WO2016123510A1 - Distributed uplink coordinated multipoint transmission via symmetric two-stage equalization - Google Patents

Abstract

A receiver includes a first equalizer and a second equalizer. The first equalizer receiving a signal representing at least one symbol transmitted on an orthogonal frequency division multiplexing (OFDM) network by a first user equipment (UE) and at least one additional UE. The transmission is by the first UE to the receiver. The first equalizer computing, based on the signal, at least one first partially-equalized symbol of the first UE. The second equalizer is communicatively coupled to the first equalizer. The second equalizer receiving the at least one first partially-equalized symbols of the first UE, receiving at least one second partially-equalized symbol, and computing at least one equalized-symbol of the at least one first UE. The at least one second partially-equalized symbol is received from at least one additional base station being part of a coordinated multipoint (CoMP) group. Related apparatus, systems, techniques, and articles are also described.

Description

Distributed Uplink Coordinated Multipoint Transmission via

Symmetric Two-Stage Equalization

RELATED APPLICATIONS

[0001] This application claims priority under 35 U.S.C. § 119 to U.S. Provisional

Application No. 62/110, 137 filed January 30, 2015, and U.S. Provisional Application No.

62/153,180 filed April 27, 2015, the entire contents of each of which are hereby incorporated by reference herein.

TECHNICAL FIELD

[0002] The subject matter described herein relates to dynamic interference management using distributed uplink coordinated multipoint (CoMP) transmission via two- stage equalization in a multi-cellular multiple-input and multiple-output (MEVIO) orthogonal frequency division multiplexing (OFDM) network.

BACKGROUND

[0003] Mutual interference is unavoidable in modern spectrally efficient wireless systems: even when using orthogonal multiplexing systems such as Time Division Multiple Access (TDMA), synchronous Code Division Multiple Access (CDMA) or Orthogonal Frequency Division Multiple Access (OFDMA), multiuser interference can originate from channel distortion and from multiple user equipment (UEs) communicating on the same time and frequency resources to different evolved node Bs (eNBs), base-stations, or access points.

SUMMARY

[0004] In an aspect, a low-complexity distributed uplink CoMP via symmetric two-stage equalization (S-TSEQ) is for decoding the data of multiple users within a CoMP group. It is shown that the S-TSEQ algorithm can uniformly improve the post-processing signal to interference plus noise ratio (PP-SINR) in comparison with a log-likelihood ratio exchange (LLR-E) approach, whereas the latency requirements of S-TSEQ can be substantially lessened compared with a two-stage equalization - minimum mean squared error (TSEQ-MMSE) approach. In the symmetric version of TSEQ, in the first stage each e B can perform its own joint equalization of all UEs. However, in the second stage each e B decodes only the UE that is connected to it while treating the contribution from other UEs as interference.

[0005] In an aspect, a first base station belonging to a coordinated multipoint transmission group having other base stations, can receive data characterizing symbols equalized by the first base station and of a transmission from a user equipment to the first base station as received by the first base station. The first base station can receive data characterizing symbols equalized by one or more of the other base stations and of the transmission from the user equipment to the first base station as received by the one or more other base stations. The first base station can determine, using the received data, equalized symbols for the transmission from the user equipment. Contributions from additional user equipment are treated as interference.

[0006] In another aspect, interference reduction and/or cancellation can include at least one of: Joint Reception in Frequency-Domain (JRFD); Log-Likelihood Ratio (LLR) Exchange (LLR-E); Transport Block Selection Diversity (TBSD); Two-Stage Equalization (TSEQ), and coordinative scheduling with adaptive interference mitigation (CS-AIM). In some implementations, a system may dynamically manage interference using one or more of these approaches. In some implementations, the system may determine which approach to use for decoding signals transmitted by multiple user equipments.

[0007] In yet another aspect, a receiver includes a first equalizer and a second equalizer. The first equalizer receiving a signal representing at least one symbol transmitted on an orthogonal frequency division multiplexing (OFDM) network by a first user equipment (UE) and at least one additional UE. The transmission is by the first UE to the receiver. The first equalizer computing, based on the signal, at least one first partially-equalized symbol of the first UE. The second equalizer is communicatively coupled to the first equalizer. The second equalizer receiving the at least one first partially-equalized symbols of the first UE, receiving at least one second partially-equalized symbol, and computing at least one equalized-symbol of the at least one first UE. The at least one second partially-equalized symbol is received from at least one additional base station being part of a coordinated multipoint (CoMP) group. The at least one second partially-equalized symbol characterizes the signal transmitted on the OFDM network by the first UE and as received by the at least one additional base station.

[0008] In yet another aspect, a method for implementation by a first base station belonging to a coordinated multipoint transmission (CoMP) group, the CoMP group having at least one other CoMP member base station, is provided. Data is received characterizing at least one partially-equalized first symbol by the first base station and of a transmission from a first user equipment (UE) to the first base station as received by the first base station. Data is received characterizing at least one partially-equalized second symbol by the at least one other CoMP member base station and of the transmission from the first UE to the first base station as received by the at least one CoMP member base station. Using the received data, at least one equalized symbol for the transmission from the first UE is determined.

Contributions from additional UE to the transmission from the first UE to the first base station are processed as interference. The receiving data characterizing partially-equalized first symbols, the receiving data characterizing partially-equalized second symbols, and the determining are performed by at least one data processor forming part of at least one computing system of the first base station. [0009] One or more of the following features can be included in any feasible combination. The first equalizer can implement joint equalization. The second equalizer can implement adaptive interference-rejection combining (IRC). The first equalizer can compute at least one equalized-symbol of the one or more additional UEs for provision to the at least one additional base station being part of the CoMP group. The at least one symbol transmitted on the OFDM network by the first UE and at least one symbol transmitted on the OFDM network by the at least one additional UE can share time and frequency resources. The second equalizer can process contributions to the signal from the additional UE as interference when it computes the equalized-symbols of the first UE.

[0010] The transmission from the first UE to the first base station and a transmission from the additional UE to the at least one other CoMP member base station can share time and frequency resources. A second transmission can be received from one of the additional UE, the second transmission to one of the at least one other CoMP member base station. Symbols of the received transmission from the one of the at least one additional UE can be partially equalized. Partially-equalizing symbols can be performed using joint- equalization. The partially-equalized symbols can be transmitted to one of the at least one other CoMP member base stations.

[0011] Auxiliary channel information of the at least one additional UEs can be exchanged with the at least one CoMP member base station. The partially-equalized second symbols can be received from the at least one CoMP member base station on a per time- frequency resource basis. Complex valued scale factors can be exchanged per demodulated reference symbol (DMRS) tone. Complex valued scale factors can be computed as beamformed channel characteristics. Noise variance values can be exchanged per

demodulated reference signal (DRMS). The noise variance values can be of effective noise plus interference. Noise variance values of effective noise plus interference can be computed. The noise variance values can be computed as a function of beamforming values, channel noise covariance, and channel characteristics.

[0012] The determined equalized symbols can be provided. Providing can include transmitting, storing, or processing.

[0013] Articles are also described that comprise a tangibly embodied machine- readable medium embodying instructions that, when performed, cause one or more machines (e.g., computers, etc.) to result in operations described herein. Similarly, computer systems are also described that can include a processor and a memory coupled to the processor. The memory can include one or more programs that cause the processor to perform one or more of the operations described herein. Additionally, computer systems may include additional specialized processing units that are able to apply a single instruction to multiple data points in parallel. Such units include but are not limited to so-called "Graphics Processing Units (GPU) "

[0014] The details of one or more variations of the subject matter described herein are set forth in the accompanying drawings and the description below. Other features and advantages of the subject matter described herein will be apparent from the description and drawings, and from the claims.

DESCRIPTION OF DRAWINGS

[0015] The accompanying drawings, which are incorporated in and constitute a part of this specification, show certain aspects of the subject matter disclosed herein and, together with the description, help explain some of the principles associated with the disclosed implementations. In the drawings,

[0016] FIG. 1 is an exemplary system, according to some implementations of the current subject matter; [0017] FIG. 2 is a process flow diagram of a method according to some implementations of the current subject matter;

[0018] FIGs. 3-10 are complimentary cumulative distribution functions (CCDF) for different example simulations according to the current subject matter;

[0019] FIG. 11 illustrates an exemplary system, according to some implementations of the current subject matter;

[0020] FIG. 12 is an example system diagram illustrating an example scenario in which a CoMP group and associated algorithms can be used to improve system performance;

[0021] FIG. 13 is a system block diagram of an example e B illustrating processing flow of S-TSEQ;

[0022] FIG. 14 is a data flow diagram illustrating data flow among e Bs of the example CoMP group of FIG. 12 during transmission decoding;

[0023] FIGs. 15-18 are complimentary cumulative distribution functions (CCDF) for different example simulations according to the current subject matter; and

[0024] FIG. 19 is a process flow diagram illustrating a method of equalizing signals having symbols where multiple UEs transmit on the same time-frequency resource.

DETAILED DESCRIPTION

[0025] FIG. 1 illustrates an exemplary system 100, according to some implementations of the current subject matter. The system 100 can be implemented as a centralized cloud radio access network ("C-RAN"). The system 100 can include at least one intelligent remote radio head ("iRRH") unit 102 and an intelligent baseband unit ("iBBU)

104. The iRRH 102 and iBBU 104 can be connected using Ethernet fronthaul ("FH") communication 106 and the iBBU 104 can be connected to an evolved packet core ("EPC") using backhaul ("BH") communication 108. The user equipment (not shown in FIG. 1) can communicate with the iRRH 102. [0026] In some implementations, the iRRH 102 can include the power amplifier ("PA") module 112, the radio frequency ("RF") module 114, LTE layer LI (or PHY layer) 116, and a portion 118 of the LTE layer L2. The portion 118 of the LTE layer L2 can include the MAC layer and can further include some functionalities/protocols associated with radio link control ("RLC") and a packet data convergence protocol ("PDCP"). The iBBU 104 can be a centralized unit that can communicate with a plurality of iRRH and can include LTE layer L3 122 (e.g., RRC, RRM, etc.) and can also include a portion 120 of the LTE layer L2. Similar to portion 118, the portion 120 can include various functionalities/protocols associated with PDCP. Thus, the system 100 can be configured to split

functionalities/protocols associated with PDCP between iRRH 102 and the iBBU 104.

[0027] In some implementations, the system 100 can form part of an evolved Node B ("eNodeB" or "eNB") and there may be a plurality of eNBs interconnected with one another using an interface (e.g., X2). The interface can be established between two eNodeBs in order to provide an exchange of signals, which can include a load- or interference-related information as well as handover-related information. The eNodeBs can communicate with the evolved packet core EPC 108

[0028] In some implementation, the system 100 can implement coordinated multipoint ("CoMP") transmission features. In some implementations, system 100 can perform dynamic interference management via coordination transmission and joint reception. Such features can increase data throughput rate and can work with 4G LTE-Advanced.

[0029] The uplink ("UL") CoMP features in LTE enables joint decoding of multiple users connected to different eNBs but transmitting on same time-frequency resources. For example, FIG. 12 is an example system diagram illustrating an example scenario in which a CoMP group and associated algorithms can be used to improve system performance. The example CoMP group includes three eNBs denoted as eNBl 1205, eNB2 1210, and eNB 3 1215. Each eNB in the CoMP group can communicate with one another and each eNB has an effective coverage area or footprint (e.g., the eNB cell), represented by circles 1220, 1225, and 1230, respectively. Three UEs, denoted by UE1 1235, UE2 1240, and UE3 1245, reside within a cell overlap region, are connected to their respective eNBs (denoted by dotted lines), and transmit on same time-frequency resources (e.g., symbols). Thus, in the illustrated scenario, each eNB (1205, 1210, and 1215) is connected to one UE (1235, 1240, and 1245, respectively) but receives transmissions from all three UEs (1235, 1240, and 1245). In order to decode each UE transmission, coordination among eNBs can occur. Depending on a specific CoMP decoding algorithm, one eNB or node may serve as a master node for decoding transmissions. Some CoMP decoding algorithms are distributed and do not have a master node.

[0030] UL-CoMP may provide significant gains in the cell-edge spectral efficiency by performing joint reception in frequency-domain (JRFD) of the cell-edge users in the overlapping coverage areas. However, a challenge of implementing JRFD is the higher CoMP-specific feedback rate requirements, and lower latencies, from participating CoMP nodes to the master CoMP node where the joint decoding takes place. Because of this, JRFD appears to be more suitable to a centralized RAN solution in which RRHs are connected with the central BBU via dedicated fiber links. However, some eNB architectures have the Radio, LI, and some of L2 processing hosted on iRRH 102, and the L3 and higher layers are hosted on iBBU 104 and therefore have feedback rate and latency limits. Various UL-CoMP approaches are described herein with a detailed analysis of the CoMP-specific overhead (in Mbps) and the post-processing SINR gains relative to a baseline non-CoMP adaptive interference-rejection combining (IRC) receiver.

[0031] A low-complexity distributed uplink CoMP via symmetric two-stage equalization (S-TSEQ) is described for decoding the data of multiple users within a CoMP group. It is shown that the S-TSEQ algorithm uniformly improves the post-processing signal to interference plus noise ratio (SINR) in comparison with a log-likelihood ratio exchange (LLR-E) approach described more fully below, whereas latency requirements of S-TSEQ are substantially lessened compared with a TSEQ-MMSE approach, also described more fully below.

[0032] The two-stage equalization (TSEQ) approach to UL CoMP is described more fully below. Described here is a symmetric version of TSEQ that has relaxed latency requirements. In the symmetric version of TSEQ, in the first stage each eNB can perform its own joint equalization of all UEs. However, in the second stage each eNB can decode only the UE that is connected to it while treating the contribution from other UEs as interference.

[0033] FIG. 13 is a system block diagram of an example eNB 1300 illustrating processing flow of S-TSEQ. CE denotes the channel estimation unit 1330„, and N is the number of UEs within the CoMP group. The example eNB 1300 includes a number of receive antenna's 1305, which receive transmissions or signals from multiple UEs operating on shared resources. One of the UEs is connected to the eNB 1300 while others of the UEs can be connected to other eNBs in the UL CoMP group. Each antenna 1305 can have corresponding digital down conversion units 1310 and fast Fourier transform units 1315 for processing received signals. An allocation extraction unit 1320 can receive the received and processed transmission signal (e.g., y,) as well as UE information 1325, and extract allocation information. The allocation information can include, for example, the number of resource blocks allocated for CoMP, the modulation information, the code rate, whether frequency- hopping is enabled or not, whether additional sounding symbols are allocated, the redundancy version of the current transmission, and the like. UE information 1325 can include, for example, the radio network temporary identity (RNTI), the UE capability, and the like. [0034] Channel estimation units 1330„ n= 1, N, estimate the respective channel information using the allocation information and the processed received transmission signal, which is supplied to stage- 1 joint equalizer 1335 along with the allocation information. Stage-1 joint equalizer 1335, for example can perform stage 1 equalization, as described more fully below with reference to TSEQ to produce soft equalized symbols (also referred to herein as partially-equalized symbols). Soft equalized symbols can be complex- valued symbols, which, when properly sliced using the constellation information can provide symbols belonging to the constellation. In some sense, the soft equalized symbols can be viewed as perturbed versions of the constellation symbols. Soft equalized symbols can be determined for all UEs, including one connected to eNB 1300 and those connected to other eNBs in the UL CoMP group. As more fully described below with reference to TSEQ, these soft equalized symbols represent intermediate solutions, which can be used to determine final equalized symbols. Stage-2 equalizer 1340 receives the soft equalized symbols for the UE connected to the eNB 1300, while the soft equalized symbols for the other UEs connected to the other eNBs belonging to the UL CoMP group can be quantized at 1345 and sent to the other eNBs belonging to the UL CoMP group. Stage-2 equalizer 1340 can also receive auxiliary channel information and soft equalized symbols of the UE connected to eNB 1300 from the other eNBs belonging to the UL CoMP group. Auxiliary channel information can include additional channel state information (e.g., instant estimated channel gains, noise variance, covariance matrices, and the like) that may be needed at a specific CoMP node to enable it perform a final decision. Auxiliary channel information can include the combination of noise variance (a real quantity), and the inner product of the equalizer weight with its local channel estimate (a complex quantity). Auxiliary channel information can depend on the specific CoMP scheme being used. For example, with JRFD, each eNB simply shares its received samples with the master eNB, and the master eNB estimates the channels between each UE-eNB link. So, the auxiliary channel information is zero.

[0035] Stage-2 equalizer 1340 can determine finalized equalized symbols for the connected UE while treating contributions from all other UEs as interference. The behavior of stage-2 equalizer 1340 is described more fully below.

[0036] FIG. 14 is a data flow diagram illustrating data flow among eNBs of the example CoMP group of FIG. 12 during transmission decoding. At 1405 each eNB (eNBl 1205, eNB2 1210, and eNB3 1215), after having received transmissions from each UE (UE1 1220, UE2 1225, and UE3 1230) on a shared time and frequency resource, performs stage-1 joint equalization (e.g., via stage-1 joint equalizer 1335) . Specifically, each eNB computes soft equalized symbols for UE1 1235, UE2 1240, and UE3 1245. Thus eNBl 1205 computes a soft equalized symbol for each UE from the signals received by eNBl 1205.

[0037] At 1410, eNBl 1205 sends the soft equalized symbol that eNBl 1205 determined for UE2 1240 and based on the signal received by eNBl 1205 to eNB2 1210 and at 1415, eNBl 1205 sends the soft equalized symbols that eNBl 1205 determined for UE3 1245 and based on the signal received by eNBl 1205 to eNB 3 1215.

[0038] At 1420 and 1425, eNBl 1205 receives the soft equalized symbols that eNB2 1210 and eNB3 1215 determined for UEl 1235 based on the respective signals received by eNB2 1210 and eNB 3 1215, respectively. At 1430 and 1435 eNB2 1210 and eNB 3 1215 also exchange their respectively computed soft equalized symbols.

[0039] At 1440, each of eNBl 1205, eNB2 1210, and eNB 3 1215 can perform stage-2 equalization using the soft equalized symbols for their respectively connected UE as received by each eNB (e.g., and using stage-2 receiver/equalizer 1340). Each eNB can perform stage-2 equalization while treating the contributions from all the other UEs as interference. [0040] Thus, according to some implementations of the current subject matter, a first eNBs belonging to a coordinated multipoint transmission group having other eNBs can receive symbols equalized by the first eNB and of a transmission from a UE to the first eNB as received by the first eNB. The first eNB can also receive symbols equalized by one or more of the other eNBs and of the transmission from the UE to the first eNB as received by the one or more other eNBs. The first eNB can determine, using the received symbol data, equalized symbols for the transmission from the UE where contributions from additional UE are treated as interference.

[0041] A detailed discussion of Stage-2 equalization behavior follows.

[0042] Consider a CoMP group that is defined as a group of N eNBs that collectively decodes N user equipments (UEs). (In the example of FIG. 12, there are 3 eNBs and 3 UEs). In the absence of UL-CoMP, UE-/^' is served by eNB-/^', z^'=l, . . . ,N. The modulation symbol of UE-y on a given time-frequency resource is denoted by Sj. With this, the received signal at eNB-/^' on a given tone-frequency resource is

V.; :::: ^ 1*·; Si._; , >

Where h,j is the channel from UE-y to eNB-/^', z^{' '}=l, . . . ,N, and n, is the noise plus interference covariance matrix at eNB-/^'. Assuming that n, has zero-mean and covariance matrix R_;. With NR_X antennas at eNB-/, y, is of size

and R, is of size VR*_,,-

[0043] A note on the notation: column vectors are denoted by lower case bold faced letters (e.g., JC), whereas matrices are denoted by upper case bold faced letters (e.g., A).

[0044] Vector denotes a weight vector employed at eNB-/^' to estimate Sj and Sj^ denotes the estimate of Sj at eNB-/^'. Now, Sj^ is given by where ^¾ is the Hermitian (conjugate

is a complex- valued signal coefficient of UE-/ at eNB-/^' when UE-y's signal is processed, and v J, i) = wj_j ii is the effective noise plus interference random variable that has mean zero and variance o_v(j_iVi = w- ^ Wj^. Also noted is that the beamformer Wj_:i in Eq. 2 is given by

[0045] In stage-2 of symmetric-TSEQ (e.g., at stage-2 equalizer 1340), each eNB decodes its own UE while treating the contributions from all the other UEs as interference. This reduces overall decoding complexity and eliminates the need to exchange the decoded transport blocks. To enable this at eNB-/^', in addition to having its own local estimate s_i also needed are estimates s_{i rn} of UE-/^' from eNB-w, m=l, . . . ,N, m≠i. That is, to enable the second stage processing, each eNB exchanges N-1 estimates of Eq. 2 on a per time-frequency resource basis. Some auxiliary channel information can be exchanged, which will be described shortly.

[0046] Focusing on processing at eNB-/^', the N estimates s_{i m}, m = 1, ... , N, can be stacked in a column vector as

[0047] From the above equation, eNB-/^' has N equations and N unknowns (¾ 1=1, . . . ,N) and it only desires a final estimate of s,. For this, an IRC receiver can be employed at eNB-/^' that treats the remaining N-l UE's signals as part of interference. With an IRC receiver in the second stage at eNB-/^', the MMSE weight vector is denoted by < ,·, and is given by

where R_v. is the covariance matrix of v,. And the final estimate of s, is

[0048] A note on computing various factors appearing in the weight vector Eq. 5. First, computation of R_v.. Note that v(i,i) is local to eNB-/^' whereas v(i,m), m=l, . . . ,N, m≠/^', appears at eNB-m. Also, due to spatial independence, for i≠m, v(i,i) and v(i,m) are uncorrelated. So, R_v. is

*

where diag(-) is the diagonal matrix that is formed using its input vector argument. In effect, to i?_v.eNB-/^' needs to estimate

to be exchanged from eN -m, m=l, ... , N, m≠i. This amounts of exchanging N-l positive quantities from the other e Bs.

[0049] Next, the quantities c, , 1=1, ... , N, in the formation of the final weight vector can be reformulated as in Eq. 5. Note that c(i,i,l), 1=1, ... , N, are available locally at eNB-z, whereas c(i,j,l), j=\, ... , N, j≠m, l=\,... ,N, should be estimated at eNB-y and transmitted to eNB-z^'. These are N ^x (N-l) complex-valued quantities.

[0050] In summary, to enable estimation of s, at eNB-z in the second stage, the following information exchange can occur: (1) per time-frequency resource: Sj^j = 1, ... , N,j≠ i, which are complex -valued; (2) per DMRS (demodulation reference symbol) tone: the N ^x (N-l) scale factors c(i,j,l), j=\, ... ,N, j≠i, l=\,... ,N, which are complex-valued; and (3) per DMRS tone: the N-l noise variance σ^_,ΐ) ^{= w}}i^-i^wj, j⁼^_>— _>N, j≠i , which are strictly positive.

[0051] Assuming an allocation of K resource blocks for CoMP and assuming that channel estimation is performed once per physical resource block (PRB) per slot, on a per- subframe basis, 288 ^χ (N-l) ^χΚ complex- valued equalized symbols are needed, 2*N(N-1) xK complex -valued scale factors and 2x (N-l) xK positive noise variances to be transmitted to eNB-z from the remaining N-l eNBs.

[0052] FIG. 19 is a process flow diagram illustrating a method 1900 of equalizing signals having symbols where multiple UEs transmit on the same time-frequency resource. The method 1900 can be implemented by a base station belonging to a CoMP group (for example, as illustrated in FIG. 12) and in communication with a first UE. The method 1900 of equalizing signals can eliminate the need to exchange the decoded transport blocks enabling the current subject matter to meet latency and data rate requirements.

[0053] At 1910, a first base station or receiver receives data characterizing partially (or soft) equalized symbols related to a signal received by the first base station. The signal can include a transmission from the first UE as well as contributions from additional UE in the network not connected to the first base station but transmitting on the same time- frequency resource. The base station can have computed or determined the partially- equalized symbols using, for example, joint equalization.

[0054] At 1920, the base station can receive additional partially-equalized symbols from other base stations in the CoMP group. The additional partially-equalized symbols can have been partially-equalized by the other base stations in the CoMP group (for example, by using joint equalization). The additional partially-equalized symbols can relate to the transmission from the first UE as received by the other base stations in the CoMP group.

[0055] At 1930, completely (or finally) equalized symbols can be determined for the transmission from the UE. The determination can be performed by treating the contributions from the additional UE as interference (in contrast to receiving data from other base stations in the CoMP group in order to have additional information to equalize or otherwise identify and process the contributions from the additional UE). These completely equalized symbols can be provided, for example, storing, processing, or transmitting.

[0056] In some implementations, the first base station can determine or compute partially-equalized symbols for the additional UEs from the signal as received by the first base station. These partially-equalized symbols can be transmitted to the other base stations in the CoMP group. Auxiliary channel information (of the first UE and/or additional UE) may be exchanged with the other base stations in the CoMP group.

[0057] Some simulation results are presented that describe the relative performances of S-TSEQ, TSEQ-MMSE, TSEQ-ZF and LLR-E. Note that the schemes TSEQ-ZF, TSEQ- MMSE and LLR-E are described in detail below. In the simulations, two eNBs and three UEs are considered. UE-1 is attached to eNB-1, UE-2 is attached to eNB-2, and UE-3 is the common uplink interference that overlaps the allocations of UE-1 and UE-2. The CoMP group includes these two e Bs with the goal of jointly (and in a distributed fashion) decoding UEs 1 and 2. Each eNB is assumed to have 4 receive antennas, and each UE has one transmit antenna.

[0058] The channel models between the eNB-UE links are described in Table 1 and all the channels are spatially independent. Table 1 shows multipath channel models between eNB-/^' and UE-y, i=l,2 and j= 1,2,3 and note that all the channels are independent across the eNB-UE links.

[0059] fi is defined as the average received signal-to-noise ratio (SNR) from UE-y to eNB-/^' (i=l,2 and j=l,2,3). With this definition, two simulation scenarios are considered. In the first scenario, = y_2j2 = 1 dB, f_{1 2} = 7_2,i = 7i_,3 = 7_3,1 = 0 dB. Additionally, for each UE fluctuations in transmit power (for example, due to power control errors) are modeled as independently and uniformly distributed within [-5,5] dB. In the second scenario, 7i_,i ⁼ 7_2,2 ⁼ 1 dB, 7_{1 2} = 7_{2 1} =— 5 dB, f_{1 3} = f_{3 1} = 0 dB and there are no fluctuations in transmit power. The post-processing (PP) signal to noise plus interference ratio (SINR) for UEs 1 and 2 with various UL-CoMP algorithms is of interest. In particular, the complimentary cumulative distribution function (CCDF) of the PP-SINR is of interest as it captures the cell-edge SINR improvement.

[0060] FIGs. 15 and 16 show the CCDF of PP-SINR for UEs 1 and 2, respectively, under the simulation scenario 1. Similarly, in FIGs. 17 and 18 are plots of the CCDF of PP- SF R of UEs 1 and 2, respectively, under the simulation scenario 2. From FIGs. 15 through 18 it can be observed that the proposed S-TSEQ approach uniformly improves the PP-SF R performance in comparison with the LLR-E approach described below. On the other hand, although S-TSEQ is found to be somewhat inferior to the TSEQ-MMSE approach, it is noted that S-TSEQ offers improved latency performance (or the latency requirements are relaxed for S-TSEQ) over the TSEQ-MMSE approach. This is because the TSEQ-MMSE as described below may require that in the second-stage all the UEs have to be decoded at a CoMP leader node, and the decoded transport blocks (TBs) of the UEs that are not connected to the master CoMP node have to be fed back to their respective eNBs, which leads to an increase in latency. On the other hand, with S-TSEQ, at the second stage, each eNB decodes only the UEs that are connected to itself while treating the contribution of other UEs as interference.

[0061] Although a few variations have been described in detail above, other modifications or additions are possible. For example, S-TSEQ can be applied for any number of UEs and any number of eNBs. S-TSEQ can be applied for any number of receiver antennas at each eNBs. While the example implementation of FIG. 13 includes an MMSE receiver for stage-1 equalization and an IRC receiver for stage-2 equalization, S-TSEQ can be applied using other linear processing for stage-1 and stage-2. Other implementations are possible.

[0062] The following four approaches of UL-CoMP are described in further detail: Joint Reception in frequency domain (JRFD); Log-Likelihood Ratio (LLR) Exchange (LLR- E); Transport Block Selection Diversity (TBSD); and Two-Stage Equalization (TSEQ). Another approach to interference cancellation is also described as coordinative scheduling with adaptive interference mitigation (CS-AFM).

[0063] In some implementations, the system 100 may adaptively apply different approaches based on coupling between eNBs and UEs. Coupling can include the level of carrier-to-interference ratio (C/I, CIR), also known as the signal-to-interference ratio (S/I or SIR), which is the quotient between the average received modulated carrier power S or C and the average received co-channel interference power I, i.e. cross-talk, from other transmitters than the useful signal. For example, FIG. 2 is a process flow diagram illustrating a method 200 of adaptively and dynamically managing interference.

[0064] At 210, data can be received characterizing carrier to interference ratio for a plurality of user equipment in a multiple-input multiple output (MEVIO) orthogonal frequency division multiplexing (OFDM) network.

[0065] At 220, based on the carrier to interference ratio, an interference reduction approach can be determined. The interference reduction approach can include at least one of: Joint Reception in Frequency Domain (JRFD); Log-Likelihood Ratio (LLR) Exchange (LLR- E); Transport Block Selection Diversity (TBSD); Two-Stage Equalization (TSEQ), and coordinative scheduling with adaptive interference mitigation (CS-AFM).

[0066] At 230, signals received from the plurality of user equipment can be jointly decoded. The decoding may occur, e.g., at system 100, and/or distributed across multiple systems 100 and/or eNodeBs.

[0067] Each approach to interference reduction or cancellation is described below in more detail.

[0068] Joint Reception in Frequency Domain (JRFD).

[0069] In JRFD, the participating CoMP nodes transfer their received samples in frequency-domain (FD) to the master node. The bit rate per node is a function of the number of receiver antennas, the allocation size, and the quantization resolution of the complex- valued frequency-domain samples. Assuming 16 bits for the real and 16 bits for the imaginary parts of a complex -valued FD sample, the bit rate per PRB per subframe is given in the following table:

[0070] Table 2: Bit rate requirement per participating CoMP node per PRB with FD I/Q sample exchange. Quantization 2 Rx Antennas 4 Rx Antennas

Resolution

8 bits per 2* 12* 14* 16 bits/msec = 4* 12* 14* 16 bits/msec =

(real/imaginary) 5.376 Mbps 10.752 Mbps

16 bits per 2* 12* 14*32 bits/msec = 4* 12* 14*32 bits/msec =

(real/imaginary) 10.752 Mbps 21.504 Mbps

[0071] The above calculation shows that with a typical quantization of 32 bits per complex- valued sample, the rate per CoMP node is more than 210 Mbps for an allocation of 10 RBs. For each increase in the allocation size, the rate increases by more than 21 Mbps. The rate requirements double in going from 2 to 4 Rx antennas. The rate requirement may not be functions of the number of users occupying a given PRB, and may also not be a function of the modulation formats used by the users within the CoMP allocation.

[0072] Below is a model for JRFD when two UEs communicate with two eNBs. Note that the notation used below is slightly different from the notation used earlier to describe the symmetric- TSEQ approach. UE-l is connected to eNBl having Nl Rx antennas whereas UE-2 is connected to eNB2 having N2 antennas. UE-l and UE-2 are both cell-edge users. At eNBl, the received signal on tone k when UE-l transmits Sl(k) and UE-2 transmits S2(k) is: y_i(k) = hMSM + h₂ (k)S₂ (k) + l^k In the above, ii(k) is the channel from UE-l to eNBl, /i₂(k) is the channel from UE-2 to eNBl, and ii(k) is the interference plus noise term at eNBl . [0073] Similarly, the received signal at eNB2 is: y₂(fc) = + g₂ (k)s₂ (k) + i₂ (k)

In the above, gi( ) is the channel from UE-l to eNB2, g₂(k) is the channel from UE-2 to eNB2, and 7₂(k) is the interference plus noise term at eNB2. [0074] It can be assumed that the eNB with the smallest number of rx antennas forwards its frequency-domain samples to the other eNB. This makes sense since the overhead increases with the number of Rx antennas. With Nl <= N2, the FD samples from eNB l and eNB2 are then processed at eNB2. Upon stacking the signals ji(k) and j₂(k) together, the equalized signal for UE-l and UE-2 can be obtained in 3 steps. In the first step, the overall covariance matrix can be computed as

Ryy(k> [h?(k) g«(k)]- [h (k) g«(k)]

[0075] In the second step, the weight vectors of the two users can be obtained. For UE-l it can be

whereas for UE-2 it can be w(2,k)=i?-ⁱ(/c)

g₂ (* The equalized symbol for UE-l is

And for UE-2, the equalized symbol is y₂ (k)

[0076] The above calculations reveal that the key to JRFD is accurate estimation of not only the per-tone channel, but also the cross-correlation of the channel gains across the CoMP nodes and the covariance matrices of noise plus interference on each CoMP node.

[0077] Log Likelihood Ratio Exchange (LLR-E)

[0078] In this approach, each CoMP node can perform its own channel estimation, equalization, and LLR generation of the bits within the constellation symbol. The participating CoMP nodes can transmit their LLRs to the master CoMP node, which in turn combines these LLRs (per-user) before performing per-user FEC decoding. The master node also maintains the HARQ buffer for each user across the retransmissions. Unlike the per- sample quantization requirements of FD I/Q sample exchange, the quantization requirements for LLR exchange are moderately use. In some implementations, up to 8 bits are used to quantize an LLR. The number of LLRs can depend on the modulation format of the user; with QPSK we have 2 LLRs, with 16QAM there can be 4 and with 64QAM there can be 6 LLRs. Also the LLR exchange may not be a function of the number of receive antennas and can be a function of the number of users the CoMP node decodes. So, assuming 4 and 8 bits per real-valued LLR and two users per CoMP node, the below table summarizes the rate requirements for LLR exchange:

[0079] Table 3 : Bit rate requirement per participating CoMP node per PRB with LLR exchange

[0080] Since the mapping of bits to a modulation symbol is nonlinear, the LLR generation of a given bit from the equalized symbols is also nonlinear. And, it is generally a hard problem to model the received SF R of a given user with nonlinear processing such as the LLR operation. So, to arrive at a tractable approach to post-processing SFNR, the following equivalent system model can be used. In this model, the equalized symbols of e Bl and e B2, for a given UE, are added, thus mimicking the LLR addition operation of the LLR-E approach.

[0081] Let w(i,j, k) denote the weight vector used by eNB-z to extract the equalized symbol of UE-y on tone k. This weight vector can be expressed as follows

[(h, (k)h (k)+ h₂ (k)h^H ₂ (k) + R_l (k))¹h_J (k) i = 1

w

(_gl (k)g (k)+g₂ (*)g? (k) + R_1A (k)y_gj (k) i = 2

With the above notation, the sum of the equalized symbols of UE-l is:

S, (k) = w^H (1,1, k)y ₁ (k) + w^H (2,1, k)y ₂ (k)

= (w^H (1,1, k)h, (k) + w^H (2,1, k]g, ( , (k) + (w^H (1,1, k)h₂ (k)+w^H (2,1, k)g₂ (k))S₂ (k) + (w^H(\Xk)l₁(k)+w^H(2Xk)l₂(k))

[0082] From the above equation, the effective equalized symbol for UE-l has interference from the symbol of UE-2. Additionally, the non-CoMP interference plus noise contribution from both the eNBs are also present as part of the effective interference plus noise of UE-l.

[0083] The effective post-combining SINR of UE-l can now be written as

In a similar way, the equalized symbol for UE-2 is

S₂ (k) = w^H (1,2, k)y ₁ (k) + w^H (2,2, k)y ₂ (k)

= (w^H (1,2, k)h₂(k)+w^H(2,2,k)g₂ (k))S₂ (k)+(w^H (l,2, k) (k) + w^H (2,2, k)g, ( , (k) + {w^H (1,2, k) (k) + w^H (2,2, k)l₂ (k))

And the corresponding SINR of UE-2 is

[0084] Transport Block Selection Diversity (TBSD) [0085] Here, each node in the CoMP group decodes the two users up to transport block decoding. If the initial transmission is successful at one or more nodes, then no retransmission is scheduled, and the HARQ process ends for that user. On the other hand, if none of the CoMP nodes is able to decode the first transmission, then the retransmission is scheduled for the UE, and this process continues for the additional retransmissions as well.

[0086] There is no CoMP-specific overhead in the packet selection approach. However, the decoded transport blocks of a given UE can be fed back to the eNB that the UE is connected to. Since the maximum MCS index in UL is 23, and the corresponding ITBS index is 21 and with an allocation of 10 RBs for the CoMP users, the TB size is 4968 bits (with ITBS = 21). This means that up to 5 Mbps rate is required to exchange each user's TB to the eNB that UE is connected to. However, if the eNB is able to decode its own UE, then there is no TB exchange involved in this process. Below table illustrates the maximum rate requirements assuming that TB exchange is needed for both the users (i.e., UE 1 is decoded by eNB2 and UE 2 is decoded by eNBl).

[0087] Table 4: Bit rate requirements with transport block selection

[0088] A note on the performance expected from transport block selection: Assume that each eNB has 2 Rx antennas, and each eNB decodes two users each transmitting with single antenna. Since a jointly linear receiver (or joint MMSE receiver) is most likely used (to reduce implementation complexity), after the decoding process, each user can effectively see a single receive antenna (as the second antenna at the eNB is used to suppress the interference due to the other user), and the performance of this scheme at eNB may be worse compared with the JRFD and the LLR-E approaches. However, since the TB error rate at each eNB is spatially independent, the selection diversity nevertheless may introduce some gains. The exact gains of transport block selection are a function of the relative power levels of each UE at each eNB, the channel estimation accuracy, and the type of receiver employed by each eNB.

[0089] The effective SINR of the TBSD can be modeled as follows. As before, let w(i,j, k) denote the weight vector used by eNB-/^' to extract the equalized symbol of UE-y on tone k. This weight vector can be expressed as follows

(h, (k)h (k) + h₂ (*)h? (k) + R _l (k)Y h_J (k) i = 1

w

(_gl (k)g (k) + g₂ (*)g? (k) + R_IA (k)y_gj (k) i = 2

[0090] Let SINR(i,j,k) denote the post-processing SINR of UE-j at eNB-i on tone k. With the ideal MMSE (or adaptive IRC) receiver, an expression for SINR(i,l,k) of UE-l can be given by

Likewise, the post-processing SINR of UE-2 can be

[0091] The post-processing SINR of each UE with TBSD can be

SINR, (k) =

SINR₂ (k) =

[0092] Two-Stage Equalization (TSEQ).

[0093] In the two-stage approach, one of the eNBs performs joint equalization of the two users and transmits the equalized un-sliced symbols and the corresponding noise covariance information to the second eNB. The second eNB combines its own frequency- domain samples with the information received from the first eNB to perform the overall joint detection. To explain this operation in detail, a simple joint equalization algorithm can be assumed at the first eNB; the zero-forcing (ZF) equalizer (ZFEQ).

[0094] With ZFEQ, the received signal at eNB-1 can be expressed as

The output of the ZFEQ can be s(k) = (ll^H(kMk)yil^H(kUk)

= s(k)+ \_l(k)

[0095] In the above, the effective interference plus noise matrix Vi(&) can be colored with the covariance matrix

R _{I I} (k) = (H" (k)u(k)Y H- (k)R _li (k)ll(kill^H (k)u(k)Y

[0096] The above covariance matrix can be of size 2-by-2, and the diagonal elements of this matrix can be strictly positive and the off-diagonal elements can be complex- conjugates of each other. That is, two real numbers and two positive numbers can be needed to describe this matrix. [0097] Quantifying the rate requirements for feeding back the ZFEQ samples to the master eNB is possible. Assuming quantization of the equalized samples is performed using Qi bits per real/imaginary part, whereas quantization of the interference plus noise covariance matrix is performed using Q₂ bits. Note that if the channel varies once per tone then this covariance matrix can be fed back per tone, otherwise the feedback can be relaxed to F times per PRB bandwidth. The below table considers worst case covariance feedback frequency (i.e., once per tone for highly selective channels).

[0098] Table 5 : Bit rate requirement per participating CoMP node per PRB with TSEQ and ZF in the first stage

[0099] Upon receiving s(k) and _VlVl (k) from eNB-1, eNB-2 can perform joint equalization as follows. First, upon stacking the samples of eNB-2 along with the equalized samples of eNB-1, the effective received signal at eNB-2 can be

G(k)s(k) + l(k)

With the above e uivalent signal model, the weight vectors for UE-7 and UE-2 are given by "

With the two-user joint-MMSE receiver, the final equalized symbols of UE-y,y=/,2, is written as

[00100] With the effective channel G(k), the receiver weights Wj(&), and the block-diagonal covariance matrix of I(k), it is then straightforward to obtain the postprocessing SINRs of UE-1 and UE-2.

[00101] Note that it can also be possible to perform joint MMSE equalization in the first stage itself. This approach can be referred to as TSEQ-MMSE (to denote the first stage as MMSE equalization). In this case, the equalized signal at eNB-1 can take the following form: gW = H^ff (* H(*)H^J?(*) ₊ R_iiii( ¹y₁W

= C{k)s{k) + Y₂{k)

[00102] In the above, the matrix C(k) can be the MMSE-specific scaling matrix which can be given by

C(k) = H- (£)(ΐ )Η* (*) ₊ R_/IA ( ¹ H(*) and the covariance of matrix

can be

R_V2V2 (k) = ( )(ΐ )Η* (*) + R_VI (^))^"1 R_/IA ( )(H( )H* (k) + R _l (k)Y H(*)

[00103] Both C(k) and

can be Hermitian 2-by-2 matrices and hence each one may need 4 real parameters to represent them.

[00104] The joint MMSE receiver at eNB-2 can takes the following form. First, a small change in the overall received signal at eNB-2 can be expressed as

= G₀(k)s(k)+ l(k) [00105] With the above equivalent signal model, the weight vector matrix for

UE-7 and UE-2 can be given by

W(*) = k(*) w₂ (*)] = Ο₀ (*)θ? (*) G₀ (k)

o

[00106] And the final equalized symbols of XJE-j,j=l,2, can be written as

*, (*) = ? (*M*)

[00107] The feedback rate to account for the feedback of the scale matrix C(k) can now be given by Table 6: Bit rate requirement per participating CoMP node per PRB with TSEQ and MMSE in the first stage

[00108] Coordinated Scheduling with Adaptive Interference Mitigation (CS-

AEVI).

[00109] CS-AEVI is another approach to interference cancellation that does not use the joint reception paradigm that is typically associated with UL-CoMP. Instead, the schedulers at each eNodeB cooperate in a distributed manner to pair the users such that each eNB cancels the UE that is not connected to it but is interfering with its own users. For example, consider two EnodeB's (eNB l and eNB2) with overlapping cells having, within the overlapping cell region, two user equipment (UEl and UE2) transmitting on the same time- frequency resources. UEl is connected to eNB l and UE2 is connected to eNB2. In order for each eNodeB (eNB 1 and eNB2, respectively) to cancel (or reduce) interference from the non- connected user equipment (UE2 and UEl, respectively), with proper pairing of these two users that differ significantly in the carrier to interference ratio, the eNodeB can use a single user adaptive interference cancellation receiver. Thus, by having each eNodeB cancel interference of the non-connected user equipment, the signals received from the connected user equipment can be decoded.

[00110] Performance Comparison.

[00111] Some results are presented on the comparative performance of four UL-CoMP algorithms considered above, and the relative gains of CoMP over the baseline adaptive IRC receiver.

[00112] Simulation Setup 1.

[00113] First a note on the simulation model: considered is a system with two eNBs and 3 UEs. UE-l is attached to eNB-1 and UE-2 is attached to eNB-2. Additionally, UE-l and UE-2 are in the coverage overlap of these two eNBs, and hence are part of the CoMP group. To model potential interference to the CoMP group, a third UE (UE-3) is introduced whose interference to the users of eNB-1 and eNB-2 can be controlled by its received SNR at each of these two eNBs. This simulation setup is configured by the 6 SNRs, SNR(i,j), where i is the eNB index and j is the UE index and SNR(i,j) is the average received in the absence of any other transmission. Likewise, 6 channel profiles can be used to describe these links. The values of these parameters are listed below.

[00115] Note that all the multipath channels are independent across the UE- eNB links. To further model the dynamic variation of the transmit power due to the user's location within its cell, each UE's transmit power is uniformly varied from -5 dB to +5dB, independently across each subframe.

[00116] The number of transmission time intervals (TTIs) considered for simulation is 10K. Within each TTI, two post-processing SINRs per slot are computed per PRB yielding a total of 4 SINRs per PRB per subframe. These SINRs are aggregated across the TTIs to plot the complimentary cumulative distribution function (CCDF). The SINR corresponding to the 95% CCDF value can be viewed as a performance measure of the cell- edge users (at 5% level).

[00117] With 2 receive antennas at each eNB, the following post-processing SINR values can be seen: Table 8: Cell-edge SINR with 2 receive antennas at each eNB

[00118] With 2 receive antennas, TSEQ-ZF performs as good as JRFD, and the improvement of TSEQ-ZF over LLR-E is 0.85 dB. Around 6.5 dB gain be had with TSEQ- ZF over the CS-AFM receiver.

[00119] With 4 rx antennas at each eNB, the following performance numbers can be observed: Table 9: Cell-edge SINR with 4 receive antennas at each eNB Receiver type Post-processing SINR (dB)

CS-AFM -2.28

TBSD CoMP 0.81

LLR-E CoMP 2.71

TSEQ-ZF CoMP 2.82

TSEQ-MMSE CoMP 3.4

JRFD CoMP 3.4

[00120] The gains are relatively smaller for TSEQ-ZF over LLR-E. However, both of these provide around 2 dB gain over the TBSD CoMP. Further, there is a 5 dB gain possible with TSEQ-ZF over the CS-AFM receiver.

[00121] In the third case, 2 rx antennas at eNB-1 and 4 rx antennas at eNB-2 are considered. The TSEQ-ZF and TSEQ-MMSE are performed at eNB-1. The SFNR performance is listed in the below table at UE-l : Table 10: Cell-edge SINR of UE-l with 2 receive antennas at eNB-1 and 4 receive antennas at eNB-2

[00122] In this case, around 0.65 dB with TSEQ-ZF over the LLR-E can be seen and more than 9 dB gain over the baseline CS-AFM receiver. [00123] The SINR performance for UE-2 is given in the following table: Table

1 : Cell-edge SINR of UE-2 with 2 receive antennas at eNB-1 and 4 receive antennas at eNB- 2

[00124] It can be seen that UE-2 performance can also be improved with TSEQ-ZF by 0.7 dB over the LLR-E approach. The gain of TSEQ-ZF over the baseline CS- AFM receiver is more than 3 dB.

[00125] In the fourth case, 4 rx antennas at eNB-1 and 2 rx antennas at eNB-2 is considered. The TSEQ-ZF and TSEQ-MMSE are performed at eNB-1. The SFNR performance is listed in the below table at UE-l : Table 2: Cell-edge SFNR of UE-l with 4 receive antennas at eNB-1 and 2 receive antennas at eNB-2

[00126] In this case, it is seen that TSEQ-ZF inferior to LLR-E by 0.2 dB.

However, TSEQ-MMSE receiver extracts the full potential of JR, and is better than the baseline CS-AFM receiver by more than 3 dB.

[00127] The SINR performance for UE-2 is given in the following table: Table

3 : Cell-edge SINR of UE-2 with 4 receive antennas at eNB-1 and 2 receive antennas at eNB- 2

[00128] Similar to UE-l 's performance, it is seen that TSEQ-ZF may be inferior to LLR-E by 0.1 dB for UE-2. On the other hand, TSEQ-MMSE may be superior to TSEQ-ZF by 0.8 dB, and may be better than the CS-AFM receiver by more than 9 dB.

[00129] Simulation Setup 2. A set of results is presented under a different set of assumptions on the relative SNRs of each UE as received at each eNB. The channel models are identical to the one shown in Table . The SNRs are set as per following table: Table 4: Average received SNR at each eNB from the UEs UE index eNB-1 eNB -2

1 1 dB -5 dB

2 -5 dB 1 dB

3 O dB O dB

[00130] Note that there are no UE-specific power fluctuations (apart from the fluctuations due to the fading channels) in this part of the simulation study. With 4 receive antennas at each eNB, the CCDF curves of the PP-SINR for each UE are plotted for all the algorithms considered in this paper.

[00131] Figure 3 : CCDF of PP-SF R for UE-1 with 4 Rx Antennas at each eNB. Figure 4: CCDF of PP-SFNR for UE-2 with 4 Rx Antennas at each eNB. It can be seen from Figure 3 and Figure 4 that the TSEQ-MMSE provides performance virtually identical to the optimal JRFD, and TSEQ-ZF is monotonically better than LLR-E and CS-AEVI receivers. The LLR-E has inferior performance compared with the CS-AEVI receiver, for both users, for higher levels of PP-SINR, and the CS-AEVI receiver is as good as the TBSD. These results indicate that it is possible to make the CS-AEVI approach effective in canceling the other-user interference provided the two users are paired such that their SNRs differ by at least 6 dB. For example, pairing the cell-edge user of eNB-1 and the cell-center user of eNB-2 can help the CS-AEVI receiver to decode UE-1 at eNB-1.

[00132] Similar set of conclusions can be drawn for the cases in which eNB-1 having 2 receive antennas and eNB-2 having 2 receive antennas. The PP-SINR CCDF curves of UE-1 and UE-2 are shown in Figure 5 and Figure 6 wherein we see virtually identical performances among the JRFD, TSEQ-ZF and TSEQ-MMSE receiver.

[00133] Figure 5: CCDF of PP-SFNR for UE-1 with 2 Rx Antennas at each eNB. Figure 6: CCDF of PP-SFNR for UE-2 with 2 Rx Antennas at each eNB. Figure 7: CCDF of PP-SINR for UE-1 with 2 Rx Antennas at eNB-1 and 4 Rx Antennas at eNB-2. Figure 8: CCDF of PP-SINR for UE-2 with 2 Rx Antennas at eNB-1 and 4 Rx Antennas at eNB-2. Figure 9: CCDF of PP-SFNR for UE-1 with 4 Rx Antennas at eNB-1 and 2 Rx Antennas at eNB-2. Figure 10: CCDF of PP-SINR for UE-2 with 4 Rx Antennas at eNB-1 and 2 Rx Antennas at eNB-2.

[00134] When eNB-1 has 2 receive antennas and eNB-2 has 4 receive antennas, it can be seen that the CS-AEVI receiver is less effective in canceling UE-2 at eNB-1. This can be seen from Figure 7 for UE-1 wherein the PP-SFNR of UE-1 is almost always better with LLR-E compared with the CS-AEVI approach. On the other hand, due to eNB-2 having 4 receive antennas, as shown in Figure 8, the CS-AEVI is able to cancel the interference of UE-1 at eNB-2, and thus making this approach more attractive compared with the LLR-E approach. The situation, and the resulting conclusions, simply reverse when eNB-1 has 4 receive antennas and eNB-2 has 2 receive antennas. In this case, as shown in Figure 9 and Figure 10, CS-AEVI is less effective for canceling UE-1 interference at eNB-2, whereas it is more effective compared with the LLR-E for canceling UE-2 interference at eNB-1.

[00135] In some implementations, the current subject matter can be configured to be implemented in a system 1100, as shown in FIG. 11. The system 1100 can include one or more of a processor 1110, a memory 1120, a storage device 1130, and an input/output device 1140. Each of the components 1110, 1120, 1130 and 1140 can be interconnected using a system bus 1150. The processor 1110 can be configured to process instructions for execution within the system 1100. In some implementations, the processor 1110 can be a single-threaded processor. In alternate implementations, the processor 1110 can be a multithreaded processor. The processor 1110 can be further configured to process instructions stored in the memory 1 120 or on the storage device 1130, including receiving or sending information through the input/output device 1140. The memory 1120 can store information within the system 1100. In some implementations, the memory 1120 can be a computer- readable medium. In alternate implementations, the memory 1120 can be a volatile memory unit. In yet some implementations, the memory 1 120 can be a non-volatile memory unit. The storage device 1130 can be capable of providing mass storage for the system 1100. In some implementations, the storage device 1130 can be a computer-readable medium. In alternate implementations, the storage device 1130 can be a floppy disk device, a hard disk device, an optical disk device, a tape device, non-volatile solid state memory, or any other type of storage device. The input/output device 1140 can be configured to provide input/output operations for the system 1100. In some implementations, the input/output device 1140 can include a keyboard and/or pointing device. In alternate implementations, the input/output device 1140 can include a display unit for displaying graphical user interfaces.

[00136] The systems and methods disclosed herein can be embodied in various forms including, for example, a data processor, such as a computer that also includes a database, digital electronic circuitry, firmware, software, or in combinations of them. Moreover, the above-noted features and other aspects and principles of the present disclosed implementations can be implemented in various environments. Such environments and related applications can be specially constructed for performing the various processes and operations according to the disclosed implementations or they can include a general-purpose computer or computing platform selectively activated or reconfigured by code to provide the necessary functionality. The processes disclosed herein are not inherently related to any particular computer, network, architecture, environment, or other apparatus, and can be implemented by a suitable combination of hardware, software, and/or firmware. For example, various general-purpose machines can be used with programs written in accordance with teachings of the disclosed implementations, or it can be more convenient to construct a specialized apparatus or system to perform the required methods and techniques. [00137] The systems and methods disclosed herein can be implemented as a computer program product, i.e., a computer program tangibly embodied in an information carrier, e.g., in a machine readable storage device or in a propagated signal, for execution by, or to control the operation of, data processing apparatus, e.g., a programmable processor, a computer, or multiple computers. A computer program can be written in any form of programming language, including compiled or interpreted languages, and it can be deployed in any form, including as a stand-alone program or as a module, component, subroutine, or other unit suitable for use in a computing environment. A computer program can be deployed to be executed on one computer or on multiple computers at one site or distributed across multiple sites and interconnected by a communication network.

[00138] As used herein, the term "user" can refer to any entity including a person or a computer.

[00139] Although ordinal numbers such as first, second, and the like can, in some situations, relate to an order; as used in this document ordinal numbers do not necessarily imply an order. For example, ordinal numbers can be merely used to distinguish one item from another. For example, to distinguish a first event from a second event, but need not imply any chronological ordering or a fixed reference system (such that a first event in one paragraph of the description can be different from a first event in another paragraph of the description).

[00140] The foregoing description is intended to illustrate but not to limit the scope of the invention, which is defined by the scope of the appended claims. Other implementations are within the scope of the following claims.

[00141] These computer programs, which can also be referred to programs, software, software applications, applications, components, or code, include machine instructions for a programmable processor, and can be implemented in a high-level procedural and/or object-oriented programming language, and/or in assembly/machine language. As used herein, the term "machine-readable medium" refers to any computer program product, apparatus and/or device, such as for example magnetic discs, optical disks, memory, and Programmable Logic Devices (PLDs), used to provide machine instructions and/or data to a programmable processor, including a machine-readable medium that receives machine instructions as a machine-readable signal. The term "machine-readable signal" refers to any signal used to provide machine instructions and/or data to a programmable processor. The machine-readable medium can store such machine instructions non- transitorily, such as for example as would a non-transient solid state memory or a magnetic hard drive or any equivalent storage medium. The machine-readable medium can alternatively or additionally store such machine instructions in a transient manner, such as for example as would a processor cache or other random access memory associated with one or more physical processor cores.

[00142] To provide for interaction with a user, the subject matter described herein can be implemented on a computer having a display device, such as for example a cathode ray tube (CRT) or a liquid crystal display (LCD) monitor for displaying information to the user and a keyboard and a pointing device, such as for example a mouse or a trackball, by which the user can provide input to the computer. Other kinds of devices can be used to provide for interaction with a user as well. For example, feedback provided to the user can be any form of sensory feedback, such as for example visual feedback, auditory feedback, or tactile feedback; and input from the user can be received in any form, including, but not limited to, acoustic, speech, or tactile input.

[00143] The subject matter described herein can be implemented in a computing system that includes a back-end component, such as for example one or more data servers, or that includes a middleware component, such as for example one or more application servers, or that includes a front-end component, such as for example one or more client computers having a graphical user interface or a Web browser through which a user can interact with an implementation of the subject matter described herein, or any combination of such back-end, middleware, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication, such as for example a communication network. Examples of communication networks include, but are not limited to, a local area network ("LAN"), a wide area network ("WAN"), and the Internet.

[00144] The computing system can include clients and servers. A client and server are generally, but not exclusively, remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other.

[00145] The implementations set forth in the foregoing description do not represent all implementations consistent with the subject matter described herein. Instead, they are merely some examples consistent with aspects related to the described subject matter. Although a few variations have been described in detail above, other modifications or additions are possible. In particular, further features and/or variations can be provided in addition to those set forth herein. For example, the implementations described above can be directed to various combinations and sub-combinations of the disclosed features and/or combinations and sub-combinations of several further features disclosed above. In addition, the logic flows depicted in the accompanying figures and/or described herein do not necessarily require the particular order shown, or sequential order, to achieve desirable results. Other implementations can be within the scope of the following claims.

Claims

WHAT IS CLAIMED IS:

1. A receiver compri sing :

a first equalizer receiving a signal representing at least one symbol transmitted on an orthogonal frequency division multiplexing (OFDM) network by a first user equipment (UE) and at least one additional UE, the transmission by the first UE to the receiver, the first equalizer computing, based on the signal, at least one first partially-equalized symbol of the first UE; and

a second equalizer communicatively coupled to the first equalizer, receiving the at least one first partially-equalized symbols of the first UE, receiving at least one second partially-equalized symbol, and computing at least one equalized-symbol of the at least one first UE, the at least one second partially-equalized symbol received from at least one additional base station being part of a coordinated multipoint (CoMP) group, the at least one second partially-equalized symbol characterizing the signal transmitted on the OFDM network by the first UE and as received by the at least one additional base station.

2. The receiver of claim 1, wherein the first equalizer implements joint equalization.

3. The receiver of claim 1, wherein the second equalizer implements adaptive interference-rejection combining (IRC).

4. The receiver of claim 1, the first equalizer computing at least one equalized-symbol of the one or more additional UEs for provision to the at least one additional base station being part of the CoMP group.

5. The receiver of claim 1, wherein the at least one symbol transmitted on the OFDM network by the first UE and at least one symbol transmitted on the OFDM network by the at least one additional UE share time and frequency resources.

6. The receiver of claim 1, the second equalizer processing contributions to the signal from the additional UE as interference when it computes the equalized-symbols of the first UE.

7. A method for implementation by a first base station belonging to a coordinated multipoint transmission (CoMP) group, the CoMP group having at least one other CoMP member base station, the method comprising:

receiving data characterizing at least one partially-equalized first symbol by the first base station and of a transmission from a first user equipment (UE) to the first base station as received by the first base station;

receiving data characterizing at least one partially-equalized second symbol by the at least one other CoMP member base station and of the transmission from the first UE to the first base station as received by the at least one CoMP member base station; and

determining, using the received data, at least one equalized symbol for the transmission from the first UE, wherein contributions from additional UE to the transmission from the first UE to the first base station are processed as interference;

wherein the receiving data characterizing partially-equalized first symbols, the receiving data characterizing partially-equalized second symbols, and the determining are performed by at least one data processor forming part of at least one computing system of the first base station.

8. The method of claim 7 wherein the transmission from the first UE to the first base station and a transmission from the additional UE to the at least one other CoMP member base station share time and frequency resources.

9. The method of claim 7 further comprising:

receiving a second transmission from one of the additional UE, the second transmission to one of the at least one other CoMP member base station.

10. The method of claim 9 further comprising:

partially-equalizing symbols of the received transmission from the one of the at least one additional UE.

11. The method of claim 10 wherein partially-equalizing symbols is performed using j oint-equalization.

12. The method of claim 9 further comprising: transmitting, to one of the at least one other CoMP member base stations, the partially-equalized symbols.

13. The method of claim 7 further comprising exchanging auxiliary channel information of the at least one additional UEs with the at least one CoMP member base station.

14. The method of claim 7 wherein the partially-equalized second symbols are received from the at least one CoMP member base station on a per time-frequency resource basis.

15. The method of claim 7 further comprising:

exchanging complex valued scale factors per demodulated reference symbol (DMRS) tone.

16. The method of claim 7 further comprising:

computing complex valued scale factors as beamformed channel characteristics.

17. The method of claim 7 further comprising:

exchanging noise variance values per demodulated reference signal (DRMS), wherein the noise variance values are of effective noise plus interference.

18. The method of claim 7 further comprising:

computing noise variance values of effective noise plus interference, the noise variance values computed as a function of beamforming values, channel noise covariance, and channel characteristics.

19. The method of claim 7 further comprising:

providing the determined equalized symbols, wherein providing includes transmitting, storing, or processing.

20. A non-transitory computer program product storing instructions, which when executed by at least one data processor of at least one computing system, implements a method, the at least one computing system of a first base station belonging to a coordinated multipoint transmission (CoMP) group, the CoMP group having at least one other CoMP member base station, the method comprising:

receiving data characterizing at least one partially-equalized first symbol by the first base station and of a transmission from a first user equipment (UE) to the first base station as received by the first base station;

receiving data characterizing at least one partially-equalized second symbol by the at least one other CoMP member base station and of the transmission from the first UE to the first base station as received by the at least one CoMP member base station; and

determining, using the received data, at least one equalized symbol for the transmission from the first UE, wherein contributions from additional UE to the transmission from the first UE to the first base station are processed as interference.