dependence on the transmit filter chosen in the given iteration. Accordingly, there is a simple computational rule for the determination of the receive filter, which again improves the efficiency of the algorithm.
An embodiment according to the invention creates a method for calculating at least a transmit filter for a communication from a transmit communication device to a receive communication device. The method is based on the same ideas as the transmission filter calculator.
An embodiment according to the invention creates a communication device (for example, a receive communication device, a mobile communication device, or the like) for communicating with a transmit communication device in the presence of one or more interferer communication devices. The communication device is configured to determine a channel matrix describing MIMO channel characteristics between the communication device and the transmit communication device, and to also determine one or more channel matrices describing MIMO channel characteristics between the communication device and the one or more interferer communication devices. The communication device is also configured to determine a matrix product which is based on the channel matrix describing a MIMO channel between the communication device and the transmit communication device and on a distortion covariance matrix describing a noise, an inter-cell interference and an upper limit of an intra-cell interference, to obtain a product matrix. The communication device is configured to transmit the product matrix to the transmit communication device.
This communication device is based on the finding that it is particularly efficient to transmit the matrix product, as defined above, from the communication device (for example, receive communication device, mobile communication device, or the like) to the transmit communication device. In other words, it has been found that it is not necessitated to transmit each individual information item (like, for example, the individual channel matrices or the individual distortion covariance matrix) from the communication device (for example, receive communication device, mobile communication device, or the like) to the transmit communication device. Rather, it has been found that the combined information is sufficient for the transmit communication device to obtain a transmit filter information (like, for example, a transmit filter vector describing a transmit filter) which brings along an efficient usage of the radio resources.
In an embodiment, the communication device (for example, receive communication device, mobile communication device, or the like) is configured to obtain the distortion covariance matrix on the basis of a noise information describing a noise effective for the communication device (for example, receive communication device, mobile communication device, or the like), on the basis of the channel matrix describing the MIMO channel characteristics between the communication device (for example, receive communication device, mobile communication device, or the like) and the transmit communication device, and on the basis of the one or more channel matrices describing the MIMO channel characteristics between the communication device (for example, receive communication device, mobile communication device, or the like) and the one or more interferer communication devices. It has been found that the derivation of an appropriate distortion covariance matrix can be done on the basis of information which can be obtained by a communication device like, for example, a mobile communication device or the like, with relatively small effort.
An embodiment according to the invention creates a method for operating a communication device. Said method is based on the same ideas as the communication device discussed above.
Other embodiments according to the invention create a computer program for performing one of the above discussed methods when the computer program runs on a computer.
Embodiments of the present invention will be detailed subsequently referring to the appended drawings, in which:
A brief introduction into the problem underlying the embodiments according to the invention will be provided and alternative solutions will be discussed. Subsequently, a detailed discussion of embodiments according to the invention will be provided and simulation results will be provided. Finally, conclusions will be given.
The scenario described in the following may be a scenario in which a communication device comprising the transmission filter calculator according to the present invention may be operated. Moreover, the communication devices defined herein may operate in the scenario described in the following.
It should be noted here that the description sometimes distinguishes between “transmitters” (also designated as “transmit communication devices”) and receivers (also designated as “receive communication devices” or “user communication devices” or briefly “communication devices”). However, it should be noted that even a device designated as a transmitter or a transmit communication device may naturally also comprise some reception functionality. In many cases, a transmitter or a transmit communication device is part of a transceiver or a transceiver communication device. Similarly, a receiver or a receive communication device may be part of a respective transceiver communication device. Accordingly, the designation “transmit communication device” and “receive communication device” is primarily used to describe a current role of the respective communication device, wherein however the roles may change. However, a “transmit communication device” may typically be configured to establish communication with a plurality of “receive communication devices”. Accordingly, in some embodiments a “transmit communication device” may be a base station of a communication network, or the like, and a “receive communication device” may be a mobile station of a communication network.
In wireless networks the quality of the received signal, and therefore the data rate of the user, depends on the transmit strategy chosen, the properties of the wireless channel, noise level, and interference. The interdependence of the data rates of the users due to interference and limited resources makes it difficult to optimize the physical layer parameters, which is especially challenging in MIMO networks, as the availability of additional degrees of freedom is directly reflected in the spatial signature of the interference.
In general, the exact transmit covariance matrix of the other transmitters cannot be predicted correctly in advance as the optimal transmission strategies mutually depend on each other. Here we regard the case were users are served while being affected by inter-cell interference and assume that the set of interfering transmitters I is given.
We do not regard further interference coordination among the involved transmitters and therefore each transmitter (for example, a transmit communication device comprising the transmit filter calculator according to the present invention) chooses the best transmit strategy for its users egoistically.
For networks where each transmitter is equipped with a single antenna it is clear that the strategy of each transmitter is to use all available power, which makes the interference caused to each user very well predictable, for example by measurements. As in the single antenna case, for multiple transmit antennas each transmitter will use its full power budget, however the spatial signature of the interference is difficult to predict. The noise plus inter-cell interference covariance matrix CkεN
where σ2I is the receiver noise covariance, QtεN
N
An alternative approach is to use a prediction of the inter-cell interference based on either measured inter-channel-interference ICI in the past [2], or based on knowledge of the cross-channels.
It has been found that in case the users are able to gather correct channel state information on the channels to the interfering transmitters, a prediction that performs well is to assume a white transmit covariance using the full power budget:
The predicted interference plus noise covariance is then given by
In Section 1.1, it will be discussed how MIMO downlink strategies can be parameterized, that is the mapping of physical layer parameters to data rates. In Section 1.2., algorithms to select transmission strategies in order to maximize a weighted sum of the data rates (WSR) will be briefly discussed. Further, it will be discussed how the incorrect prediction of the interference affects the performance. To reduce the uncertainty in the prediction we can enforce a white transmit covariance at every transmitter. Several known strategies that result in a white transmit covariance matrix will be discussed in Section 1.2. Finally, alternative methods will be summarized and the relationship to the embodiments according to the invention will be discussed.
In the following, some assumptions and definitions underlying embodiments according to the invention will be described.
As described in the previous Section, we assume that the interdependence of the data rates due to interference is considered by a prediction of the interference. This decouples the problem of finding optimal downlink strategies into a problem per transmitter, in the following we therefore omit the affiliation to a specific transmitter and regard a single cell, where the set of users is K, K=|K|.
The transmitter (which may be considered as a communication device or transmit communication device, and which may comprise a transmission filter calculator) has Ntx antennas and the number of receive antennas of user k is Nrx,k. The receiver of user k may be considered as a “communication device” or “user communication device” or “receive communication device”. For clearer notation, but without loss of generality, we assume that all users (or all receive communication devices, or all user communication devices) have the same number of receive antennas Nrx,k=Nrx∀kεK. The downlink channel matrices are H1, . . . , HKεN
Gaussian input signals and DPC are capacity achieving for the Gaussian multiple input multiple output (MIMO) broadcast channel [4], therefore the downlink strategies are fully described by the Ntx×Ntx downlink transmit covariance matrices Q1, . . . , QK and the encoding order. The covariances are necessitated to be positive semidefinite, which we denote as Qk≧0∀kεK. Assuming the users are sorted according to the encoding order, the data rate of user k is given by
In case of linear precoding the data rate of user k is given by
For users equipped with a single antenna (MISO), Nrx=1, channel coefficients are usually noted as column vectors. Instead of Hkε1×N
N
and for linear precoding the rate expression (1.3) becomes
Any rate configuration achievable by the transmit covariance parameterization, see Section 1.1.1, can alternatively be achieved by transmitting data streams over scalar channels that are established by transmit and receive filters. Low-complexity solutions often target to find an efficient data stream set directly. The filters are chosen as normalized vectors, i.e. udHud=1 and vdHvd=1, so each stream d is described by the assignment to a user sdεK, receive filter udεN
N
+. We define a set of data streams D=(S, U, V, p), D=|D|, where the set of active users is given by S={s1, . . . , sD}, the set of receive filters is U={u1, . . . , uD}, the transmit filters are collected in a matrix V=[v1, . . . , vD]ε
N
+D.
For DPC the rate of a data stream d can be calculated as
where we assumed that the data streams are sorted according to the encoding order. For linear precoding, the data rate is
For single antenna receivers we have
for the case of DPC and for linear precoding the rate is
Note that some relevant downlink strategies investigated are designed for receivers with a single receive antenna. We describe these algorithms for single antenna receivers, for the numerical simulations we extend these results to handle multiple receive antennas.
It should be noted that a transmit filter vd (or, more precisely, a transmit filter described by a transmit filter vector vd) associated to a data stream describes how a signal representing the data stream d is assigned to a plurality of antennas. Each entry of the transmit filter vector vd describes how a signal representing the data stream is assigned to a single antenna out of a plurality of antennas.
Usually the weighted sum rate optimization is subject to a power constraint, that is
for the parameterization in transmit covariances, see Section 1.1.1, and
for the data stream formulation, see Section 1.1.2.
In the following, some alternative algorithms will be described to facilitate the understanding.
In case of DPC the global optimal transmit covariances can be efficiently computed via BCMAC Duality [4, 5], for both multiple receive antennas (MIMO) and single receive antennas (MISO). For linear precoding and MIMO the global optimal solution is an open problem. For MISO case a global optimal solution can be computed by methods for global optimization [6, 7]. However, these methods have high complexity and are not applicable for a larger number of users or online implementation. An approach for local optimal solutions, for both MIMO and MISO, can be found in [8].
SESAM [9, 10] and LISA [11, 12, 13] are low complexity algorithms that directly select the data streams, where transmit and receiver filters are chosen to completely eliminate intra-cell interference (zero-forcing). In the following, we briefly describe the algorithms, for details we refer to [9, 10] and [11, 12, 13]. Both algorithms successively allocate data streams under the condition that the newly added stream does not interfere with the previous data streams, this implies orthogonal transmit filters. This orthogonality is assured by a projector matrix Π, which is also successively updated. For SESAM the remaining interference is canceled by DPC. For LISA an additional zero-forcing step to update the transmit filters (the receive filters are kept fixed) is necessitated. The optimal power allocation over the now orthogonal scalar channels is found by waterfilling. SESAM will allocate the maximal number of data streams, usually Ntx, as adding new streams does not decrease the weighted sum-rate as the gains of scalar channels of the previous data streams are not altered. This is different for the LISA algorithm, where the zero-forcing step affects all channels, which could lead to a decrease of the weighted sum-rate. Therefore, after a new data stream is added a check if the performance is increased is performed and in case of a decrease, meaning
w
T
r(D(d))<wTr(D(d−1)). (1.8)
the algorithm is interrupted and the latest data stream is discarded. SESAM is summarized in Algorithm 1, which is represented in
It has been found that one possibility to reduce the negative effects of uncertainty in the interference is to constrain the transmission strategies such that the interference caused to other cells is constant. For example one could restrain to a single precoder, or necessitate the covariance of the transmitted signal to have a certain structure, for example a scaled identity matrix. This means to remove the power constraint and force a white transmit covariance
that uses the full power budget
This reduces the uncertainty in interference to the uncertainty in the cross-channels. In case the cross-channels are perfectly known, the prediction (1.1) is correct. The weighted sum-rate optimization under a white transmit covariance constraint is:
As a remark, it should be noted that the problem is feasible, for example by setting
for any k, this is what is called stabilization in [2]. An algorithm to compute a global optimal solution to the WSR optimization with white transmit covariance constraint can be derived as a special case from the results in [14]. But, the high computational complexity, for DPC and finding a solution, prevent a practical implementation.
A practical approach is the use of unitary precoders, for example Per-User Unitary Rate Control (PU2RC) [15]. Using the stream based formulation, unitary (or orthogonal) precoding, means to select precoders such that V is unitary, that is VVH=I. In case every transmitter serves exactly Ntx datastreams and allocates equal power among them,
the transmit covariance is
In the following, some algorithms using unitary precoding will be described. wherein the algorithms described the following assume linear precoding.
It has been found that unitary precoding is attractive as the intra-cell interference Cintra-cell,d of stream d, and therefore the achievable data rate, only depends on the own precoder vd,
In the above,
may represent an upper limit of intra-cell interference.
For single antenna receivers (remember hk=HkT) this is
This means that for PU2RC with a fixed unitary precoder codebook, every user can decide for the codebook entry vd that maximizes his data rate, which is given by
Advanced downlink strategies with adaptive precoding are also available for unitary precoding. In case the user set S is given, such that |S|=Ntx, the weighted sum-rate optimization is
which is unfortunately a non-convex problem. A global optimal algorithm is only known for the two user case [16]. Algorithms to find a local optimal solution can be found in [17, 18]. The user set has to be determined by an additional user selection scheme, which motivates algorithms that find the user set and precoders jointly, see [19, 20] for MISO and [21] for MIMO. These algorithms are all designed for sum-rate maximization, but in all cases an extension to WSR is possible.
In the following, some algorithms will be described in a consistent (and in some cases more sophisticated) notation and to allow for a comparison in the relevant scenarios the extension to weighted sum-rate will be shown. Additionally, in case not already present in the original work, a stopping criteria is included, see (1.8), to interrupt the successive allocation if the new configuration has a lower weighted sum-rate.
In the following, some algorithms will be described which are somewhat similar to the algorithms in reference [19]. In [19] the precoder v1 of the first data stream is chosen in the direction of the users channel
In this case the first stream has no intra-cell interference, see Equation (1.13). The first user is selected as the one that achieves the maximal weighted rate:
In a next step the remaining Ntx−1 precoders of the unitary precoding matrix V are determined via Gram-Schmidt orthogonalization starting with v1. Finally, the remaining users are successively assigned to the precoders, see Algorithm 3, which is represented in
In the following, an algorithm will be described which is somewhat similar to the algorithm shown in reference [20]. In [20] users and precoders are determined successively, where in every step the weighted sumrate is maximally increased and the precoder is orthogonal to the already determined precoders, see Algorithm 5, which is represented in
In the following, an algorithm will be described which is similar to the algorithm described in reference [21]. Multiple receive antennas are considered in [21], where receive and transmit filters are determined jointly. First, for every user an eigenvalue decomposition (EVD) HkHHk is computed and the user with the highest maximal eigenvalue is selected as the first user. Assuming the eigenvalue decomposition returns a full unitary basis Vk, it is used as the precoder matrix. In a second step user to precoder allocation is performed successively. Maximum ratio combining (MRC) receive filters are used, for data stream d that is Hs
Please note that vdHHs
In the following, the presented alternative methods are summarized and the relationship to the embodiments according to the invention will be discussed. For an overview see Table 1.1, which is represented in
In the power constraint case, see Section 1.2, the global optimal transmission strategies are found by the MIMO BC-MAC duality [4, 5] in case DPC is used. There exist various low complexity approaches, such as the SESAM [9, 10] algorithm, that typically build on zero-forcing. The global optimal solution for linear precoding is an open problem for multiple receive antennas, for single antenna receivers solutions can be found by global optimization [6, 7]. An approach for local optimal solutions can be found in [8] and LISA [11, 12, 13] provides a low-complexity algorithm.
For DPC the white transmit covariance constraint an algorithm can be derived as a special case from the results in [14]. The global optimal solution for linear unitary precoding is an open problem for multiple receive antennas, for single antenna receivers a solution is available for the 2 user case [16]. Local optimal solutions for single antenna receivers can be found in [17, 18].
Further, there exist multiple low-complexity algorithms for MISO [19, 20] and the MIMO case is treated in [21].
Embodiments according to the invention create a low-complexity algorithm for unitary linear precoding (Unitary-LISA) that can be applied for multiple receive antennas. For the case of single antenna receivers it is also applicable and provides similar results as the algorithm according to [20].
Additional details regarding the functionality of the transmission filter calculator will be described subsequently, wherein it should be noted that any of the features and functionalities of the method for determining transmit filters may be included in, or performed by, the transmission filter calculator 1000. Moreover, it should be noted that the transmission filter calculator 1000 may be part of a transmitter or transceiver like, for example, a base station in a mobile communication network. It should be noted that the transmission filter calculator may, for example, receive the channel matrices from a channel estimator. Alternatively, however, the channel matrices (or information representing the channel matrices) may be received from communication devices (like, for example, receive communication devices) with which a transmit communication device comprising the transmission filter calculator 1000 desires to communicate. Moreover, the transmission filter calculator may receive the distortion covariance matrix, or an information representing the distortion covariance matrix, from a receive communication device with which the transmit communication device comprising the transmission filter calculator 1000 desires to communication. However, the transmission filter calculator 1000 may alternatively receive a combined information, which may combine the channel matrix information and the distortion covariance matrix information. This combined information (for example, in the form of a product matrix information describing a product matrix, which will be described below) may, for example, be received from a receive communication device with which the transmit communication device comprising the transmission filter calculator 1000 desires to communicate.
Moreover, it should be noted that the transmit filters or transmit filter vectors vd may be used for a transmit beamforming. In other words, the entries of a transmit filter vector may describe intensities (and possibly phases) with which a signal representing a data stream d is transmitted via multiple antennas of the transmit communication device comprising the transmission filter calculator. Accordingly, the transmit filters vd associated with multiple data streams having data stream indices d describe a “spatial diversity” with which the different data streams are transmitted.
The method 1100 receives channel matrices Hs
The method comprises iteratively determining the transmit filter for a plurality of data streams, wherein a transmit filter is chosen in a given iteration such that the transmit filter chosen in the given iteration is orthogonal to one or more transmit filters chosen in one or more previous iterations (or orthogonal to all transmit filters chosen in the one or more previous iterations), and such that a weighted or unweighted transmission rate obtainable using the transmit filter chosen in the given iteration is maximized (or, at least approximately maximized).
When iteratively determining the transmit filters, the transmit filter is chosen in dependence on a matrix product term which is based on a channel matrix Hs
Further details regarding the method represented in
It should be noted that the method 1200, which is represented in
The method 1200 comprises a step 1210 of initializing a projector matrix to be a full rank matrix having equal Eigenvalues and orthogonal Eigenvectors. Accordingly, the projector matrix Π(1) of the first iteration (indicated by “(1)”) may be initialized to be the identity matrix I or a multiple thereof.
In a step 1220, a rning variable d may be initialized to designate first data stream and at the same time to designate a first iteration (wherein it is should be noted that a d-th data stream is typically allocated in a d-th iteration). For example, the running variable d may be set to take the value of 1.
In a step 1230, a user (or receive communication device) sd and the transmit filter or transmit filter vector vd are determined such that a weighted data rate (also designated as transmission rate) rd obtained in the d-th iteration is maximized. Here, the condition is observed that the transmit filter is chosen in a given iteration such that the chosen transmit filter is orthogonal to one or more (and advantageously all) transmit filter chosen in one or more previous iterations. Furthermore, it is ensured that the norm of the transmit filter vector describing the transmit filter takes a predetermined value, for example, 1 (which may be worded as the condition vdHvd=1).
For example, an optimization (or maximum search) as indicated in block 1230 of
Further details regarding a possible determination of the user sd and the transmit filter vd will be described below taking reference to
The method 1200 optionally comprises a step 1240 of computing a receive filter ud, which may be applied, for example, in the receive communication device sd. The receive filter ud may, for example, define how the signals from a plurality of antennas of a receive communication device are combined to obtain a signal for the extraction of the data of the data stream having data stream index d, which is directed to the receive communication device (or user) having receive communication device index (or user index) sd. For example, the equation shown in block 1240 of
Subsequently, an optional check 1250 may be performed whether a stopping criterion is fulfilled or not. For example, if the stopping criterion is fulfilled, the further execution of the algorithm may be avoided. Otherwise, the algorithm may proceed with a step 1260 of updating the projector matrix. However, it should be noted that the check 1250 is not necessitated.
In step 1260, the projector matrix II(d) is updated such that a transmit filter described by a transmit filter vector vd is an Eigenvector with associated Eigenvalue of 0 for the updated projector matrix II(d+1), or such that Eigenvectors associated with non-zero Eigenvalues of the updated projector matrix are orthogonal to transmit filter vectors describing transmit filters chosen in previous iterations and in the current iteration. For example, the projector matrix II(d+1) for the next iteration may be derived from the projector matrix II(d) for the current iteration according to the equation shown in block 1260 of the method 1200.
Moreover, the algorithm 1260 also comprises checking, in a step 1270, whether another data stream is to be processed. For example, it may be checked whether the iteration index or data stream index d is smaller than the number Ntx of transmit antennas of the transmit communication device in which the method 1200 is performed. If the checking 1270 indicates that another data stream is to be processed, the data stream index or iteration index d may be incremented in a step 1280 and the algorithm may proceed with step 1230. Otherwise, if it is found that there is no further data stream to be processed, the algorithm may terminate.
Moreover, it should be noted that the order of the steps may be changed to some degree. For example, the check 1270 may be performed before or after the update 1260 of the projector matrix, such that the update of the projector matrix is only performed if another data stream is to be processed. Similarly, the order of the steps 1240 and 1260 may be changed.
In the following, a possible algorithm for determining the user sd and the transmit filter vd will be described taking reference to
The algorithm 1300 may take the place of the step 1230 of the algorithm 1200 and may constitute a particularly computationally efficient way of determining the user sd and the transmit filter vd.
The algorithm 1300 comprises a step 1310, which is executed for a plurality of users sd (or, equivalently, for a plurality of receive communication devices sd) out of a set of users (or receive communication devices). In the step 1310 a largest Eigenvalue esd(d) of a matrix product term, which is based on the projector matrix II(d) of the current iteration, on the channel matrix Hs
Subsequently, in step 1320, the weighted or unweighted transmission rates rsd(d) obtainable for a plurality of users (or receive communication devices) may be compared, and a selected user (or selected receive communication device) may be chosen on the basis of said comparison. In other words, multiple weighted or unweighted transmission rates obtainable for a plurality of users (or receive communication devices) are compared and the user (or receive communication device) which brings along the largest weighted or unweighted obtainable transmission rate is selected. Accordingly, there is a very simple and computationally efficient mechanism for said selection of a user wherein well-known algorithms for the determination of Eigenvalues can be exploited to improve computational efficiency.
Subsequent to the selection of a user (or a receive communication device) in step 1320, an Eigenvector associated with the largest Eigenvalue of the matrix product term based on the projector matrix of the current iteration, the channel matrix associated with the currently considered (selected) one of the users (or receive communication devices) and the distortion covariance matrix associated with the currently considered (selected) one of the users (or receive communication devices) is determined in step 1330, in order to obtain the transmit filter associated with a data stream to the user (receive communication device) selected in the current iteration. In other words, the Eigenvector associated with said matrix product term may be used, for example, in a normalized form, as the transmit filter vector describing the transmit filter. Thus, the determination of an Eigenvector, which can be implemented using one of many highly efficient algorithms known in the art, may be used in order to determine a transmit filter vector which describes a high precision transmit filter.
To conclude, the algorithm (or method) 1300 provides for a computationally very efficient way of determining a user sd and transmit filter vd. Moreover, it should be noted that additional implementation details will subsequently be described taking reference to
In the following, some details regarding the so-called “Unitary-LISA” algorithm according to an embodiment of the invention will be described.
Said algorithm termed “Unitary-LISA” is a significant improvement of the “LISA”-algorithm described above, and is a low-complexity algorithm. To facilitate the understanding, in a first step a rate expression for unitary precoding will be derived which generalizes (or improves) an expression for single antenna receivers (1.14) to the multiple receive antenna case. It has been found that this improved rate expression allows to design a successive stream allocation: the so-called “Unitary-LISA” algorithm.
Even though said Unitary-LISA algorithm is also applicable for the special case of single antenna receivers, the unitary-LISA algorithm brings along specific advantages when used in combination with multiple antenna receivers (i.e. receive communication devices).
Contrary to the concept described, for example, in reference [21], a correct prediction of the noise plus interference covariance (which is achieved by the noise covariance matrix Cη) allows to find the rate maximizing (MMSE) receive filters. Also, it is important to note that the new algorithm according to the embodiments of the invention does not need additional feedback of cross channels or an interference prediction. The embodiments according to the invention thus provide the following contributions:
“Unitary-LISA”, a low-complexity solution for linear unitary precoding;
A method to find optimal receive filters; and
A method for the associated channel feedback.
In the following, it will be briefly described how the distortion covariance matrix Cη,sd is obtained.
Depending on the antenna configuration unitary precoding and a zero-forcing solution contradict. This means, for unitary precoding one has to accept intra-cell interference, which however is known at the time the transmit beamformer for a datastream is selected. By using (1.1), the total predicted noise plus interference (inter-cell and intra-cell) covariance for a stream d assigned to user sd is
Regarding the above definition of Cη,sd, it should be noted that the distortion covariance matrix Cη,sd is associated with a receive communication device (or user) having index sd, and is typically independent from the iteration. Moreover, it should be noted that σ designates a noise (including random noise and noise introduced by components of the receiver). Moreover, it should be noted that Hs
describes an upper limit of an intra-cell interference, i.e. an interference caused by the transmission of the transmit communication device for which the transmit filter is currently determined. Thus, it can be seen that the total noise plus interference Cnoise+interference,d is somewhat smaller than the noise plus interference described by the distortion covariance matrix Cη,sd.
However, as will be described in the following, the usage of the distortion covariance matrix Cη,sd results in a very efficient algorithm which brings along good results.
Therefore by starting from (1.3) and assuming a rank one covariance given by
we obtain for the rate:
Note that,
is the MSE (mean squared error) and this result can also be derived by using the MMSE receiver [22]. Up to a normalizing factor, the MMSE (minimum mean squared error) filter is given by Cη,s
The optimal precoder vd for data stream d is therefore the normalized eigenvector of Hs
In the following, the “Unitary-LISA” algorithm will be briefly discussed taking reference to
The algorithm 1400 according to
The loop 1420 also comprises a selection 1440 of a receive filter described by a receive filter vector ud.
Moreover, the loop 1420 optionally comprises the evaluation of a stopping criterion, which is shown in reference numeral 1450. For example, the loop may be aborted if sufficient data rate could be allocated. Also, the loop may be aborted if it is found that the allocation of additional links (data streams) does no longer significantly improve the data rate.
Moreover, the projector is updated for the next loop iteration, as shown at reference numeral 1460.
It should be noted that step 1410 corresponds to step 1210, that the loop 1420 corresponds to steps 1220 and 1270, that step 1430 corresponds to step 1230, that step 1440 corresponds to step 1240, that step 1450 corresponds to step 1250 and that step 1460 corresponds to step 1260.
Moreover, it should be noted that the algorithm may also be applied in the special case of single antenna receivers, which may also be considered as a multiple-input-single-output case or MISO case. For the special case of single antenna receivers, the rate expression (1.16) can be shown to be equivalent to equation (1.14). For MISO, an arbitrary scalar receive filter can be selected. Accordingly, algorithm 7 remains applicable with some simplifications for the special case of single antenna receivers.
In the following, the main advantages of the algorithm 1400 according to an embodiment of the invention will be summarized:
multiple data streams per user
The multiple receive antennas allow multiple independent data streams for each user, which is supported in the Unitary-LISA algorithm but not in [21]. Note that these streams are in general not orthogonal.
receive filters and link rate adaptation
Note that the optimal receive filter for a data stream can be found by providing information on the used precoder to the user, which is also necessitated for [21]. For non-unitary approaches such as LISA to compute the optimal receive filters, also knowledge on the transmit covariances of the neighbouring transmitters is necessitated, which usually is obtained by additional reference signals, that are also necessitated for improved link rate adaption. Details are provided with the description of the reference algorithms in Section 7.
As both intra and inter-cell interference are considered correctly when selecting the next data stream, the assumed SINR is also correct and is known without additional feedback of the users. Therefore the link rate can be optimally adapted.
channel feedback
Note that Cη,k−1 can be computed at the receiver, so the user can calculate an efficient channel Cη,k−0.5Hk and provide (quantized) feedback to the transmitter. This means, the Unitary-LISA algorithm does not need feedback of cross-channels or ICI measurements, which makes the algorithm especially attractive for deployable networks and inoperable with existing signaling and feedback methods.
In the following, some simulation results will be discussed taking reference to
Although the restriction to a white transmit covariance means a performance loss for an isolated cell, we do expect a significant gain for interference networks. We compare the Unitary-LISA algorithm to the existing LISA algorithm in a interference network. Both are applied in a scenario with time variant channels and fractional reuse, as in [23], but without dynamic cell selection. This means we have more users that are affected by interference, which should be in favor of our method. For the LISA algorithm we use a prediction of the inter-cell interference according to (1.1). As the interference is not as expected, link rate adaption needs additional processing [2]. The first idea is to multiply the expected rates by a constant factor α that is between zero and one to avoid outages. This approach is called gambling in and by experiments we found that the best performing factor is α=0.1. Alternatively, additional reference signals (2nd pilots) can be used to either estimate the correct SINR, which allows for optimal link rate adaption, or to estimate the real inter-cell interference which would allow to find the optimal receive filters, but this necessitates even more overhead by longer training sequences.
The local optimal approaches in [17, 18] are easily extended to multiple receive antennas, by considering the rate formulation (1.16). But these local optimal approaches can only be used if the user set is given. Running the algorithm for all possible user selections is not practical, we therefore add a simple user selection by deciding for the user set that promises the highest sumrate. In general one would expect the best performance from this optimization based approach, but the non-optimal users selection seams to have a drastic impact. The work in [21] seams to be an (weak and somewhat inaccurate) attempt to generalize the result in [19] to multiple receive antennas. Instead, we correctly extend the work in [19] to the multiple receive antenna case. The work in [20] is for single receive antennas and for this scenario the Unitary-LISA is identical. An accurate generalization to multiple-receive antennas results in Unitary-LISA. Further, we include PU2RC [15] and assume the codebook as in LTE, but only consider multi-user transmission. One has to consider, that PU2RC works with very low channel feedback overhead, while all other attempts assume full CSI at the transmitter. For Unitary-LISA, the inter-cell interference is known and both the optimal link rate and the optimal receive filter can be computed.
In the following, a communication device (for example, a receive communication device) 1500 will be described, which is particularly well-suited to usage in combination with the transmission filter calculation discussed above.
The communication device 1500 is adapted for communicating with a transmit communication device in the presence of one or more interferer communication devices. The communication device 1500 is configured to determine a channel matrix Hs
The communication 1500 is also configured to determine a matrix product which is based on the channel matrix Hs
In other words, communication device 1500 may, for example, comprise a channel matrix determination 1510 and a distortion covariance determination 1520. Also, the communication device 1500 may comprise a matrix product determinator 1530, which is configured to determine said matrix product on the basis of the matrix Hs
For example, the matrix product determinator 1530 may be configured to compute the product matrix M as shown in block 1530 of the communication device 1500.
Naturally, the communication device 1500 may comprise additional functionality to transfer data, encoded audio or video signals, or the like. Also, the communication device 1500 is typically configured to receive data from a transmit communication device (for example, from a transmit communication device comprising the transmission filter calculator discussed above).
The method 1600 comprises a step 1610 of determining a channel matrix describing MIMO channel characteristics between the communication device and the transmit communication device (with which the communication is desired and which is typically in the same communication cell as the communication device performing the method 1600). The method 1600 also comprises a step 1620 of determining one or more channel matrices describing MIMO channel characteristics between the communication device and the one or more interferer communication devices (for example, base stations arranged in different communication cells, which different communication cells are adjacent to the communication cell in which the communication device performing the method 1600 is arranged). Steps 1610 and 1620 may be performed simultaneously in some embodiments, and subsequently in other embodiments.
The method 1600 also comprises a step 1630 of determining a matrix product which is based on the channel matrix describing a MIMO channel between the communication device and the transmit communication device with which a communication is desired and on a distortion covariance matrix describing a noise, an inter-cell interference and an upper limit of an intra-cell interference to obtain a product matrix. For details, reference is made, for example, to the functionality of the matrix product determinator 1530, which may perform the determination 1630.
Moreover, the method 1600 comprises a step 1640 of transmitting the product matrix (or an information describing the product matrix itself, rather than the channel matrix or the distortion covariance matrix individually) to the transmit communication device with which a communication is desired (i.e. the communication device performing the method 1600).
It should be noted that the method 1600 implements the functionality of the communication device 1500.
To conclude, it should be noted that embodiments according to the invention introduce a so-called “unitary-LISA” algorithm, a low-complexity solution for linear unitary precoding in a multi-user MIMO downlink. Unitary precoding results in a white sum-transmit covariance, which allows for a better approximation of interference in the neighboring cells and enables to compute optimal receive filters and perfect link rate adaptation. Moreover, an embodiment according to the invention creates a method for the associated channel feedback. In other words, a method for the associated channel feedback is included in the invention.
Moreover, it should be noted that embodiments of the invention are based on the finding that, for unitary precoding, interference depends on the own precoder. Accordingly, anew method named “Adaptive Unitary Precoding—Unitary LISA” is created. Embodiments according to the invention create an extension to multi-antenna receivers. In the embodiments, interference plus noise covariance of a stream d are considered. For details, reference is made to the above-described formula or the calculation of said covariance Cd.
Moreover, embodiments according to the invention are based on a computation rule for the data rate of the stream d. For details, reference is made to the above equations which describe the computation of the data rate rd.
In some embodiments of the invention, the fact that
H
s
H
C
η,s
−1
H
s
can be computed at the receiver is exploited. In other words, it is advantageous in some embodiments to have a new type of feedback from the receive communication device to the transmit communication device. However, the new feedback, which brings along efficiency improvement, is different from conventional feedback (provided conventionally from a receive communication device to a transmit communication device), but is not increased. In other words, the feedback in the form of a product matrix typically does not necessitate a larger bitrate than conventional feedback.
Some embodiments according to the invention are based on the main idea to use a projection matrix to obtain unitary precoders. For details, reference is made, for example, to the detailed discussion of the algorithm, which was made with respect to
Some embodiments according to the invention create a unitary precoding for users with multiple receive antennas. Some main features of the embodiments of the invention are as follows:
Thus, embodiments according to the invention differ significantly from alternative solutions.
In the following, a multi-cell/cooperative multipoint (CoMP) simulation setup will be described. The setup is as follows:
57 cells wrap-around;
Average 10 mobile users per cell randomly distributed
4×4 MIMO;
Rayleigh fading (vehicular users);
Fractional reuse/proportional fair.
The following Algorithms are compared:
PU2RC (LTE-codebook)
LISA (second pilot);
Local optimal (heuristic user selection);
Gram-Schmidt (extended to multi-user); and
Unitary LISA (according to the present invention).
For results, reference is made to
To further conclude, a simple embodiment according to the invention creates a linear successive user allocation method jointly precoding and selecting users, comprising unitary precoders at the transmitters.
An improvement of said linear successive user allocation further comprises optimal receive filters and optimal link-rate adaptation.
A further improvement of the above-described methods further comprises a channel feedback scheme where the noise covariance matrix is incorporated without increasing the feedback amount.
To further summarize the above, embodiments according to the invention can be applied in multi-cell MIMO communications.
Embodiments of the invention are based on an investigation of different paradigms of cooperation in multi-cell communication networks, on an identification of the potential of those cooperation strategies in cellular communication networks based on MIMO systems, and on a development of respective concepts and algorithms in order to achieve significant gains compared over conventional techniques.
To further conclude, embodiments according to the invention can be used in wireless access, and, in particular, in radio access networks.
To further conclude and briefly summarize, embodiments according to the invention create an adaptive unitary precoding, unitary-LISA. Embodiments according to the invention comprise an extension of unitary precoding to multiple receive antennas. Embodiments according to the invention do not need additional feedback. Embodiments according to the invention provide an optimal rate adaptation and/or provide for optimal receive filters (wherein the latter may, in some cases, cause some overhead).
Embodiments according to the invention outperform alternative solutions for linear unitary precoding like, for example:
It should be noted that the above solutions 2) and 3) are only applicable for single antenna receivers, while embodiments according to the invention are applicable in environments in which receivers comprise a multiple antenna.
To further conclude, embodiments according to the invention reach the goal to reduce uncertainty in interference to obtain more robust downlink strategies.
Although some aspects have been described in the context of an apparatus, it is clear that these aspects also represent a description of the corresponding method, where a block or device corresponds to a method step or a feature of a method step. Analogously, aspects described in the context of a method step also represent a description of a corresponding block or item or feature of a corresponding apparatus. Some or all of the method steps may be executed by (or using) a hardware apparatus, like for example, a microprocessor, a programmable computer or an electronic circuit. In some embodiments, some one or more of the most important method steps may be executed by such an apparatus.
Depending on certain implementation requirements, embodiments of the invention can be implemented in hardware or in software. The implementation can be performed using a digital storage medium, for example a floppy disk, a DVD, a Blu-Ray, a CD, a ROM, a PROM, an EPROM, an EEPROM or a FLASH memory, having electronically readable control signals stored thereon, which cooperate (or are capable of cooperating) with a programmable computer system such that the respective method is performed. Therefore, the digital storage medium may be computer readable.
Some embodiments according to the invention comprise a data carrier having electronically readable control signals, which are capable of cooperating with a programmable computer system, such that one of the methods described herein is performed.
Generally, embodiments of the present invention can be implemented as a computer program product with a program code, the program code being operative for performing one of the methods when the computer program product runs on a computer. The program code may for example be stored on a machine readable carrier.
Other embodiments comprise the computer program for performing one of the methods described herein, stored on a machine readable carrier.
In other words, an embodiment of the inventive method is, therefore, a computer program having a program code for performing one of the methods described herein, when the computer program runs on a computer.
A further embodiment of the inventive methods is, therefore, a data carrier (or a digital storage medium, or a computer-readable medium) comprising, recorded thereon, the computer program for performing one of the methods described herein. The data carrier, the digital storage medium or the recorded medium are typically tangible and/or non-transitionary.
A further embodiment of the inventive method is, therefore, a data stream or a sequence of signals representing the computer program for performing one of the methods described herein. The data stream or the sequence of signals may for example be configured to be transferred via a data communication connection, for example via the Internet.
A further embodiment comprises a processing means, for example a computer, or a programmable logic device, configured to or adapted to perform one of the methods described herein.
A further embodiment comprises a computer having installed thereon the computer program for performing one of the methods described herein.
A further embodiment according to the invention comprises an apparatus or a system configured to transfer (for example, electronically or optically) a computer program for performing one of the methods described herein to a receiver. The receiver may, for example, be a computer, a mobile device, a memory device or the like. The apparatus or system may, for example, comprise a file server for transferring the computer program to the receiver.
In some embodiments, a programmable logic device (for example a field programmable gate array) may be used to perform some or all of the functionalities of the methods described herein. In some embodiments, a field programmable gate array may cooperate with a microprocessor in order to perform one of the methods described herein. Generally, the methods are performed by any hardware apparatus.
While this invention has been described in terms of several advantageous embodiments, there are alterations, permutations, and equivalents which fall within the scope of this invention. It should also be noted that there are many alternative ways of implementing the methods and compositions of the present invention. It is therefore intended that the following appended claims be interpreted as including all such alterations, permutations, and equivalents as fall within the true spirit and scope of the present invention.
Number | Date | Country | Kind |
---|---|---|---|
12162634.5 | Mar 2012 | EP | regional |