WO2014071210A1

WO2014071210A1 - Coding schemes for cognitive overlay radios

Info

Publication number: WO2014071210A1
Application number: PCT/US2013/068105
Authority: WO
Inventors: Andrea Goldsmith; Ernest KURNIAWAN; Stefano RINI
Original assignee: The Board Of Trustees Of The Leland Stanford Junior University
Priority date: 2012-11-02
Filing date: 2013-11-01
Publication date: 2014-05-08

Abstract

Methods and apparatus, including computer program products, are provided for decoder/encoder selection in cognitive radios. In one aspect there is provided a method, which may include receiving an indication of whether an interference in a cognitive radio network is at least one of a weak interference regime, a very strong interference regime, and a primary decodes cognitive interference regime; selecting, when the interference comprises the weak interference regime, a pre-canceling decoder to cancel interference caused by a primary transmitter at a secondary receiver; selecting, when the interference comprises the very strong interference regime, a first joint decoder to decode a first message superposed on a second message; and selecting, when the interference comprises the primary decodes cognitive interference regime, a second joint decoder to decode the first message and the second message, the first message having a pre-canceling code decodable by the second joint decoder.

Description

CODING SCHEMES FOR COGNITIVE OVERLAY RADIOS

CROSS-REFERNCE TO RELATED APPLICATIONS

[001] This application claims priority under 35 U.S.C 119(e) to U.S. Provisional Patent Application Serial No. 61/722,079, filed November 2, 2012, titled, "PRACTICAL CODING SCHEMES FOR COGNITIVE OVERLAY RADIOS." Priority of the filing date of the Provisional Patent Application is hereby claimed. The disclosure of the Provisional Patent Application is incorporated by reference herein in its entirety.

FIELD

[002] The subject matter disclosed herein relates to cognitive radios and, in particular, coding.

BACKGROUND

[003] Cognitive radio is widely considered an enabling technology for increasing the spectral efficiency of wireless networks. An underlying principle of cognitive radio is to allow a set of cognitive users to access the spectrum that belongs to the primary users (also known as the licensed users), without compromising the primary users' link quality. For the cognitive overlay radios, this is achieved through cooperation, whereby the primary user is willing to grant access to the medium under the conditions that the cognitive user assists its transmission, allowing for faster and more reliable communication.

SUMMARY

[004] Methods and apparatus, including computer program products, are provided for coding schemes which can be used for cognitive overlay radios.

[005] In some example embodiments, there is a method. The method may include receiving an indication of whether an interference in a cognitive radio network is at least one of a weak interference regime, a very strong interference regime, and a primary decodes cognitive interference regime; selecting, when the interference comprises the weak interference regime, a pre-canceling decoder to cancel interference caused by a primary transmitter at a secondary receiver; selecting, when the interference comprises the very strong interference regime, a first joint decoder to decode a first message superposed on a second message; and selecting, when the interference comprises the primary decodes cognitive interference regime, a second joint decoder to decode the first message and the second message, the first message having a pre-canceling code decodable by the second joint decoder.

[006] In some implementations, the above-noted aspects may further include additional features described herein including one or more of the following. The weak interference regime may represent the interference originating primarily from the primary transmitter, rather than a secondary transmitter. A cognitive radio may include the secondary transmitter, the cognitive radio having access to the second message being transmitted by the primary transmitter. The very strong interference regime may represent the interference originating primarily from a secondary transmitter, rather than the primary transmitter. The primary decodes cognitive interference regime may represent substantial interference from both the primary transmitter and a secondary transmitter. The pre- canceling decoder may include a dirty paper decoder. The second message may originate at the primary transmitter, and the first message may originate at a secondary transmitter. A cognitive radio may include the secondary receiver, the secondary receiver may implement the receiving, the selecting the pre-canceling decoder, the selecting the first joint decoder, and the selecting the second joint decoder.

[007] It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive. Further features and/or variations may be provided in addition to those set forth herein. For example, the implementations described herein may be directed to various combinations and subcombinations of the disclosed features and/or combinations and subcombinations of several further features disclosed below in the detailed description.

DESCRIPTION OF THE DRAWINGS

[008] 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 subject matter disclosed herein. In the drawings,

[009] FIGs. 1A-1B depict block diagrams of examples of cognitive radio systems, in accordance with some example embodiments;

[010] FIG. 2 depicts a plot of an example of interference regime, in accordance with some example embodiments;

[011] FIG. 3 depicts a flow chart of an example process for selecting coders based on the interference regime, in accordance with some example embodiments;

[001] FIG. 4 depicts a block diagram of an example of a dirty paper encoder;

[012] FIG. 5 depicts a block diagram of an example of a dirty paper decoder;

[013] FIG. 6 depicts a portion of a multi-user joint decoder, the portion may be used as the soft in, soft out a priori probability decoder (SISO app) depicted at FIG. 7;

[014] FIG. 7 depicts an example implementation of a multi-user joint decoder which can be used at a primary receiver and/or a secondary receiver;

[015] FIG. 8 depicts a block diagram of an example of a dirty paper decoder and irregular repeat accumulate joint decoder for the primary receiver in the primary decodes cognitive receiver; and

[016] FIG. 9 depicts an example of a radio, in accordance with some example embodiments. [017] Like labels are used to refer to same or similar items in the drawings.

DETAILED DESCRIPTION

[018] The subject matter described herein relates to coding schemes for the cognitive radio networks including cognitive overlay radios, such as primary and secondary radios. In these cognitive radio systems, a primary transmitter and primary receiver pair are willing to share spectrum with a secondary transmitter and secondary receiver pair.

Moreover, the secondary transmitter includes, in some example embodiments, knowledge regarding the information being transmitted by the primary transmitter. The secondary transmitter's knowledge may comprise some (if not all) of primary transmitter's message (or, for example, symbols). This non-causal knowledge may be obtained when the primary user's message is public, such as a TV program in a broadcasting network, and/or when the secondary transmitter is provided access to the message before transmission.

[019] FIG. 1A depicts an example of a cognitive radio overlay system 100 including a secondary transmitter/receiver pair 103A-B and a primary transmitter/receiver pair 105A-B. Secondary transmitter 103 A and secondary receiver 103B may also be referred to as a cognitive transmitter and a cognitive receiver, respectively.

[020] In the example of system 100, primary transmitter 105 A may send a message, such as codeword X₂ 106A, via a communication channel, such as a cognitive interference channel (CIC) 150, to primary receiver 105B, which is received as Y₂. While secondary transmitter 103 A sends a message, Xi 106C, via cognitive interference channel 150 to secondary receiver 103B, which is received as Yi. Secondary transmitter 103 A may have anticipatory access to the information, X₂ 106B, being transmitted by primary transmitter.

[021] FIG. IB depicts cognitive radio overlay system 100 with additional details with respect to the cognitive interference channel (CIC) 150. The CIC 150 (which may be Gaussian) depicted at FIG. IB may include a secondary transmitter 103 A (Si), a secondary receiver 103B (Di), a primary transmitter 105 A (S₂), and primary receiver 105B (D₂) sharing the same time-frequency resources at CIC 150 to transmit/receive information, such as messages provided to the encoders at the transmitters and decoders at the receivers, Xi, X₂, and the like.

[022] Given for example a CIC channel 150 having for example unitary gain, the received signals Yi and Y₂ at secondary receiver 103B (Di) and primary receiver 105B (D₂) may be determined based on the following discrete time signal model:

wherein X_j (e.g., Xi and X₂) is a transmitted codeword from a transmitter S_j (e.g., Si and S₂) which satisfy a power constraint E[\X_j | ] < P_j, and Z_{ denotes the additive Gaussian noise at D{ with variance Ni; P_s is the transmit power used at transmitter S_j. The interference channel gains a and b (for example, ^ ^{δ Λ ai SM} ^ ) respectively may be assumed to be known by all nodes in the network 100; R denotes a set of real numbers, and ' denotes the set of positive real numbers. The non-causal knowledge of the primary transmitter's message, X₂, may be used by the cognitive (or secondary) transmitter 103 A to allocate a fraction of its power to assist the primary transmitter 105 A. With this cooperative strategy and denoting 0 < a < 1 as a fraction of transmit power used by the secondary transmitter 103 A to transmit its own message X_{l s} the received signal model for Yi and Y₂ can be expressed as follows:

<2bi wherein ·'^*> - i '^Λ

[023] In some example embodiments, a capacity of the cognitive interference channel 150 may be characterized as being in a weak interference regime, a very strong interference regime, and a primary decodes cognitive regime. The capacity may represent the largest set of rates that can be achieved with a vanishing probability of error.

[024] In the weak interference regime, capacity may be realized by treating interference (e.g., caused by link b and also referred to as secondary interference at FIG. IB) at the primary receiver 105B as noise with respect to decoding. Moreover, the interference (e.g., caused by link a and also referred to as primary interference at FIG. IB) at the secondary receiver 103B may be pre-canceled by a coder at the secondary transmitter 103 A. This pre-cancelation is possible due to secondary transmitter 103A's access to the message being transmitted by primary transmitter 105 A (which causes interference through link a). In some example embodiments, the secondary transmitter-receiver pair 103A/B may include a pre-canceling coder/decoder, such as a dirty paper coder/decoder, and the primary transmitter-receiver pair 105A/B may use another coder, such as a point-to-point coder/decoder that does not take into account interference. In the very strong interference regime, capacity may be attained by having both received messages Xi and X₂ decoded by the primary receiver 105B and the decoder at secondary receiver 103B. For example, the primary transmitter 105 A transmits its message, X₂. However, the secondary transmitter 103 A may perform superposition coding to superpose the secondary message, Xi, over the primary message, X₂. At the receivers 103B/105B, joint decoding of both messages is performed.

[025] In the primary decodes cognitive regime, capacity may be realized with a mixture of the coding techniques that achieve capacity in the weak and very strong interference regimes. At the transmitters, secondary transmitter 103 A implements dirty paper coding and the primary transmitter 105 A implements a point-to-point coder. At the receivers 105B/103B, joint decoding is used. At the primary receiver 105B, it encounters a combination of the intended codeword/message, X₂, and a dirty paper codeword, Xi, sent by secondary transmitter 103 A to pre-cancel the secondary codeword against the interference experienced at the cognitive receiver 103B.

[026] FIG. 2 depicts an example plot characterizing whether the interference regime is weak, very strong, or primary decodes cognitive. For channels outside the colored area, either strategy for the very strong or primary decodes cognitive can be used, but optimality may not be guaranteed. The horizontal axis represents absolute values of b, which represents the secondary transmitter's 103 A interference gain at the primary receiver 105B as depicted at FIG. IB and Equations lb and 2b. The vertical axis represents the value of a, which represents the primary transmitter's 105 A interference gain at the secondary receiver 103B as depicted at FIG. IB and Equations la and 2a. As such, estimates for a and b may be used, in some example embodiments, to assess whether the regime is weak, very strong, or primary decodes cognitive. Moreover, the determined regime may be used to select a coder(s) and/or a decoder(s) to be used at secondary transmitter/receiver pair 103A-B and/or primary transmitter/receiver pair 105A-B.

[027] FIG. 3 depicts an example process 300 for selecting coders in a cognitive overlay network 300, in accordance with some example embodiments.

[028] At 305, a determination may be made regarding the interference regime of the CIC 150, in accordance with some example embodiments. For example, a processor at one or more of the radios 103A-B/105A-B may determine whether the interference regime at the cognitive interference channel 150 can be characterized as a weak interference regime, a very strong interference regime, and/or a primary decodes cognitive regime. This determination may be performed by for example calculating estimates for the power of the interference associated with the secondary transmitter on the primary receiver (for example, b, b , and/or Pi as shown in Equations lb and 2b) and/or the interference associated with the primary transmitter on the secondary receiver (for example, a, a , P₂ as shown in Equations la and2a). In some implementations, the processor may estimate the values of a and/or b as noted above with respect to Equations la-2b and FIG. 2. For example, the values of a or b for a fixed power at the transmitters P I and P_2 may map to a region in the plot of FIG. 3 indicating a weak interference, a very strong interference, and/or a primary decodes cognitive regime.

[029] When the results of the determination at 305 indicate weak interference, a coder may be selected for use at secondary transmitter to enable pre-cancelation of the interference caused by the primary transmitter 105 A at the secondary (or cognitive) receiver 103B (yes at 310 and 315). An example of a coder that can provide the pre-cancelation is a dirty paper coder (DPC), although other types of coders, such as deterministic coders, Tomlinson Harashima precoder, and vector perturbation techniques may be used as well.

[030] To illustrate, when in the weak interference regime, the direct link from secondary transmitter 103 A (Si) to secondary receiver 103B (Di) is more capable than the interference link, b, from secondary transmitter 103 A (Si) to primary receiver 105B (i¾). This weak interference condition may correspond to the following equation: wherein N₂ is the noise variance at the primary receiver 105B (D₂), and Ni is the noise variance at the secondary receiver 103B (Di).

[031] The DPC coding used at 315 may include secondary transmitter 103A (Si) pre-canceling, via the DPC, the total interference caused by the primary user's message X₂ at secondary receiver 103B (Di), while the primary receiver 105B (D₂) treats the interference as noise. The decoding at the cognitive/secondary receiver 103B (Di) may be performed using an interference pre-cancelling decoder, such as a dirty paper decoder. This pre-cancelling scheme may exploit the non-causal/anticipatory interference knowledge at the secondary transmitter to facilitate the transmission of the intended message in accordance with DPC at 315. [032] A deterministic implementation of DPC may include interference quantization and channel coding adapted to the quantization output, although other DPC implementations and other types of pre-canceling coders may be used as well. This DPC implementation for the cognitive/secondary transmitter 103 A may include a combination of Trellis Coded Quantization (TCQ) and Irregular Repeat Accumulate (IRA) codes (which may be used to achieve shaping and/or coding gain). The primary transmitter 105 A (which may have no knowledge of the interference) may also employ at 315 an Irregular Repeat Accumulate (IRA) code. The structure of IRA code (specifically the use of accumulator as an inner code) enables a computationally simple calculation of the posteriori probability over the combined trellis at the decoder. Although the previous example describes use of the IRA code, channel coding schemes other than IRA code may be used as well.

[033] FIG. 4 depicts an example block diagram of an example of a DPC 400 which can be used as interference pre-cancelling encoder at a transmitter, such as at the secondary transmitter 103 A, when in the weak regime as well as other regimes as well.

[034] Given the interference sequence *^' = («^■»· _v « ^? from the transmitter (which is known non-causally at the cognitive/secondary transmitter 103 A), the encoder therein may generate + D using a multiplication factor λ and a pseudo random dither sequence D, and forward the generated S + D to the input of a trellis coded quantizer (TCQ). Here, the dither sequence D is generated through a pseudo random process that is available at both the transmitter and receiver. The K-bits message, u, to be transmitted is first encoded using a non-systematic rate- 1/4 IRA code into m, and then used to determine the coset of the TCQ codebook. The TCQ may be performed using the Viterbi algorithm over a rate- 1/2 convolutional code with modulo distance metric and output alphabet {0,1,2,3} corresponding to the two-bits output. The coset codebook is obtained by applying a shift of 0 < Δ < 1 (a tunable parameter) to the convolutional code output. The TCQ codeword obtained in this manner, Λ, satisfies Λ ;: A ^'iK -, where the set A = {0, Δ, 1,1 + Δ, 2,2 + Δ, 3,3 + Δ} . The actual transmitted sequence Xi is then the quantization error between ( S + D) mod 4 and the TCQ output codeword Λ. Following the lattice property of TCQ, the sequence Xi is approximately Gaussian.

[035] Denoting |A£ ^·¾^· D, m _rco as the quantization operation of XS + D over the TCQ coset codebook specified by m, the transmitted sequence Xi may be expressed as follows:

V : :=== M l - aS : D}_{sili;(i i}

[036] Upon receiving Y_\ = X_\ + S + Z_\, the receiver calculates

which is equivalent to a transmission of the TCQ coset codeword Λ in the presence of additive Gaussian noise Z over a modulo channel. To minimize the effective noise variance

N, λ is set to the minimum mean square error scaling ... -· ^{; ';} · '.which may coincides with Costa's DPC scaling.

[037] To recover the message u, a decoder may be used at for example the secondary receiver 103B, when in the weak regime and the like. FIG. 5 depicts an example implementation of a decoder 500, such as a dirty paper decoder, which can be used at a receiver to recover message, u, at secondary receiver 103B. Specifically, to recover the message u, the coset of Λ is identified as specified by m. From Equation (5), the likelihood value that the i^th symbol of the TCQ codeword Λ [i] takes on a specific value ¾ - given the i^th observed symbol Y [i] is

[038] Due to the properties of the modulo channel, the first numerator term ; ί._{Γ !} above may be calculated as follows:

wherein the approximation is performed by taking only the most significant element in the summation. The second term Pr{Aj«f ^::::: <t\ represents the a priori probability of A[z] = q, which (for a given trellis state) is directly related to the a priori of the bit at the accumulator

(ACC) input, while the normalization term P ^' !Ai.) ensures that .·,. . : ^Py i ------- -------- *

[039] The likelihood ^{Ρτ ( Λ} i^{f "}' > may then be used to calculate the symbol maximum a posteriori probability of the ACC input sequence w. This may be performed using BCJR algorithm, which accumulates the likelihood of all the symbols in the codeword sequence according to the structure of the combined trellis of TCQ and the ACC trellis of the IRA code. The decoding of the message sequence u can then be performed using the IRA sum-product algorithm through several iterations of extrinsic information exchange with the BCJR decoder.

[040] Referring again to FIG. 3, when 305 results in a determination of a very strong interference, a second coder, such as superposition coder (SPC), may be selected to enable both the primary receiver 105B and the secondary receiver 103B to decode both messages transmitted by the primary and secondary transmitters 103 A and 105 A (yes at 317 and 319).

[041] The very strong interference regime may correspond to the regime in which the capacity of the channel reduces to the capacity of the compound multiple access channel. In this regime, each decoder (at primary receiver 105B and secondary receiver 103B) decodes all the messages, Yi and Y₂, in the network 100. For the CIC 150, this very strong interference regime may be expressed by the following inequalities: [042] For the above condition to hold irrespective of the value of power splitting parameter a, the channel parameters may need to satisfy the following:

[043] At 319, the secondary transmitter 103 A may implement superposition coding (SPC) where all codewords are superposed on top of each other, and the primary receiver 105B and cognitive/secondary receiver 103B may decode both received messages, Yi and Y₂ (yes at 317 and 319). At the transmitter's coders, X_\ and ₂ may be generated according to a channel code, such as an IRA code, and the cognitive/secondary transmitter 103 A {S_\) may send a weighted sum of the two codewords X_\ and ₂ according to the power splitting parameter a.

[044] The received signals at the primary receiver 105B and cognitive/secondary receiver 103B may be expressed as noted above with respect to Equations (2a) and (2b), and take the form of the following equation:

Y = <·: V; r <¾¾ ^; ;r.

wherein c_\ and c₂ are the effective gain of X_\ and ₂, respectively. Since both X_\ and ₂ are IRA codewords, a joint decoding technique may be used to combine the trellises of the ACC from the IRA code used at both the cognitive and the primary encoders (e.g., the encoders at the primary and secondary transmitters).

[045] FIG. 6 depicts an example of a Trellis diagram of (a) a single accumulator (ACC) for the decoding of a single IRA codeword and (b) a combination of two ACCs for decoding of a superposition of two IRA codewords. Given that in this example both IRA codewords employ binary phase shift keying (BPSK), the term ci i + c_{2 2} may be regarded as a super-symbol ψ drawn from a size-4 alphabet ~~ i^±c ν^'*^*» ± <- -*^'¾.^? · The likelihood value may then be calculated with respect to this super-symbol rather than the individual codeword bit. Given the i^th observed symbol Y[i], the likelihood that y [i] takes on a particular value q E B in accordance with the following:

Fr { \ί\ ::::: ΐίΠ'Ίΐϊ % = ' ^{ί ' i} ' ^{~ ~} ^ — . i W)

[046] In a similar manner, the normalization term I ensures that

∑-,;::. ·- - ! ^"" !*J,^{! ::::} ^ Since Z is Gaussian with variance N, the term i^¾¾bH^fS ¾) may be calculated as follows:

. exp( -- f YUi -- e?}^" /3ΛΠ . f 1 )

[047] For a given trellis state, instead of one bit as in the case of decoding a single IRA code, the a priori probability vr ¾? > may be determined by the a priori of the bits at the ACC input of both IRA encoders. As an example consider the combined trellis FIG. 6(b) at state 0, the a priori probability ψ) ^~ «*vft may be given by the product of two a priori probabilities: i t an -~ Fri¾¾— 0)- The calculation of the a posteriori probability of the bits at the ACC inputs may then be performed using the generalization of the BCJR algorithm, which is explained in the following. For a given edge e on the trellis,

S E

denote s (e) and s (e) as its starting and ending state; and w\(e), w₂(e), and ψ(ε) as the corresponding w\, w₂, and ψ values, respectively. Define the edge transition probability at the i^th stage as follows: wherein ξ is the normalization factor. The letter / indicates that the probabilities act as the input to the algorithm. Note that for a given edge, Equation (10) and (12) are equivalent. The following is also defined:

[048] The initial value of 1 is set for the initial and final state of Ao( ) and B_K(s), respectively; and value 0 is used for other states. The a posteriori probability of H : tm t i J .2 _may be calculated as follows:

P: {Wt hi ^ $; O) ξ' A_s~. ; j) H-i*. } wherein ξ' is another normalization factor, and the letter O signifies it as an output. The corresponding extrinsic Log Likelihood Ratio (LLR) of W_t[i] may then be calculated as follows:

[049] The obtained extrinsic LLR is then passed to the corresponding sum-product decoder of its respective IRA, and several iterations of extrinsic information exchange are performed before a hard decision is made.

[050] FIG. 7 depicts an example implementation of a joint decoder implementation which is in accordance with the aforementioned description and can be used at primary receiver 105B and/or secondary receiver 103B. Referring to FIG. 7, the received signal may be decoded by first processing it through a likelihood calculator to compute the likelihood value of the super symbol. At the SISO APP block, this likelihood value together with the a priori probability of the bits at the ACC inputs of both IRA codes (which are initialized to zero at the beginning of the iteration), will then be used to calculate the a posteriori probability using the generalized BCJR algorithm on the combined trellis. The extrinsic LLR is then exchanged with the corresponding sum-product IRA decoder block for several iterations before the hard decision on the desired message is finally made.

[051] Referring again to FIG. 3, when 305 results in a determination of primary decodes cognitive (yes in 321 and 323), the secondary transmitter 103 A may implement dirty paper coding, and the primary transmitter 105 A may implement a point-to-point coder (which is similar to the weak regime noted above). The secondary receiver 103B may implement DPC decoder just as in the weak regime. However, the primary receiver 105B may select joint decoding instead of treating the interference as noise. Here, the received signal at the primary receiver 105B comprises of a superposition coding of the intended message codeword and the dirty paper codeword from the secondary transmitter 103 A. Since the dirty paper codeword is designed to pre cancel the interference caused by the message codeword for the primary receiver 105B, decoding it would help the primary receiver better decode its intended message codeword.

[052] The set of channel parameters in the primary decodes cognitive regime may satisfy the follow conditions: s!id f{ a) > 0, ¥0 a < } vfh&s /(a)—

[053] A transmission strategy for the primary decodes cognitive regime may be treated similar to that of the weak interference regime for the cognitive user, wherein the cognitive/secondary transmitter 103 A (Si) applies DPC against the total interference caused by X₂ at cognitive/secondary receiver 103B (Di), while the primary transmitter 105 A (S₂) implements conventional channel coding to generate X₂. The decoding at cognitive/secondary receiver 103B (Di) may also be performed using the DPC decoder to achieve interference-free capacity (or substantially so). For this reason, the cognitive/secondary transmitter's decoder may be the same, or similar to, the weak interference regime.

[054] At the primary receiver 105B (D₂), a joint decoder may be used in the primary decodes cognitive regime, and the joint decoder may be implemented using a similar approach as for the joint decoding in the very strong interference regime.

[055] To simplify some of the mathematical expressions, denote the following:

[056] FIG. 8 depicts a block diagram of an example of a DPC and IRA joint decoder 800. Even though decoding u₂ is of primary interest, the sum-product decoder of ui may also be used to refine the a priori probability at each decoding iteration. In terms of decoding flow, it is similar to the joint decoder in the very strong regime in FIG. 7. Namely, the received signal Y₂ is first used to calculate the likelihood value of the super symbol. The following SISO APP block will then use the likelihood value to compute the a posteriori probability of the bits at the ACC inputs of the IRA codes using generalized BCJR on the combined trellis. Finally, after several iterations of extrinsic information exchange with the corresponding sum-product IRA decoder block, a hard decision is made. The main difference between this joint decoder and the one used for the very strong regime is on how the likelihood value of the super symbol is calculated. Since the received signal Y₂ contains a superposition of the desired message codeword and the DPC codeword, it is possible to obtain three different likelihood values representing the super symbol and its components (namely the TCQ coset codeword A and the desired message codeword ₂), which are then combined using Maximum Ratio Combining (MRC) block. Another difference is on the combined trellis used in the SISO APP block (see, for example, FIG. 6). Here, the combined trellis is from the ACC of the primary user IRA code and the cognitive/secondary user DPC (which is itself a combined trellis of the TCQ and the ACC of cognitive/secondary user IRA code). As such, the combined trellis used for the generalized BCJR in SISO APP block here is larger than the one used in the very strong regime. Referring to FIG. 8, soft values of both the TCQ codeword and ₂ may be calculated. Firstly, knowing that v¾ is a DPC codeword designed to cancel the total interference S = QX₂ at D_\, Equation (2b) may be used to compute

wherein = { ½ ? ¾ - ίΧ / } -- 1 ; v¾ is the effective noise which is approximately Gaussian with variance Λ^¹-* = - - - ifaPt - . The likelihood value of the i^th

TCQ symbol ^Vv iHA - ^&s::¾ ·^·· ^rnay then be calculated using Equation (6), with the noise variance N replaced by N^^l

[057] Moreover, exploiting the fact that X_\ is approximately Gaussian, the LLR of Xi can be calculated from Equation (2b) as follows:

where -- ¥ * ¾? «¾ ^ ¾^ "^"¾is the variance of the effective noise.

[058] Furthermore, using a different scaling factor ^{Λ ?: A>:} the received signal may be transformed into the following:

[059] The received signal model in Equation (23) may represent a transmission of a TCQ coset codeword and a primary user codeword (which are scaled by a factor of Γ and μ, respectively) over a modulo additive Gaussian channel. Considering Φ - ^§"A t as the super-symbol drawn from the concatenated alphabet C il A *^> f the likelihood value of =· for any *^! « "ca be calculated using the same technique as Equation (6).

[060] The scaling factor λ and Γ may be optimized to improve the decoding performance. A good choice of λ and Γ may minimize the effective noise variance and may maximize the minimum modulo-4 Γ-distance d*_mi_n(C) between two elements in C. Rather than performing optimization of either λ or Γ, λ or Γ values are selected to approximate the noise minimizing criterion as follows:

and at the same time maximizing the following metric:

which may, in some implementations, produce good error performance.

[061] To combine the three soft values obtained so far, a Maximum Ratio Combining (MRC) may be applied and then the following calculation may be performed: wherein <ί s A■ø ^■ i * *? C^, m$ r¾ * Note that although all of the soft values above are derived from the same F₂[z], only two of them involve a modulo operation (with different modulus), therefore the MRC computation will generally produce a better likelihood value.

[062] Next, the likelihood value obtained in Equation (24) may be used to calculate the state transition probability required by the generalized BCJR algorithm, which is to be executed on the combined trellis of the ACC of the primary user IRA code and the cognitive user DPC (which is itself a combined trellis of the TCQ and the ACC of cognitive user IRA code). The a posteriori probability of the bits at the ACC input of the IRA codes and their extrinsic LLR can then be calculated in a similar way as Equation (15) and Equation (16), respectively. The only distinction from the previous scenario is that the number of edges to be considered is larger, as the combined trellis includes the TCQ trellis from DPC. The final estimate of w₂ is obtained after several iterations of extrinsic information exchange with the sum-product decoder of both IRA codes.

[063] In some example embodiments, the subject matter disclosed herein may thus allow a cognitive radio overlay system to use coders (typically used with single-user systems) in cognitive multi-terminal systems. Moreover, the coders/decoders may be selected based on a characterization of the interference regime, such as whether the interference corresponds to weak, very strong, or primary decodes cognitive.

FIG. 9 depicts a block diagram of a radio 900, which may be used at one or more of the radios at 103A-B and 105A-B. The radio 900 may include one or more antennas, such as antennas 920A-B for receiving a downlink and transmitting via an uplink. The radio 900 may also include a radio transceiver 940 coupled to the antennas 920 A-B. The radio interface 940 may include other components, such as filters, converters (for example, digital- to-analog converters and the like), symbol demappers, signal shaping components, an Inverse Fast Fourier Transform (IFFT) module, coders, decoders, the like, to process symbols carried by a downlink or an uplink. The radio 900 may further include at least one processor, such as processor 930 for controlling radio 900 and for accessing and executing program code stored in memory 935. In some example embodiments, the memory 935 includes code, which when executed by at least one processor causes one or more of the operations described herein. In some example embodiments, the radio 900 may comprise a mobile station, a mobile unit, a subscriber station, a wireless terminal, a tablet, a smart phone, a base station, a wireless access point, and/or any other type of wireless device configured as a cognitive overlay radio. In some example implementations, the coders and decoders disclosed herein may be performed by at least radio transceiver 940, which may be under the control of processor 930. For example, the radio 900 including processor 900 may receive an indication of whether an interference in a cognitive radio network is at least one of a weak interference regime, a very strong interference regime, and a primary cognitive decodes interference regime. Processor 900 (or a controller at radio 900) may select, when the interference comprises the weak interference regime, a pre-canceling decoder to cancel interference caused by a primary transmitter at a secondary receiver; select, when the interference comprises the very strong interference regime, a first joint decoder to decode a first message superposed on a second message; and select, when the interference comprises the primary decodes cognitive interference regime, a second joint decoder to decode the first message and the second message, the first message having a pre-canceling code decodable by the second joint decoder.

[064] Although FIG. 1 depicts a certain quantity and configuration of radios 103A-B and 105A-B, other quantities and configurations of radios may be used as well. Moreover, the radios 103A-B and 105A-B may also include one or more channels to feedback information, feed-forward information, and/or provide control information to other radios. [065] The subject matter described herein may be embodied in systems, apparatus, methods, and/or articles depending on the desired configuration. For example, the base stations and user equipment (or one or more components therein) and/or the processes described herein can be implemented using one or more of the following: a processor executing program code, an application-specific integrated circuit (ASIC), a digital signal processor (DSP), an embedded processor, a field programmable gate array (FPGA), and/or combinations thereof. These various implementations may include implementation in one or more computer programs that are executable and/or interpretable on a programmable system including at least one programmable processor, which may be special or general purpose, coupled to receive data and instructions from, and to transmit data and instructions to, a storage system, at least one input device, and at least one output device. These computer programs (also known as programs, software, software applications, applications, components, program code, or code) include machine instructions for a programmable processor, and may 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, computer-readable medium, computer-readable storage medium, apparatus and/or device (e.g., magnetic discs, optical disks, memory, 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. Similarly, systems are also described herein that may include a processor and a memory coupled to the processor. The memory may include one or more programs that cause the processor to perform one or more of the operations described herein.

[066] Although a few variations have been described in detail above, other modifications or additions are possible. In particular, further features and/or variations may be provided in addition to those set forth herein. Moreover, the implementations described above may be directed to various combinations and subcombinations of the disclosed features and/or combinations and subcombinations of several further features disclosed above. In addition, the logic flow depicted in the accompanying figures and/or described herein does not require the particular order shown, or sequential order, to achieve desirable results. Other embodiments may be within the scope of the following claims.

[067] Although various aspects of the invention are set out in the independent claims, other aspects of the invention comprise other combinations of features from the described embodiments and/or the dependent claims with the features of the independent claims, and not solely the combinations explicitly set out in the claims.

[068] It is also noted herein that while the above describes example embodiments of the invention, these descriptions should not be viewed in a limiting sense. Rather, there are several variations and modifications which may be made without departing from the scope of the present invention as defined in the appended claims.

Claims

WHAT IS CLAIMED IS

1. A method comprising:

receiving an indication of whether an interference in a cognitive radio network is at least one of a weak interference regime, a very strong interference regime, and a primary decodes cognitive interference regime;

selecting, when the interference comprises the weak interference regime, a pre- canceling decoder to cancel interference caused by a primary transmitter at a secondary receiver;

selecting, when the interference comprises the very strong interference regime, a first joint decoder to decode a first message superposed on a second message; and

selecting, when the interference comprises the primary decodes cognitive interference regime, a second joint decoder to decode the first message and the second message, the first message having a pre-canceling code decodable by the second joint decoder.

2. The method of claim 1, wherein the weak interference regime represents the interference originating primarily from the primary transmitter, rather than a secondary transmitter.

3. The method of claim 2, wherein a cognitive radio comprises the secondary transmitter, the cognitive radio having access to the second message being transmitted by the primary transmitter.

4. The method of claim 1, wherein the very strong interference regime represents the interference originating primarily from a secondary transmitter, rather than the primary transmitter.

5. The method of claim 1, wherein the primary decodes cognitive regime represents substantial interference from both the primary transmitter and a secondary transmitter.

6. The method of claim 1, wherein the pre-canceling decoder comprises a dirty paper decoder.

7. The method of claim 1, wherein the second message originates at the primary transmitter, and the first message originates at a secondary transmitter.

8. The method of claim 1, wherein a cognitive radio comprises the secondary receiver, the secondary receiver implementing the receiving, the selecting the pre-canceling decoder, the selecting the first joint decoder, and the selecting the second joint decoder.

9. An apparatus comprising:

at least one processor circuitry; and

at least one memory circuitry including computer program code, the at least one memory and the computer program code configured to, with the at least one processor, cause the apparatus to perform at least the following:

receive an indication of whether an interference in a cognitive radio network is at least one of a weak interference regime, a very strong interference regime, and a primary decodes cognitive interference regime;

select, when the interference comprises the weak interference regime, a pre-canceling decoder to cancel interference caused by a primary transmitter at a secondary receiver, wherein the apparatus comprises the secondary receiver;

select, when the interference comprises the very strong interference regime, a first joint decoder to decode a first message superposed on a second message; and

select, when the interference comprises the primary decodes cognitive interference regime, a second joint decoder to decode the first message and the second message, the first message having a pre-canceling code decodable by the second joint decoder.

10. The apparatus of claim 9, wherein the weak interference regime represents the interference originating primarily from the primary transmitter, rather than a secondary transmitter.

11. The apparatus of claim 10, wherein a cognitive radio comprises the secondary transmitter, the cognitive radio having access to the second message being transmitted by the primary transmitter.

12. The apparatus of claim 9, wherein the very strong interference regime represents the interference originating primarily from a secondary transmitter, rather than the primary transmitter.

13. The apparatus of claim 9, wherein the primary decodes cognitive regime represents substantial interference from both the primary transmitter and a secondary transmitter.

14. The apparatus of claim 9, wherein the pre-canceling decoder comprises a dirty paper decoder.

15. The apparatus of claim 9, wherein the second message originates at the primary transmitter, and the first message originates at a secondary transmitter.

16. The apparatus of claim 10, wherein a cognitive radio comprises the secondary receiver, the secondary receiver implementing the receiving, the selecting the pre-canceling decoder, the selecting the first joint decoder, and the selecting the second joint decoder.

17. An apparatus comprising:

means for receiving an indication of whether an interference in a cognitive radio network is at least one of a weak interference regime, a very strong interference regime, and a primary decodes cognitive interference regime; means for selecting, when the interference comprises the weak interference regime, a pre-canceling decoder to cancel interference caused by a primary transmitter at a secondary receiver;

means for selecting, when the interference comprises the very strong interference regime, a first joint decoder to decode a first message superposed on a second message; and means for selecting, when the interference comprises the primary decodes cognitive interference regime, a second joint decoder to decode the first message and the second message, the first message having a pre-canceling code decodable by the second joint decoder.

18. A non-transitory computer readable storage medium including computer program

code, which when executed by at least one processor circuitry, causes operations comprising:

selecting, when the interference comprises the weak interference regime, a pre-canceling decoder to cancel interference caused by a primary transmitter at a secondary receiver;

selecting, when the interference comprises the very strong interference regime, a first joint decoder to decode a first message superposed on a second message; and selecting, when the interference comprises the primary decodes cognitive interference regime, a second joint decoder to decode the first message and the second message, the first message having a pre-canceling code decodable by the second joint decoder.