CN102237954A - Soft de-mapping method and device and communication system thereof - Google Patents

Abstract

The invention discloses a soft de-mapping method and device and a communication system thereof. The method comprises the following steps of: calculating a minimum Euclidean distance between all probably transmitted signal vectors corresponding to a part of bits which are not obtained and received signals during signal detection by using channel state information and a modulating coefficient to accordingly establish a complete bit-vector-minimum distance mapping table; and obtaining a logarithm likehood ratio corresponding to every bit through the bit-vector-minimum distance mapping table. The soft de-mapping method can be matched with the signal detection modes of different types to complete a task for decoding a received signal vector into a bit vector, wherein the signal detection modes can be a maximum likehood detection mode and a spherical decoding mode.

Description

Soft de-mapping method and device and its communication system

Technical field

The invention relates to a kind of communication system, particularly relevant for a kind of soft de-mapping method and device of receiver of communication system.

Background technology

In recent years, wireless or wire communication technology develops apace, and people can see through the communication device with communication function whenever and wherever possible and surf the Net, and perhaps converse with other people.Exceedingly be subjected to the influence of wireless channel for fear of signal vector, and cause receiver received signal vector smoothly, multi-input multi-output system is used especially widely at present.How receiver generally can use hybrid sphere decoding (Sphere Decoding abbreviates SD as) or the most general like detecting (Maximum LikelihoodDetection abbreviates MLD as) when carrying out signal detection.

The most general like the detecting meeting find out might the set of signal vector from comprising one with received signal vector apart from the most close signal vector, as if the situation of not considering to misread, this signal vector the most close with the received signal vector distance is the signal vector that conveyer transmits.The most general signal vector of being found out like detecting can be expressed as

Wherein y is a received signal vector, and H is the system channel matrix, and x is arbitrary group a signal vector in the S set, and S set has comprised that institute might a signal vector.The most general signal vector that is solved like detecting

Be optimum solution.

Hybrid sphere decoding is different from the most probably must be searched all signal vectors like detecting, and hybrid sphere decoding is only searched the signal vector of part, and from the signal of being searched, find out one with received signal vector apart from the most close signal vector.Hybrid sphere decoding is only searched the signal vector of part, and therefore the signal vector of a part and the Euclidean distance (Euclidean Distance) of received signal vector can not calculated in addition.The signal vector that hybrid sphere decoding solved is a suboptimal solution.

Receiver solve signal vector after, can separate mapping to signal vector, with the weight of entrained each bit of picked up signal vector (bit), or claim to transmit possibility.For instance, if communication system adopts orthogonal phase shift modulation (QPSK), as if being separated, solid part signal and imaginary signals carry out demodulation, when solid part signal pairing signal on planisphere (Constellation Map) real number axis is-1, the bit of this solid part signal correspondence is 0, when solid part signal pairing signal on the planisphere real number axis is 1, the bit of this solid part signal correspondence is 1.Similarly, when imaginary signals pairing signal on the planisphere axis of imaginaries is-1, the bit of this imaginary signals correspondence is 0, and when the pairing signal of imaginary signals on the planisphere axis of imaginaries is 1, the bit of this imaginary signals correspondence is 1.

For instance, when the single input single output system of communication system for employing orthogonal phase shift modulation, just, when the conveyer of communication system and receiver only have an antenna respectively, if receiver is solving signal vector

Afterwards, just can be with signal vector

Separate and be mapped as bit vector b=[b ₁b ₂] ^T=[0 1] ^T, signal wherein

With Be respectively solid part signal and imaginary signals that receiver solves.At the orthogonal phase shift modulating signal, real part or imaginary signals correspond to a bit separately, promptly

With

Separate mapping and can be divided into and rigidly separate mapping and separate mapping, rigidly separate mapping and separate a plurality of bit value that are mapped as corresponding bit vector with the signal vector that the mapping relations of signal vector directly solve receiver according to the bit vector with soft.It is soft that to separate mapping then be that a plurality of logarithms of signal calculated vector are general like ratio (LogLikelihood Ratio abbreviates LLR as), and by the general a plurality of bit value that solve the pairing bit vector of signal vector like ratio of described a plurality of logarithms.Described logarithm is general to be defined as follows like ratio,

$\hat{L}\left(b_n\right)={\min }_{x\∈S_{b_n=1}}\left(\frac{1}{{\mathrm{\σ}}^2}\right){\left|\right|y-\mathrm{Hx}\left|\right|}^2-{\min }_{x\∈S_{b_n=0}}\left(\frac{1}{{\mathrm{\σ}}^2}\right){\left|\right|y-\mathrm{Hx}\left|\right|}^2,$

Wherein, y is a received signal vector, and H is the system channel matrix, and x is a signal vector, σ ²Be noise power,

For working as bit b _nPairing set that might signal vector in=1 o'clock, For working as bit b _nPairing set that might signal vector in=0 o'clock, n=1 ..., N _TM _c, N _TBe to transmit antenna number, M _cBe each real part on the planisphere or bit number that imaginary signals had.

Please refer to Figure 1A and Figure 1B, Figure 1A is that tradition is used and the most probably to be obtained the method schematic diagram of bit vector like detecting and the soft receiver of separating mapping, and Figure 1B is the schematic diagram that comprises the distance of each signal vector in the set of all possible signal vector and received signal vector.It is the most general like detecting that receiver the most general carried out like the vectorial to received signal y of detecting module 100, with from the S set that comprises all signal vectors, find out one with received signal vector y apart from the most close signal vector

Signal wherein With

Be respectively solid part signal and the imaginary signals that receiver solves.

In Figure 1B, communication system is for adopting the orthogonal phase shift modulation, and its receiver and conveyer all have an antenna.The most general Euclidean distance like each signal vector and received signal vector y in the detecting module 100 meeting calculating S set is to find out the immediate signal vector with received signal vector y Shown in Figure 1B, for signal vector x=[x ₁x ₂] ^T=[1 1] ^T, signal vector x=[1 1] ^TWith the Euclidean distance of received signal vector y be 0.1; For signal vector x=[1-1] ^T, signal vector x=[1-1] ^TWith the Euclidean distance of received signal vector y be 0.5; For signal vector x=[-1 1] ^T, signal vector x=[-1 1] ^TWith the Euclidean distance of received signal vector y be 0.3; For signal vector x=[-1-1] ^TThe time, signal vector x=[-1-1] ^TWith the Euclidean distance of received signal vector y be 0.7.Therefore, the most general seemingly detecting module 100 solves the immediate signal vector with received signal vector y

In order to carry out the soft mapping of separating, receiver must be learnt respectively at bit b ₁=1, b ₁=0, b ₂=1, b ₂The minimum Eustachian distance of=0 o'clock pairing each signal vector and received signal vector y.Therefore, receiver can be recorded in the Euclidean distance of all bit vector pairing all signal vectors of b and received signal vector y in the bit vector distance correspondence table 110.For instance, at the vectorial b=[0 0 of the bit of bit vector distance correspondence table 110] ^TThe signal vector x=[-1-1 of place correspondence] ^TWith the Euclidean distance of received signal vector y be 0.7.

The signal vector of the Euclidean distance minimum that the most general seemingly detecting module 100 is found out

Therefore receiver is at the bit b of bit minimum range correspondence table 120 ₁=1 place, record bit b ₁=1 o'clock pairing might signal vector and the minimum Eustachian distance of received signal vector y be

Similarly, at the bit b of bit vector minimum range correspondence table 120 ₂=1 place, record bit b ₂=1 o'clock pairing might signal vector and the minimum Eustachian distance of received signal vector y be

In addition, can learn bit vector b=[0 1 by bit vector distance correspondence table 110] ^TThe Euclidean distance of the signal vector of place correspondence and received signal vector y is 0.3, and learns bit vector b=[0 0] ^TThe Euclidean distance of the signal vector of place correspondence and received signal vector y is 0.7.Therefore, receiver is at the bit b of bit vector minimum range correspondence table 120 ₁=0 place, record bit b ₁The minimum Eustachian distance of=0 o'clock all possible signal vector and received signal vector y is

Similarly, can learn bit vector b=[1 0 by bit vector distance correspondence table 110] ^TThe Euclidean distance of the signal vector of place correspondence and received signal vector y is 0.5, and learns bit vector b=[0 0] ^TThe Euclidean distance of the signal vector of place correspondence and received signal vector y is 0.7.Therefore, receiver is at the bit b of bit vector minimum range correspondence table 120 ₂=0 place, record bit b ₂The minimum Eustachian distance of=0 o'clock all possible signal vector and received signal vector y is

By the content of bit vector minimum range correspondence table 120, receiver just can carry out the soft mapping of separating, and is general like ratio to calculate logarithm

To estimate the bit b that conveyer may transmit whereby ₁=1.Similarly, by the content of bit vector minimum range correspondence table 120, receiver just can carry out the soft mapping of separating, and is general like ratio to calculate logarithm

To estimate the bit b that conveyer may transmit whereby ₂=1.

Please refer to Fig. 2 A～Fig. 2 C, Fig. 2 A is to use hybrid sphere decoding and the soft receiver of separating mapping to obtain the method schematic diagram of bit vector, Fig. 2 B is the schematic diagram of the distance of each signal vector and received signal vector in the set of the possible signal vector of part, and Fig. 2 C is a schematic diagram of searching the distance of each signal vector and received signal vector in the set of possible signal vector with Fig. 2 B rightabout stratum.Receiver uses two hybrid sphere decoding modules 210 and 220 vectorial to received signal y to carry out hybrid sphere decoding simultaneously, and wherein hybrid sphere decoding module 210 is with signal x when vectorial y decodes to received signal ₂Be made as the ground floor signal, decode; Hybrid sphere decoding module 210 is with signal x when vectorial y decoding to received signal ₁Be made as the ground floor signal, decode.Signal x herein ₁With x ₂Be respectively possible solid part signal and imaginary signals, and signal vector x in this example for comprising two layer signal x ₁With x ₂Signal vector.

Suppose communication system for adopting the single output system of single input of orthogonal phase shift modulation, as Fig. 2 A 210, shown in Fig. 2 B, hybrid sphere decoding module 210 is only to comprising signal vector x=[x ₁x ₂] ^T=[1 ± 1] ^TS set ₁Search, and from S set ₁In find out a signal vector the most close with received signal vector y distance

Signal vector x=[-1 ± 1] ^TCan't be searched, therefore, receiver can be with bit vector b=[b ₁b ₂] ^T=[1 1] ^T, b=[1 0] ^TThe Euclidean distance of pairing each signal vector and received signal vector y is recorded in the bit vector distance correspondence table 212.In other words, hybrid sphere decoding module 210 can't be calculated bit vector b=[0 1] ^T, b=[0 0] ^TThe Euclidean distance of pairing each signal vector and received signal vector y is so bit vector distance correspondence table 212 can't store bit vector b=[b ₁b ₂] ^T=[0 1] ^T, b=[0 0] ^TThe Euclidean distance of pairing each signal vector and received signal vector y.

In addition, shown in Fig. 2 A, Fig. 2 C, hybrid sphere decoding module 220 is only to comprising signal vector x=[± 1 1] ^TS set ₂Search, and from S set ₂In find out a signal vector the most close with received signal vector y distance

Signal vector x=[± 1-1] ^TCan't be searched, therefore, receiver can be with bit vector b=[1 1] ^T, b=[0 1] ^TThe Euclidean distance of pairing each signal vector and received signal vector y is recorded in the bit vector distance correspondence table 222.In other words, hybrid sphere decoding module 220 can't be calculated bit vector b=[1 0] ^T, b=[0 0] ^TThe Euclidean distance of pairing each signal vector and received signal vector y is so bit vector distance correspondence table 212 can't store bit vector b=[0 1] ^T, b=[0 0] ^TThe Euclidean distance of pairing each signal vector and received signal vector y.

Then, receiver can be set up out bit vector minimum range correspondence table 214 and 224 respectively according to bit vector distance correspondence table 212 and 222.Because bit vector distance correspondence table 212 can't store bit vector b=[0 1] ^T, b=[0 0] ^TThe Euclidean distance of pairing each signal vector and received signal vector y, so bit vector minimum range correspondence table 214 can't write down bit b ₁The minimum Eustachian distance P of=0 o'clock pairing all signal vector and received signal vector y ₁ ⁰Similarly, because bit vector distance correspondence table 222 can't store bit vector b=[1 0] ^T, b=[0 0] ^TThe Euclidean distance of pairing each signal vector and received signal vector y, so bit vector minimum range correspondence table 224 can't write down bit b ₂The minimum Eustachian distance P of=0 pairing all signal vectors and received signal vector y ₂ ⁰

Afterwards, receiver can be set up out complete bit vector minimum range correspondence table 230 according to bit vector minimum range correspondence table 214 and 224.By the content of bit vector minimum range correspondence table 230, receiver just can carry out the soft mapping of separating, and is general like ratio to calculate logarithm To estimate the bit b that conveyer may transmit whereby ₁=1.Similarly, by the content of bit vector minimum range correspondence table 230, receiver just can carry out the soft mapping of separating, and is general like ratio to calculate logarithm

To estimate the bit b that conveyer may transmit whereby ₂=1.

Summary of the invention

The object of the present invention is to provide a kind of soft de-mapping method and device and its communication system.When detecting by signal calculated pairing all signal vectors of part bit of not calculating and the minimum Eustachian distance between the received signal, setting up complete bit vector minimum range correspondence table whereby, and it is general like ratio to obtain the pairing logarithm of each bit by bit vector minimum range correspondence table.The soft de-mapping method of this kind can cooperate with different types of signal detection mode, to finish the task of received signal vector being decoded as the bit vector.

For realizing that purpose of the present invention provides a kind of soft de-mapping method, be applicable to the receiver of communication system, general in order to the logarithm that obtains pairing each bit of received signal vector like ratio.Described receiver receives described received signal vector

The institute that the conveyer of described communication system transmitted might be expressed as by signal vector

And the signal x of each stratum _jCorresponding a plurality of bits are [b _{(l-1) Mc+1}b _{(l-1) Mc+2}B _{(l-1) Mc+Mc}] ^T, wherein l=2j-1 is a real part, and l=2j is an imaginary part, and j is 1 to N _TInteger, N _TBe the signal sum of described received signal vector, M _cBe the signal x of j stratum _jPairing real part or imaginary part bit sum.Described soft de-mapping method may further comprise the steps: steps A, described received signal vector y is carried out signal detection, to obtain bit vector minimum range correspondence table, the integrality of this correspondence table is according to the transmission signal set size of searching of seeking, if the transmission signal set of searching of seeking too little, this table is imperfect more; Step B has the part that lacks to calculate its value at the correspondence table in the steps A, if the signal x of j stratum _jPairing part bit vector minimum range has and lacks, and its account form be each the capable vectorial h according to the system channel matrix H _j, calculate the signal x of j stratum _jThe signal x of wrong and other stratum _{I, i ≠ j}Each bit b under the correct situation _nThe minimum Eustachian distance P of pairing signal vector and received signal vector y during mistake _{J, n}, wherein n is 1 to N _TM _cInteger, the order of steps A and step B is commutative, promptly step B can first rough estimate h _jAnd in correspondence table, deposit all bit vector minimum ranges in advance, if then try to achieve certain bit vector minimum range, again the value in the correspondence table is upgraded in steps A; And step C, according to each minimum Eustachian distance P _{J, n}Set up complete bit vector minimum range correspondence table with incomplete bit vector minimum range correspondence table.

Described signal detection is a hybrid sphere decoding or the most general seemingly detecting.

When described signal is detected, solve immediate signal vector with described received signal vector y

And write down described signal vector at least With the minimum Eustachian distance of described received signal vector y, to set up out incomplete bit vector minimum range correspondence table whereby.

When described signal is detected, more write down the Euclidean distance of part signal vector with described received signal vector y, and according to described signal vector

Set up incomplete bit vector minimum range correspondence table with Euclidean distance, the part signal vector of described received signal vector y with the Euclidean distance of described received signal vector y.

Described soft de-mapping method more comprises:

Calculate each bit b according to complete bit vector minimum range correspondence table _nLogarithm general like ratio L (b _n).

This P of described calculating _{J, n}Step can carry out this signal detection with the step that obtains incomplete this bit vector minimum range correspondence table before, at this moment, first rough estimate h _jAnd in correspondence table, deposit all bit vector minimum ranges in advance, then in the step of this signal detection,, the value in this bit vector minimum range correspondence table is being upgraded if try to achieve certain bit vector minimum range.

Described minimum Eustachian distance P _{J, n}=K * E[||h _j|| ²]+E[||n|| ²], wherein n is a noise vector, K is a modulating coefficient.

Described minimum Eustachian distance

Wherein n is a noise vector,

Be modulating coefficient,

The signal that is solved with described communication system

And the modulating method of wrong bit is relevant with its position at planisphere.

For realizing that purpose of the present invention also provides a kind of soft mapping device of separating, be applicable to the receiver of communication system, general in order to the logarithm that obtains pairing each bit of received signal vector like ratio.Described receiver receives described received signal vector

The institute that the conveyer of described communication system transmitted might be expressed as by signal vector

And the signal x of each stratum _jCorresponding a plurality of bits are [b _{(l-1) Mc+1}b _{(l-1) Mc+2}B _{(l-1) Mc+Mc}] ^T, wherein l=2j-1 is a real part, and l=2j is an imaginary part, and j is 1 to N _TInteger, N _TBe the signal sum of described received signal vector, M _cBe the signal x of j stratum _jPairing real part or imaginary part bit sum.The described soft mapping device of separating comprises bit vector minimum range correspondence table module, channel condition information acquiring unit and computing unit.Bit vector minimum range correspondence table module is carried out the result of signal detection to obtain bit vector minimum range correspondence table to described received signal vector y according to the signal detection module, the integrality of this correspondence table is according to the transmission signal set size of searching of seeking, if the transmission signal set of searching of seeking too little, this table is imperfect more.Channel condition information acquiring unit self-channel estimating apparatus obtains the vectorial h of each row of system channel matrix H _jComputing unit has the part that lacks to calculate its value at correspondence table, and its account form is the vectorial h of each row according to described system channel matrix H _j, calculate the signal x of the j stratum _jThe signal x of wrong and other stratum _{I, i ≠ j}Each bit b under all correct situation _nThe minimum Eustachian distance P of pairing signal vector and received signal vector y during mistake _{J, n}, wherein k is 1 to N _TM _cInteger.The execution sequence that is noted that aforementioned bit vector minimum range correspondence table module and computing unit is commutative, and promptly computing unit can first rough estimate h _jAnd in correspondence table, deposit all bit vector minimum ranges in advance, then,, again the value in the correspondence table is upgraded if bit vector minimum range correspondence table module is tried to achieve certain bit vector minimum range.In addition, described bit vector minimum range correspondence table module is more according to each minimum Eustachian distance P _{J, n}Set up complete bit vector minimum range correspondence table with incomplete bit vector minimum range correspondence table.

Described signal detection module is a hybrid sphere decoding module or the most general seemingly detecting module.

When described signal was detected, described signal detection module solved the immediate signal vector with described received signal vector y

And write down described signal vector at least

With the Euclidean distance of described received signal vector y, to allow described bit vector minimum range correspondence table module set up out the vectorial minimum range correspondence table of incomplete unit whereby.

When described signal was detected, described signal detection module had more write down the Euclidean distance of part signal vector and described received signal vector y, described bit vector minimum range correspondence table module and according to described signal vector

Set up the vectorial minimum range correspondence table of incomplete unit with Euclidean distance, the part signal vector of described received signal vector y with the Euclidean distance of described received signal vector y.

Described bit vector minimum range correspondence table module is calculated each bit b according to complete bit vector minimum range correspondence table _nLogarithm general like ratio L (b _n).

Described computing unit just calculated the signal x of j stratum before described signal detection is carried out _jThe signal x of wrong and other stratum _{I, i ≠ j}Each bit b under all correct situation _nThe minimum Eustachian distance P of pairing signal vector and received signal vector y during mistake _{J, n}, that is this computing unit can first rough estimate h _jAnd in this bit vector minimum range correspondence table, deposit all bit vector minimum ranges in advance, then, if bit vector minimum range correspondence table module is tried to achieve certain bit vector minimum range, again the value in this bit vector minimum range correspondence table is upgraded; Perhaps, described computing unit just calculates the signal x of j stratum after described signal detection is carried out _jThe signal x of wrong and other stratum _{I, i ≠ j}Each bit b under all correct situation _nThe minimum Eustachian distance P of pairing signal vector and received signal vector y during mistake _{J, n}

Described minimum Eustachian distance P _{J, n}=K * E[||h _j|| ²]+E[||n|| ²], wherein n is a noise vector, K is a modulating coefficient.

The described soft mapping device of separating more comprises:

One modulating coefficient amending unit is in order to the described minimum Eustachian distance P with described computing unit calculated _{J, n}=K * E[||h _j|| ²]+E[||n|| ²] be modified to minimum Eustachian distance

Wherein n is a noise vector,

Be modulating coefficient,

The signal that is solved with described communication system

And the modulating method of wrong bit is relevant with its position at planisphere.

For realizing that purpose of the present invention also provides a kind of communication system, described communication system comprises receiver and conveyer, and described receiver comprises soft mapping device and the signal detection module of separating.The described soft mapping device of separating is in order to the general seemingly ratio of the logarithm that obtains pairing each bit of received signal vector, and described receiver receives described received signal vector

The institute that the conveyer of described communication system transmitted might be expressed as by signal vector

And the signal x of each stratum _jCorresponding a plurality of bits are [b _{(l-1) Mc+1}b _{(l-1) Mc+2}B _{(l-1) Mc+Mc}] ^T, wherein l=2j-1 is a real part, and l=2j is an imaginary part, and j is 1 to N _TInteger, N _TBe the signal sum of described received signal vector, M _cBe the signal x of j stratum _jPairing real part or imaginary part bit sum.The described soft mapping device of separating comprises bit vector minimum range correspondence table module, channel condition information acquiring unit and computing unit.Bit vector minimum range correspondence table module is carried out the result of signal detection to obtain bit vector minimum range correspondence table to described received signal vector y according to the signal detection module, the integrality of this correspondence table is according to the transmission signal set size of searching of seeking, if the transmission signal set of searching of seeking too little, this table is imperfect more.Channel condition information acquiring unit self-channel estimating apparatus obtains the vectorial h of each row of system channel matrix H _jComputing unit has the part that lacks to calculate its value at correspondence table, and its account form is the vectorial h of each row according to described system channel matrix H _j, calculate the signal x of the j stratum _jThe signal x of wrong and other stratum _{I, i ≠ j}Each bit b under all correct situation _nThe minimum Eustachian distance P of pairing signal vector and received signal vector y during mistake _{J, n}, wherein k is 1 to N _TM _cInteger.The execution sequence that is noted that aforementioned bit vector minimum range correspondence table module and computing unit is commutative, and promptly computing unit can first rough estimate h _jAnd in correspondence table, deposit all bit vector minimum ranges in advance, then,, again the value in the correspondence table is upgraded if bit vector minimum range correspondence table module is tried to achieve certain bit vector minimum range.In addition, described bit vector minimum range correspondence table module is more according to each minimum Eustachian distance P _{J, n}Set up complete bit vector minimum range correspondence table with incomplete bit vector minimum range correspondence table.

Described signal detection module is a hybrid sphere decoding module or the most general seemingly detecting module.

When described signal was detected, described signal detection module solved the immediate signal vector with described received signal vector y

And write down described signal vector at least With the Euclidean distance of described received signal vector y, to allow described bit vector minimum range correspondence table module set up out the vectorial minimum range correspondence table of incomplete unit whereby.

When described signal was detected, described signal detection module had more write down the Euclidean distance of part signal vector and described received signal vector y, described bit vector minimum range correspondence table module and according to described signal vector

Set up incomplete bit vector minimum range correspondence table with Euclidean distance, the part signal vector of described received signal vector y with the Euclidean distance of described received signal vector y.

Described bit vector minimum range correspondence table module is calculated each bit b according to complete bit vector minimum range correspondence table _nLogarithm general like ratio L (b _n).

Described computing unit just calculated the signal x of j stratum before described signal detection is carried out _jThe signal x of wrong and other stratum _{I, i ≠ j}Each bit b under all correct situation _nThe minimum Eustachian distance P of pairing signal vector and received signal vector y during mistake _{J, n}, that is this computing unit can first rough estimate h _jAnd in this bit vector minimum range correspondence table, deposit all bit vector minimum ranges in advance, then, if bit vector minimum range correspondence table module is tried to achieve certain bit vector minimum range, again the value in this bit vector minimum range correspondence table is upgraded; Perhaps, described computing unit just calculates the signal x of j stratum after described signal detection is carried out _jThe signal x of wrong and other stratum _{I, i ≠ j}Each bit b under all correct situation _nThe minimum Eustachian distance P of pairing signal vector and received signal vector y during mistake _{J, n}

Described minimum Eustachian distance P _{J, n}=K * E[||h _j|| ²]+E[||n|| ²], wherein n is a noise vector, K is a modulating coefficient.

The described soft mapping device of separating more comprises:

One modulating coefficient amending unit is in order to the described minimum Eustachian distance P with described computing unit calculated _{J, n}=K * E[||h _j|| ²]+E[||n|| ²] be modified to minimum Eustachian distance

Wherein n is a noise vector,

Be modulating coefficient,

The signal that is solved with described communication system

And the modulating method of wrong bit is relevant with its position at planisphere.

The soft de-mapping method that provides of the present invention can be applied in the receiver of various unlike signal detecting modes, and described soft de-mapping method can utilize simple formula to obtain null field in the incomplete bit vector minimum range correspondence table, and is general like ratio with the logarithm that successfully calculates each bit whereby.

Description of drawings

Figure 1A is that tradition is used the most general method schematic diagram that obtains the bit vector like detecting and the soft receiver of separating mapping;

Figure 1B is the schematic diagram that comprises the distance of each signal vector in the set of all possible signal vector and received signal vector;

Fig. 2 A is to use hybrid sphere decoding and the soft receiver of separating mapping to obtain the method schematic diagram of bit vector;

Fig. 2 B is the schematic diagram of the distance of each signal vector in the set of the possible signal vector of part and received signal vector;

Fig. 2 C is the schematic diagram of the distance of each signal vector in the set of the possible signal vector of another part and received signal vector;

Fig. 3 A is to use the receiver of the soft de-mapping method that hybrid sphere decoding and the embodiment of showing of the present invention provide to obtain the method schematic diagram of bit vector;

Fig. 3 B is the schematic diagram of the distance of each signal vector and received signal vector in the set of the possible signal vector of part;

Bit b when Fig. 4 A is 16 quadrature amplitude modulations of communication system employing _nWith b _N+1Pairing signal x _jPlanisphere;

Bit b when Fig. 4 B is 64 quadrature amplitude modulations of communication system employing _n, b _N+1, b _N+2Pairing signal x _jPlanisphere;

Fig. 5 is the calcspar of the communication system that provides of embodiments of the invention;

Fig. 6 is the soft calcspar of separating mapping device that embodiments of the invention provide;

Fig. 7 is the flow chart of the soft de-mapping method that provides of embodiments of the invention;

Fig. 8 is the schematic diagram of bit vector minimum range correspondence table of providing of embodiments of the invention.

Wherein, Reference numeral

100: the most general like detecting module

110: bit vector distance correspondence table

120: bit vector minimum range correspondence table

210,220: the hybrid sphere decoding module

212,222: bit vector distance correspondence table

214,224,230: bit vector minimum range correspondence table

310: the hybrid sphere decoding module

320: bit vector distance correspondence table

330: bit vector minimum range correspondence table

340: the soft mapping device of separating

350: bit vector minimum range correspondence table

500: communication system

502: conveyer

504: receiver

506: wireless channel

TX1～TX2: two transmission antennas

RX1～RX2: two reception antennas

550: the signal detection module

560: channel estimating apparatus

600: the soft mapping device of separating

610: computing unit

620: the channel condition information acquiring unit

630: the modulating coefficient amending unit

640: bit vector minimum range correspondence table module

650: multiplexer

S700～S704: step

Embodiment

In order to make purpose of the present invention, technical scheme and advantage clearer, below in conjunction with drawings and Examples, to soft de-mapping method of the present invention and the device and its communication system be further elaborated.Should be appreciated that specific embodiment described herein only in order to explanation the present invention, and be not used in qualification the present invention.

Embodiments of the invention provide a kind of soft de-mapping method, be used for signal calculated when detecting institute not pairing all signal vectors of part bit of calculating and the minimum Eustachian distance between the received signal, setting up complete bit vector minimum range correspondence table whereby, and it is general like ratio to obtain the pairing logarithm of each bit by bit vector minimum range correspondence table.The soft de-mapping method of this kind can cooperate with different types of signal detection mode, and to finish the task of received signal vector being decoded as the bit vector, wherein said signal detection mode can be the most general like detecting and hybrid sphere decoding.

Please refer to Fig. 3 A and Fig. 3 B, Fig. 3 A is to use the receiver of the soft de-mapping method that hybrid sphere decoding and embodiments of the invention provide to obtain the method schematic diagram of bit vector, and Fig. 3 B is the schematic diagram of the distance of the set of all possible signal vector and part signal vector sum received signal vector.The pairing communication system of Fig. 3 A and Fig. 3 B adopts the orthogonal phase shift modulation, and receiver and conveyer all only have an antenna.Receiver obtains received signal vector y=[y from wireless channel ₁y ₂] ^T, received signal vector y=[y ₁y ₂] ^TReceived signal y with two stratum ₁With y ₂, received signal y wherein ₁With y ₂Represent real part received signal and imaginary part received signal respectively.

In Fig. 3 A and Fig. 3 B, the hybrid sphere decoding module 310 of receiver can be carried out signal detection by vectorial to received signal y, and signal detection herein is a hybrid sphere decoding.Hybrid sphere decoding module 310 is only to comprising signal vector x=[x ₁x ₂] ^T=[± 1 1] ^TSet Search, and according to the Euclidean distance of correspondence from set

In find out a signal vector the most close with received signal vector y distance

The signal vector that receiver solved wherein In signal

With

Solid part signal and the imaginary signals representing receiver respectively and solved, and because communication system adopts orthogonal phase shift modulation, signal

With

Can correspond to a bit b respectively through after separating mapping ₁With b ₂In addition, the signal x among the signal vector x ₁With x ₂Represent solid part signal and imaginary signals respectively.

In these examples of implementation, signal vector x=[± 1-1] ^TCan't be searched, therefore, receiver can be with bit vector b=[b ₁b ₂] ^T=[1 1] ^T, b=[0 1] ^TThe Euclidean distance of pairing each signal vector and received signal vector y is recorded in the bit vector distance correspondence table 320.In other words, hybrid sphere decoding module 310 can't be calculated bit vector b=[1 0] ^T, b=[0 0] ^TThe Euclidean distance of pairing each signal vector and received signal vector y is so bit vector distance correspondence table 320 can't store bit vector b=[1 0] ^T, b=[0 0] ^TThe Euclidean distance of pairing each signal vector and received signal vector y.

Receiver can be set up out bit vector minimum range correspondence table 330 according to bit vector distance correspondence table 320.Because bit vector distance correspondence table 320 can't store bit vector b=[1 0] ^T, b=[0 0] ^TThe Euclidean distance of pairing each signal vector and received signal vector y, so bit vector minimum range correspondence table 330 can't write down bit b ₂The minimum Eustachian distance P of=0 o'clock pairing all signal vector and received signal vector y ₂ ⁰That is to say that the integrality of bit vector minimum range correspondence table 330 is according to the transmission signal set of searching of seeking

Size is if the transmission signal set of searching of seeking

Too little, then bit is imperfect more with vectorial minimum range correspondence table 330.Yet, if will make receiver can carry out the soft mapping of separating, obtain the bit vector that conveyer transmits, must obtain minimum Eustachian distance P further ₂ ⁰

In view of the above, embodiments of the invention provide a kind of soft de-mapping method, and it is in order to find out the minimum Eustachian distance of pairing all signal vectors of the bit that does not calculate and received signal vector y when carrying out signal detection.With embodiment, the soft mapping device 340 of separating of the soft de-mapping method that the use present embodiment provides more can calculate minimum Eustachian distance P ₂ ⁰Then, the soft mapping device 340 of separating just can be set up a complete bit vector minimum range correspondence table 350, and the bit vector that obtains conveyer in view of the above and transmitted.

Received signal vector y can be expressed as y=Hx+n, and wherein H is the system channel matrix, and n is a noise vector.The calculating that sees through hybrid sphere decoding module 310 with search after, can solve the signal vector that conveyer transmits and may be

The soft de-mapping method that present embodiment provided is that hypothesis is at signal to noise ratio (Signal to Noise Ratio, abbreviate SNR as) under the very high situation, use the most probably to calculate apace when carrying out hybrid sphere decoding like the judgement formula, the minimum Eustachian distance of pairing all signal vectors under the received signal vector y that can't obtain and the part bit misunderstanding situation.In addition, according to analog result, under the very low situation of signal to noise ratio, the method is suitable equally.

If the signal of first stratum that hypothesis is solved

Be correct, that is

X wherein ₁Be to transmit signal, so the most general formula like mistake in computation vector in the judgement formula can be expressed as Signal when second stratum

The time, x wherein ₂Be to transmit signal, can obtain bit b ₂=1 pairing the minimum Eustachian distance that might transmit signal vector x and received signal vector y

,

With

Condition under because P ₂ ¹Can correspond to minimum Eustachian distance, promptly || n|| ², hybrid sphere decoding module 310 can obtain this value under most of situation.On the contrary, when the signal of second stratum The time (that is, bit b ₂Wrong), then can obtain bit b ₂The minimum Eustachian distance of=0 pairing all signal vectors and received signal vector y

In this example, bit b ₂The minimum Eustachian distance P of=0 signal vector and received signal vector y ₂ ⁰Obtain in the time of can't searching and calculate by hybrid sphere decoding module 310.

If the signal of second stratum that hypothesis is solved

Be correct, that is So the most general formulate like mistake in computation vector in the judgement formula is

Signal when first stratum

The time, can obtain bit b ₁The minimum Eustachian distance of=1 pairing all signal vector x and received signal vector y

With

Condition under because P ₁ ¹Can correspond to minimum Eustachian distance, promptly || n|| ², hybrid sphere decoding module 310 can obtain this value under most of situation.On the contrary, when the signal of first stratum

The time (that is, bit b ₁Wrong), can obtain bit b ₁The minimum Eustachian distance of=0 pairing all signal vectors and received signal vector y

Bit b in this example ₁The minimum Eustachian distance P of=0 signal vector and received signal vector y ₁ ⁰Obtain in the time of can searching and calculate by hybrid sphere decoding module 310.

In view of the above, the soft de-mapping method that present embodiment provided can utilize channel condition information (ChannelState Information abbreviates CSI as) and modulating method (modulation scheme) when obtaining apace to carry out signal detection pairing all signal vectors of part bit that can't obtain and the minimum Eustachian distance of received signal vector.With top illustrative example, when hybrid sphere decoding is adopted in signal detection, then NO minimum Eustachian distance P ₂ ⁰Can use above-mentioned computing formula,

Obtain minimum Eustachian distance P fast ₂ ⁰Thus, just can obtain complete bit vector minimum range correspondence table 350, and can carry out the soft mapping of separating through bit vector minimum range correspondence table 350.

The fast calculation formulae of pairing all signal vectors of part bit that do not obtained during then, at above-mentioned signal detection and the minimum Eustachian distance of received signal vector is carried out general derivation.The fast calculation formulae of the signal vector of the part bit that acquisition is not obtained when carrying out signal detection and the minimum Eustachian distance of received signal vector can be expressed as P _j=E[||y-Hx _j|| ²], P _jExpression is to the signal x of any j of stratum _jAnd other non-x _jThe minimum Eustachian distance of the combination of all signal vectors and received signal vector y, E represents the mean value of all elements of the number of winning the confidence vector, wherein x _jBe set

In arbitrary group signal vector and j=1 ..., N _T, set

Be fixed signal x _jThe time pairing set that might signal vector.x _jCan correspond to one or more bits according to its modulating method, if hypothesis is the orthogonal phase shift modulation, a signal x _jCan correspond to a bit b _n, n=1 wherein ..., N _TM _c, and when the orthogonal phase shift modulation is used, M _c=1, thus in this example b _n=b _jIn addition, defining its mapping relations is x _jThe b of=1 o'clock correspondence _j=1, x on the contrary _jThe b of=-1 o'clock correspondence _j=0.Suppose that signal to noise ratio is very high and transmit bit is b _j=1 (is x _j=1), can obtain P via hybrid sphere decoding _j=E[||y-Hx _j|| ²]=E[||n|| ²], this value is b _j=1 bit vector minimum range.The semaphore of searching when hybrid sphere decoding has in limited time, might can't obtain b _j=0 bit vector minimum range, therefore, with aforementioned calculation formula (P _j=E[||y-Hx _j|| ²]) derive, under the very high hypothesis of signal to noise ratio, can reasonably infer other non-x _jSignal equal to transmit signal, so E[||y-Hx _j|| ²] only remaining x _jComponent and noise,

Wherein, h _jBe the capable row vector of the j of system channel matrix H.

P _jApproximate formula can continue again to be launched into again

If the capable vectorial h of row of j of supposing the system channel matrix H _jUncorrelated each other with noise vector n (Uncorrelated), then P _jFormula can be expressed as again

The vectorial h of row that wherein j of system channel matrix H is capable _jCan be obtained by channel condition information, K then is called modulating coefficient, and its numerical value is relevant with the mode of modulation, further says modulating coefficient K and the signal that is solved

And the planisphere of wrong bit is relevant.

If will set up complete bit vector minimum range correspondence table more quickly, can be at any signal x _jPairing any bit b _nBit vector minimum range have under the situation about lacking, suppose the signal x of j stratum _jThe signal x of wrong and other stratum _{I, i ≠ j}When all correct, the signal vector that this group is supposed is multiplied by its corresponding modulating coefficient K with the minimum Eustachian distance of received signal vector y, in addition, adds the influence of noise item.

Yet, if obtain comparatively accurate result, modulating coefficient K and the signal that is solved

And the modulating method of wrong bit and its bit at planisphere are equipped with the pass.In other words, for the signal x of j stratum _jThe signal x of wrong and other stratum _{I, i ≠ j}The bit b of correspondence under all correct situation _nThe minimum Eustachian distance of all signal vectors of correspondence and received signal vector y can have part difference because of modulating coefficient K during mistake.

For instance, when using the orthogonal phase shift modulation, modulating coefficient is all

Other lifts an example, please refer to Fig. 4 A, Fig. 4 A be communication system when adopting 16 quadrature amplitude modulations (QuadratureAmplitude Modulation abbreviates 16QAM as) in two bit b of real part axle _nWith b _N+1, x wherein _j=3 ,-1,1, the b that 3} corresponds to _nb _N+1=11,10,00,01}.16 quadrature amplitudes of Fig. 4 A are tuned as the modulation mode that communication system adopted of Long Term Evolution (Long Term Evolution abbreviates LTE as).By Fig. 4 A as can be known, when the solid part signal that solves

The time, with b _nDifferent bit value is not at b _nThe next door, so modulating coefficient is

When the solid part signal that solves

The time, bit b _nPairing different bit value is on its next door, so modulating coefficient is

For bit b _N+1, how many solid part signal values that no matter solves is, bit b _N+1Pairing different bit value is all at its next door, therefore corresponding bit b _N+1Modulating coefficient be all

Adopting 64 quadrature amplitude modulations (abbreviating 64QAM as) with the communication system of a Long Term Evolution again be that example illustrates, as Fig. 4 B, when Fig. 4 B is communication system 64 quadrature amplitude modulations of employing (abbreviating 64QAM as) in three bit b of real part axle _n, b _N+1, b _N+2For bit b _n, when the solid part signal that solves

The time, modulating coefficient is When the solid part signal that solves

The time, modulating coefficient is

When the solid part signal that solves

The time, modulating coefficient is

When the solid part signal that solves

The time, modulating coefficient is

For bit b _N+1, when the solid part signal that solves

Or ± 1 o'clock, modulating coefficient is When the solid part signal that solves

Or ± 3 o'clock, modulating coefficient is

For bit b _N+2, modulating coefficient is all In view of the above, the signal x of j stratum _jThe signal x of wrong and other stratum _{I, i ≠ j}Each bit under all correct situation

The minimum Eustachian distance P of pairing all signal vectors of mistake and received signal vector y _{J, n}Can be expressed as

And as noted earlier, if permission inaccuracy slightly, then can be with the signal x of j stratum _jThe signal x of wrong and other stratum _{I, i ≠ j}Each bit under all correct situation

The minimum Eustachian distance of pairing all signal vectors of mistake and received signal vector y all is made as P _{J, n}=K * E[||h _j|| ²]+E[||n|| ²], wherein modulating coefficient K can be each

One of them, its weighted average or the numerical value between its maximum and minimum value.E represents the mean value of all elements of the number of winning the confidence vector, P _jWith P _{J, n}All be the Euclidean distance of being tried to achieve, difference is P _{J, n}In the K value can be with reference to two immediate distances of different bit value of different modulating methods, and P _jIn the K value with reference to the modulating coefficient of different modulating method correspondences.

Then, please refer to Fig. 5, Fig. 5 is the calcspar of the communication system that provided of embodiments of the invention.Communication system 500 has comprised conveyer 502 and receiver 504, and conveyer 502 sees through wireless channel 506 and carries out wireless telecommunications with receiver 504.Conveyer 502 has transmission antenna TX1～TX2 of two, and receiver RX has two reception antenna RX1～RX2.Suppose that solid part signal and imaginary signals can separate processes, and communication system 500 adopts 16 quadrature amplitude modulations, then the solid part signal vector of each transmission signal can have the signal of 2 stratum, similarly, the imaginary signals vector of each transmission signal also can have the signal of 2 stratum.Therefore, two are transmitted aerial signals and comprise the solid part signal vector and formed the signal vector with 8 stratum in this embodiment with the imaginary signals vector.Described real part received signal vector and imaginary part received signal vector can pass through and separate mapping, and obtain corresponding bit vector.

Please refer to Fig. 6, Fig. 6 is the soft calcspar of separating mapping device that present embodiment provides.Receiver 504 comprises signal detection module 550, channel estimating apparatus 560 and the soft mapping device 600 of separating, and the soft mapping device 600 of separating comprises computing unit 610, channel condition information acquiring unit 620, modulating coefficient amending unit 630, bit vector minimum range correspondence table module 640 and multiplexer 650.The soft mapping device 600 of separating is connected in signal detection module 550 and channel estimating apparatus 560, and wherein channel condition information acquiring unit 620 is connected in channel estimating apparatus 560, and bit vector minimum range correspondence table module 640 is connected in signal detection module 550.Computing unit 610 is connected in channel condition information acquiring unit 620, modulating coefficient amending unit 630 and multiplexer 650.Modulating coefficient amending unit 630 is connected in multiplexer 650, and bit vector minimum range correspondence table module 640 is connected in multiplexer 650.

Signal detection module 550 comprises in the S set of a plurality of signal vectors of all or part the signal vector near received signal vector y in order to searching

And Euclidean distance, signal detection module 550 is when searching S set, also might write down the Euclidean distance of other a plurality of signal vectors of part and received signal vector y, wherein said signal detection module 550 can for example be the most general like detecting module or hybrid sphere decoding module.Channel estimating apparatus 560 is in order to estimation wireless channel 506, to obtain the system channel matrix H.Channel condition information acquiring unit 620 is in order to obtain the vectorial h of row of each row of system channel matrix H _j

Other a plurality of signal vectors of part that bit vector minimum range correspondence table module 640 can be calculated according to signal detection module 550 and the Euclidean distance of received signal vector y, the signal vector that is solved

Set up out an incomplete bit vector minimum range correspondence table with the Euclidean distance of received signal vector y, the integrality of this incomplete bit vector minimum range correspondence table is according to the transmission signal set size of searching of seeking, if institute's transmission signal set of searching of seeking is too little, then the vectorial minimum range correspondence table of this bit is imperfect more.In order to save the complexity of calculating, modulating coefficient K can be each

One of them or its weighted average, in other words, modulating coefficient K can be some particular values.At this moment, enable signal EN is a low level unit, and modulating coefficient amending unit 630 is by forbidden energy, and multiplexer 650 is directly exported the P that computing unit 610 is calculated _{J, n}=K * E[||h _j|| ²]+E[||n|| ²] to bit vector minimum range correspondence table module 640.

Computing unit 610 can be according to front P _{J, n}Formula obtain NO null field in the incomplete bit vector minimum range correspondence table.That is to say the signal x of j stratum _jThe signal x of wrong and other stratum _{I, i ≠ j}Each bit b under all correct situation _nThe minimum Eustachian distance of pairing all signal vectors and received signal vector y can be by calculating P during mistake _{J, n}Formula obtain.But as noted earlier, modulating coefficient K can with the signal that is solved Modulating method relevant, therefore, can allow enable signal EN become high levle unit in order to increase soft accuracy of separating mapping.At this moment, modulating coefficient amending unit 630 is enabled, and multiplexer 650 is exported the P that modulating coefficient amending units 630 are revised _{J, n}Give bit vector minimum range correspondence table module 640, that is modulating coefficient amending unit 630 can see through multiplexer 650 outputs

Give bit vector minimum range correspondence table module 640.After setting up good complete bit vector minimum range correspondence table, bit vector minimum range correspondence table module 640 just can calculate each bit b according to complete bit vector minimum range correspondence table _nLogarithm general like ratio L (b _n), n=1 ..., N _TM _c

Be noted that the soft mapping device 600 of separating can carry out off-line (Off-Line) computing, also can carry out online (On-Line) computing.Described off-line operation is meant when signal detection module 550 is calculated and searched, and softly separates mapping device 600 calculates j stratum in advance according to channel estimation value and noise level and modulating method signal x _jThe signal x of wrong and other stratum _{I, i ≠ j}Each bit b under all correct situation _nThe minimum Eustachian distance P of pairing signal vector and received signal vector y during mistake _j, and in advance it is recorded in the pairing field of bit vector minimum range correspondence table.Then, by the time signal detection module 550 begins to calculate when searching, and calculates with the result of search according to signal detection module 550 again and softly separates the minimum Eustachian distance P that mapping device 600 is calculated _{J, n}Upgrade before according to minimum Eustachian distance P _jThe field of the bit of inserting vector minimum range correspondence table.

In addition, described on-line operation, wait until exactly signal detection module 550 begin to calculate with search after, softly separate the signal x that mapping device 600 just begins calculating j stratum _jThe signal x of wrong and other stratum _{I, i ≠ j}Each bit b under all correct situation _nThe minimum Eustachian distance P of pairing signal vector and received signal vector y during mistake _jOr P _{J, n}, again this value is filled up the null field in the bit vector minimum range correspondence table.

In addition, need to prove that when signal detection module 550 is the most general when the detecting module, signal detection module 550 can only store the signal vector of the vectorial y of the most approaching reception

And Euclidean distance.Therefore, bit vector minimum range correspondence table has only write down corresponding this signal vector The pairing a plurality of signal vectors of each bit and the minimum Eustachian distance of received signal vector y.The minimum Eustachian distance of pairing signal vector and received signal vector then must utilize aforementioned minimum Eustachian distance P during each bit mistake _jOr P _{J, n}Computing formula could obtain.

Then, please refer to Fig. 7, Fig. 7 is the flow chart of the soft de-mapping method that provides of embodiments of the invention.This soft de-mapping method is applicable to the receiver of communication system, can be used for setting up complete bit vector minimum range table, make receiver see through bit vector minimum range table and to access the logarithm of pairing each bit of received signal vector general like ratio.The receiver received signal vector

The institute that the conveyer of communication system transmitted might be expressed as by signal vector

And the signal x of each stratum _jCorresponding a plurality of bit [b _{(l-1) Mc+1}b _{(l-1) Mc+2}B _{(l-1) Mc+Mc}] ^T, wherein l=2j-1 is a real part, and l=2j is an imaginary part, and wherein j is 1 to N _TInteger, N _TBe the signal sum of received signal vector, M _cBe the signal x of j stratum _jPairing real part or imaginary part bit sum.

At first, in step S700, vectorial to received signal y carries out signal detection, for example hybrid sphere decoding or the most general like detecting, to obtain and the immediate signal vector of received signal vector y

When signal is detected, may while tracer signal vector

The Euclidean distance of other parts signal vector and received signal vector y.Then, in step S701, according to signal vector The Euclidean distance of other parts signal vector and received signal vector y is set up an incomplete bit vector minimum range correspondence table, the integrality of this incomplete bit vector minimum range correspondence table is according to the transmission signal set size of searching of seeking, if institute's transmission signal set of searching of seeking is too little, then the vectorial minimum range correspondence table of this incomplete bit is imperfect more.Be noted that when carrying out signal detection, only the tracer signal vector

Euclidean distance with received signal vector y.

Then, in step S702, according to the vectorial h of the row of system channel matrix H _j, calculate the signal x of j stratum _jThe signal x of wrong and other stratum _{I, i ≠ j}Each bit b under all correct situation _nThe minimum Eustachian distance P of pairing signal vector and received signal vector y during mistake _{J, n}, wherein n is 1 to N _TM _cInteger.In more detail, step S702 has the part that lacks to calculate its value at the incomplete bit vector minimum range correspondence table that is produced among the step S701, if the signal x of j stratum _jPairing part bit vector minimum range has and lacks, then according to the vectorial h of the row of system channel matrix H _j, calculate the signal x of j stratum _jThe signal x of wrong and other stratum _{I ≠ j}Each bit b under all correct situation _nThe minimum Eustachian distance P of pairing signal vector and received signal vector y during mistake _{J, n}

The order that is noted that step S701 and step S702 can exchange, and promptly step S702 can first rough estimate h _jAnd in correspondence table, deposit all bit vector minimum ranges in advance, if then in step S701, try to achieve certain bit vector minimum range, again the value in the bit vector minimum range correspondence table is upgraded.Then, in step S703, according to each the minimum Eustachian distance P that is tried to achieve _{J, n}Set up complete bit vector minimum range correspondence table with incomplete bit vector minimum range correspondence table.Afterwards, in step S704, calculate each bit b according to complete bit vector minimum range correspondence table _nLogarithm general like ratio L (b _n).

At this, need to prove, if permission inaccuracy slightly, then minimum Eustachian distance P _{J, n}Can equal minimum Eustachian distance P _j, promptly do not consider the distance of each bit nearest different value of various modulating methods with it.In addition, the flow chart of Fig. 7 is not in order to the qualification soft de-mapping method that present embodiment provided, and as noted earlier, the soft de-mapping method that embodiments of the invention provided can carry out off-line or on-line operation.Therefore, step S702 can just be performed earlier before step S700, at this moment described minimum Eustachian distance P _{J, n}Equal minimum Eustachian distance P _jAn or approaching value (can be the value or a mean value of a rough calculation).

Then, please refer to Fig. 8, Fig. 8 is the schematic diagram of bit vector minimum range correspondence table of providing of present embodiment.Communication system in this embodiment adopts 16 quadrature amplitude modulations, and receiver and conveyer all have 2 antennas, and solid part signal and imaginary signals are to be separated processing.Therefore, received signal vector y has the real number received signal y of 4 stratum ₁～y ₄, the signal vector x that conveyer may transmit also has the real number signal x of 4 stratum ₁～x ₄, i.e. received signal

Can be launched into

Transmit signal

Can be launched into

When SNR is very big, after the vectorial to received signal y of receiver carries out the approximate detecting of maximum, solves signal vector and can equal to transmit the number of delivering a letter certainly

And signal vector

With the Euclidean distance of received signal vector y be 0.000053.

Please contrast Fig. 4 A and Fig. 8 simultaneously,

Corresponding bit vector is Wherein

Correspond to

With

Other by that analogy.Can learn signal vector by Fig. 4 A

Pairing bit vector

Therefore the schematic diagram of the bit of Fig. 8 vector minimum range correspondence table is in bit b ₁=0, b ₂=0, b ₃=1, b ₄=0, b ₅=0, b ₆=0, b ₇=0, b ₈The minimum Eustachian distance of=1 o'clock pairing all signal vector and received signal vector y is all 0.00053.In addition, below for convenience of description, with signal vector

Order is

If will try to achieve at b ₁=1 and the time b ₂All signal vectors of=1 correspondence and the minimum Eustachian distance of received signal vector y can be supposed the signal x of first stratum ₁Separate mistake, the signal x of other stratum ₂, x ₃, x ₄All correct, then can be according to aforementioned P _{J, n}Formula calculate bit b ₁With b ₂The minimum Eustachian distance of pairing all signal vectors and received signal vector y during mistake.In the example of Fig. 8, the bit that is solved

Therefore if bit b ₁Wrong, b then ₁=1 with the bit that solves

The signal of correspondence is x when the most close ₁=-1, so modulating coefficient is

In addition, the bit that is solved

If bit b ₂Wrong, b then ₂=1 with the bit that solves The signal of correspondence is x when the most close ₁=3, so modulating coefficient is

In view of the above, the minimum Eustachian distance that is calculated And minimum Eustachian distance

Can be filled to bit b ₁=1 and b ₂The field at=1 place.

Yet, if truly to bit b ₁Pairing all signal vectors are searched during mistake, then can obtain signal vector x=[-1-1 1 3] ^TBe bit b ₁During mistake in pairing all signal vectors near the signal vector of received signal vector y, and signal vector x=[-1-1 1 3] ^TWith the distance of received signal vector y be 0.425749.In addition, if truly to bit b ₂Pairing all signal vectors are searched during mistake, then can obtain signal vector x=[3-1 1 3] ^TBe bit b ₂During mistake in pairing all signal vectors near the signal vector of received signal vector y, and signal vector x=[3-1 1 3] ^TWith the distance of received signal vector y be 0.428814.Therefore, use the minimum Eustachian distance that formula calculated that embodiments of the invention provided

Very little with actual minimum euclidean distance error.

If will try to achieve at b ₃=0 and the time b ₄All signal vectors of=1 correspondence and the minimum Eustachian distance of received signal vector y can be supposed the signal x of second stratum ₂Separate mistake, the signal x of other stratum ₁, x ₃, x ₄All correct, then can be according to aforementioned P _{J, n}Formula calculate bit b ₃With b ₄The minimum Eustachian distance of pairing all signal vectors and received signal vector y during mistake.In the example of Fig. 8, the bit that is solved

Therefore if bit b ₃Wrong, b then ₃=0 with the bit that solves

The signal of correspondence is x when the most close ₂=1, so modulating coefficient is

In addition, the bit that is solved

Therefore if bit b ₄Wrong, b then ₄=1 with the bit that solves

The signal of correspondence is x when the most close ₂=-3, so modulating coefficient is In view of the above, the minimum Eustachian distance that is calculated

And minimum Eustachian distance

Can be filled to bit b ₃=0 and b ₄The field at=1 place.

Yet, if truly to bit b ₃Pairing all signal vectors are searched during mistake, then can obtain signal vector x=[1 11 3] ^TBe bit b ₃During mistake in pairing all signal vectors near the signal vector of received signal vector y, and signal vector x=[1 11 3] ^TWith the distance of received signal vector y be 0.436261.In addition, if truly to bit b ₄Pairing all signal vectors are searched during mistake, then can obtain signal vector x=[1-3 1 3] ^TBe bit b ₄During mistake in pairing all signal vectors near the signal vector of received signal vector y, and signal vector x=[1-3 1 3] ^TWith the distance of received signal vector y be 0.418302.Therefore, use the minimum Eustachian distance that formula calculated that embodiments of the invention provided Very little with actual minimum euclidean distance error.

If will try to achieve at b ₁=1 and the time b ₂All signal vectors of=1 correspondence and the minimum Eustachian distance of received signal vector y can be supposed the signal x of three-layered ₃Separate mistake, the signal x of other stratum ₁, x ₂, x ₄All correct, then can be according to aforementioned P _{J, n}Formula calculate bit b ₅With b ₆The minimum Eustachian distance of pairing all signal vectors and received signal vector y during mistake.In the example of Fig. 8, the bit that is solved

If bit b ₅Wrong, b then ₅=1 with the bit that solves The signal of correspondence is x when the most close ₃=-1, so modulating coefficient is

In addition, the bit that is solved If bit b ₆Wrong, b then ₆=1 with the bit that solves

The signal of correspondence is x when the most close ₃=3, so modulating coefficient is

In view of the above, the minimum Eustachian distance that is calculated

And minimum Eustachian distance

Can be filled to bit b ₅=1 and b ₆The field at=1 place.

Yet, if truly to bit b ₅Pairing all signal vectors are searched during mistake, then can obtain signal vector x=[1-1-1 3] ^TBe bit b ₅During mistake in pairing all signal vectors near the signal vector of received signal vector y, and signal vector x=[1-1-1 3] ^TWith the distance of received signal vector y be 0.298697.In addition, if truly to bit b ₆Pairing all signal vectors are searched during mistake, then can obtain signal vector x=[1-1 3 3] ^TBe bit b ₆During mistake in pairing all signal vectors near the signal vector of received signal vector y, and signal vector x=[1-1 3 3] ^TWith the distance of received signal vector y be 0.291612.Therefore, use the minimum Eustachian distance that formula calculated that embodiments of the invention provided

Very little with actual minimum euclidean distance error.

If will try to achieve at b ₇=1 and the time b ₈All signal vectors of=0 correspondence and the minimum Eustachian distance of received signal vector y can be supposed the signal x of quadravalence layer ₄Separate mistake, the signal x of other stratum ₁, x ₂, x ₃All correct, then can be according to aforementioned P _{J, n}Formula calculate bit b ₇With b ₈The minimum Eustachian distance of pairing all signal vectors and received signal vector y during mistake.In the example of Fig. 8, the bit that is solved

Therefore if bit b ₇Wrong, b then ₇=1 with the bit that solves

The signal of correspondence is x when the most close ₄=-1, so modulating coefficient is In addition, the bit that is solved Therefore if bit b ₈Wrong, b then ₈=0 with the bit that solves The signal of correspondence is x when the most close ₄=1, so modulating coefficient is

In view of the above, the minimum Eustachian distance that is calculated

With

And minimum Eustachian distance

With

Can be filled to bit b respectively ₇=1 and b ₈The field at=0 place.

Yet, if truly to bit b ₇Pairing all signal vectors are searched during mistake, then can obtain signal vector x=[1-3 1-1] ^TBe bit b ₇During mistake in pairing all signal vectors near the signal vector of received signal vector y, and signal vector x=[1-3 1-1] ^TWith the distance of received signal vector y be 1.097965.In addition, if truly to bit b ₈Pairing all signal vectors are searched during mistake, then can obtain signal vector x=[1-1 1 1] ^TBe bit b ₈During mistake in pairing all signal vectors near the signal vector of received signal vector y, and signal vector x=[1-1 1 1] ^TWith the distance of received signal vector y be 0.295709.Apparently, bit b ₇During mistake in pairing all signal vectors near the signal vector x=[1-3 1-1 of received signal vector y] ^TBe not to have only quadravalence layer signal mistake, therefore just can cause the minimum Eustachian distance that uses formula to calculate

With actual minimum Eustachian distance bigger error is arranged, but howsoever, even the error of part is arranged, the soft de-mapping method that present embodiment provided, still can be used as is P _4,7 ¹Reference value.In addition, use the minimum Eustachian distance that formula calculated that embodiments of the invention provided

Very little with actual minimum euclidean distance error.

Be noted that will obtain minimum Eustachian distance more quickly can also use P _jFormula calculate, but clearly, in this embodiment, the minimum Eustachian distance P that is calculated ₄Meeting and bit b ₁ ⁴With the minimum Eustachian distance of received signal vector y bigger error is arranged in pairing all signal vectors during mistake.In view of the above, promote the accuracy of calculating, the modulating coefficient that can use change is with formula P _jBe modified to formula P _{J, k}

In sum, the soft de-mapping method that provides of the present invention can be applied in the receiver of various unlike signal detecting modes, and described soft de-mapping method can utilize simple formula to obtain null field in the incomplete bit vector minimum range correspondence table, and is general like ratio with the logarithm that successfully calculates each bit whereby.

In conjunction with the drawings to the description of the specific embodiment of the invention, others of the present invention and feature are conspicuous to those skilled in the art.

More than specific embodiments of the invention are described and illustrate it is exemplary that these embodiment should be considered to it, and be not used in and limit the invention, the present invention should make an explanation according to appended claim.

Claims (24)

1. a soft de-mapping method is characterized in that, is applicable to a receiver of a communication system, and general like ratio in order to a logarithm that obtains pairing each bit of a received signal vector, described receiver receives described received signal vector

The institute that one conveyer transmitted of described communication system might be expressed as by signal vector

And the signal x of each stratum _jCorresponding a plurality of bits are [b _{(l-1) Mc+1}b _{(l-1) Mc+2}B _{(l-1) Mc+Mc}] ^T, wherein l=2j-1 is a real part, and l=2j is an imaginary part, and j is 1 to N _TInteger, N _TBe a signal sum of described received signal vector, M _cBe the signal x of j stratum _jPairing real part or imaginary part bit sum, described soft de-mapping method comprises:

Described received signal vector y is carried out a signal detection, to obtain incomplete bit vector minimum range correspondence table;

The vectorial h of each row according to a system channel matrix H _j, calculate the signal x of j stratum _jThe signal x of wrong and other stratum _{I, i ≠ j}Each bit b under all correct situation _nThe minimum Eustachian distance P of pairing signal vector and received signal vector y during mistake _{J, n}And

According to each minimum Eustachian distance P _{J, n}Set up complete bit vector minimum range correspondence table with incomplete bit vector minimum range correspondence table.

2. soft de-mapping method according to claim 1 is characterized in that, described signal detection is a hybrid sphere decoding or the most general seemingly detecting.

3. soft de-mapping method according to claim 1 is characterized in that, when described signal is detected, solves the immediate signal vector with described received signal vector y

And write down described signal vector at least

With the minimum Eustachian distance of described received signal vector y, to set up out incomplete bit vector minimum range correspondence table whereby.

4. soft de-mapping method according to claim 3 is characterized in that, when described signal is detected, has more write down the Euclidean distance of part signal vector with described received signal vector y, and according to described signal vector

Set up incomplete bit vector minimum range correspondence table with Euclidean distance, the part signal vector of described received signal vector y with the Euclidean distance of described received signal vector y.

5. soft de-mapping method according to claim 1 is characterized in that, more comprises:

Calculate each bit b according to complete bit vector minimum range correspondence table _nLogarithm general like ratio L (b _n).

6. soft de-mapping method according to claim 1 is characterized in that, this P of described calculating _{J, n}Step can carry out this signal detection with the step that obtains incomplete this bit vector minimum range correspondence table before, at this moment, first rough estimate h _jAnd in correspondence table, deposit all bit vector minimum ranges in advance, then in the step of this signal detection,, the value in this bit vector minimum range correspondence table is being upgraded if try to achieve certain bit vector minimum range.

7. soft de-mapping method according to claim 1 is characterized in that, described minimum Eustachian distance P _{J, n}=K * E[||h _j|| ²]+E[||n|| ²], wherein n is a noise vector, and E is the mean value of all elements of the number of winning the confidence vector, and K is a modulating coefficient.

8. soft de-mapping method according to claim 1 is characterized in that, described minimum Eustachian distance

Wherein n is a noise vector, and E is the mean value of all elements of the number of winning the confidence vector,

Be modulating coefficient,

The signal that is solved with described communication system

And the modulating method of wrong bit is relevant with its position at planisphere.

9. the soft mapping device of separating is characterized in that, is applicable to a receiver of a communication system, and general like ratio in order to a logarithm that obtains pairing each bit of a received signal vector, described receiver receives described received signal vector

The institute that one conveyer transmitted of described communication system might be expressed as by signal vector

And the signal x of each stratum _jCorresponding a plurality of bits are [b _{(l-1) Mc+1}b _{(l-1) Mc+2}B _{(l-1) Mc+Mc}] ^T, wherein l=2j-1 is a real part, and l=2j is an imaginary part, and j is 1 to N _TInteger, N _TBe a signal sum of described received signal vector, M _cBe the signal x of j stratum _jPairing real part or imaginary part bit sum, the described soft mapping device of separating comprises:

One bit vector minimum range correspondence table module, the result who described received signal vector y is carried out a signal detection according to a signal detection module is to obtain incomplete bit vector minimum range correspondence table;

One channel condition information acquiring unit, from a channel estimating apparatus obtain a system channel matrix H each the row vectorial h _jAnd

One computing unit is according to the vectorial h of each row of described system channel matrix H _j, calculate the signal x of j stratum _jThe signal x of wrong and other stratum _{I, i ≠ j}Each bit b under all correct situation _nThe minimum Eustachian distance P of pairing signal vector and received signal vector y during mistake _{J, n}

Wherein said bit vector minimum range correspondence table module is more according to each minimum Eustachian distance P _{J, n}Set up complete bit vector minimum range correspondence table with incomplete bit vector minimum range correspondence table.

10. the soft mapping device of separating according to claim 9 is characterized in that, described signal detection module is a hybrid sphere decoding module or the most general seemingly detecting module.

11. the soft mapping device of separating according to claim 9 is characterized in that, when described signal was detected, described signal detection module solved the immediate signal vector with described received signal vector y

And write down described signal vector at least

With the Euclidean distance of described received signal vector y, to allow described bit vector minimum range correspondence table module set up out the vectorial minimum range correspondence table of incomplete unit whereby.

12. the soft mapping device of separating according to claim 11, it is characterized in that, when described signal is detected, described signal detection module has more write down the Euclidean distance of part signal vector and described received signal vector y, described bit vector minimum range correspondence table module and according to described signal vector Set up the vectorial minimum range correspondence table of incomplete unit with Euclidean distance, the part signal vector of described received signal vector y with the Euclidean distance of described received signal vector y.

13. the soft mapping device of separating according to claim 9 is characterized in that, described bit vector minimum range correspondence table module is calculated each bit b according to complete bit vector minimum range correspondence table _nLogarithm general like ratio L (b _n).

14. the soft mapping device of separating according to claim 9 is characterized in that described computing unit just calculated the signal x of j stratum before described signal detection is carried out _jThe signal x of wrong and other stratum _{I, i ≠ j}Each bit b under all correct situation _nThe minimum Eustachian distance P of pairing signal vector and received signal vector y during mistake _{J, n}, that is this computing unit can first rough estimate h _jAnd in this bit vector minimum range correspondence table, deposit all bit vector minimum ranges in advance, then, if bit vector minimum range correspondence table module is tried to achieve certain bit vector minimum range, again the value in this bit vector minimum range correspondence table is upgraded; Perhaps, described computing unit just calculates the signal x of j stratum after described signal detection is carried out _jThe signal x of wrong and other stratum _{I, i ≠ j}Each bit b under all correct situation _nThe minimum Eustachian distance P of pairing signal vector and received signal vector y during mistake _{J, n}

15. the soft mapping device of separating according to claim 9 is characterized in that, described minimum Eustachian distance P _{J, n}=K * E[||h _j|| ²]+E[||n|| ²], wherein n is a noise vector, and E is the mean value of all elements of the number of winning the confidence vector, and K is a modulating coefficient.

16. the soft mapping device of separating according to claim 15 is characterized in that, more comprises:

One modulating coefficient amending unit is in order to the described minimum Eustachian distance P with described computing unit calculated _{J, n}=K * E[||h _j|| ²]+E[||n|| ²] be modified to minimum Eustachian distance

Wherein n is a noise vector, and E is the mean value of all elements of the number of winning the confidence vector,

Be modulating coefficient,

The signal that is solved with described communication system

And the modulating method of wrong bit is relevant with its position at planisphere.

17. communication system, it is characterized in that, comprise a receiver and a conveyer, described receiver comprises a soft mapping device and the signal detection module of separating, the described soft mapping device of separating is in order to the general seemingly ratio of a logarithm that obtains pairing each bit of a received signal vector, and described receiver receives described received signal vector

The institute that one conveyer transmitted of described communication system might be expressed as by signal vector

And the signal x of each stratum _jCorresponding a plurality of bits are [b _{(l-1) Mc+1}b _{(l-1) Mc+2}B _{(l-1) Mc+Mc}] ^T, wherein l=2j-1 is a real part, and l=2j is an imaginary part, and j is 1 to N _TInteger, N _TBe a signal sum of described received signal vector, M _cBe the signal x of j stratum _jPairing real part or imaginary part bit sum, the described soft mapping device of separating comprises:

One bit vector minimum range correspondence table module, the result who described received signal vector y is carried out a signal detection according to a signal detection module is to obtain incomplete bit vector minimum range correspondence table;

One channel condition information acquiring unit, from a channel estimating apparatus obtain a system channel matrix H each the row vectorial h _jAnd

One computing unit is according to the vectorial h of each row of described system channel matrix H _j, calculate the signal x of j stratum _jThe signal x of wrong and other stratum _{I, i ≠ j}Each bit b under all correct situation _nThe minimum Eustachian distance P of pairing signal vector and received signal vector y during mistake _{J, n}

Wherein said bit vector minimum range correspondence table module is more according to each minimum Eustachian distance P _{J, n}Set up complete bit vector minimum range correspondence table with incomplete bit vector minimum range correspondence table.

18. communication system according to claim 17 is characterized in that, described signal detection module is a hybrid sphere decoding module or the most general seemingly detecting module.

19. communication system according to claim 17 is characterized in that, when described signal was detected, described signal detection module solved the immediate signal vector with described received signal vector y

And write down described signal vector at least

With the Euclidean distance of described received signal vector y, to allow described bit vector minimum range correspondence table module set up out the vectorial minimum range correspondence table of incomplete unit whereby.

20. communication system according to claim 19, it is characterized in that, when described signal was detected, described signal detection module had more write down the Euclidean distance of part signal vector and described received signal vector y, described bit vector minimum range correspondence table module and according to described signal vector

Set up incomplete bit vector minimum range correspondence table with Euclidean distance, the part signal vector of described received signal vector y with the Euclidean distance of described received signal vector y.

21. communication system according to claim 17 is characterized in that, described bit vector minimum range correspondence table module is calculated each bit b according to complete bit vector minimum range correspondence table _nLogarithm general like ratio L (b _n).

22. communication system according to claim 17 is characterized in that, described computing unit just calculated the signal x of j stratum before described signal detection is carried out _jThe signal x of wrong and other stratum _{I, i ≠ j}Each bit b under all correct situation _nThe minimum Eustachian distance P of pairing signal vector and received signal vector y during mistake _{J, n}, that is this computing unit can first rough estimate h _jAnd in this bit vector minimum range correspondence table, deposit all bit vector minimum ranges in advance, then, if bit vector minimum range correspondence table module is tried to achieve certain bit vector minimum range, again the value in this bit vector minimum range correspondence table is upgraded; Perhaps, described computing unit just calculates the signal x of j stratum after described signal detection is carried out _jThe signal x of wrong and other stratum _{I, i ≠ j}Each bit b under all correct situation _nThe minimum Eustachian distance P of pairing signal vector and received signal vector y during mistake _{J, n}

23. communication system according to claim 17 is characterized in that, described minimum Eustachian distance P _{J, n}=K * E[||h _j|| ²]+E[||n|| ²], wherein n is a noise vector, and E is the mean value of all elements of the number of winning the confidence vector, and K is a modulating coefficient.

24. communication system according to claim 23 is characterized in that, the described soft mapping device of separating more comprises:

One modulating coefficient amending unit is in order to the described minimum Eustachian distance P with described computing unit calculated _{J, n}=K * E[||h _j|| ²]+E[||n|| ²] be modified to minimum Eustachian distance

Wherein n is a noise vector,

Be modulating coefficient, E is the mean value of all elements of the number of winning the confidence vector,

The signal that is solved with described communication system

And the modulating method of wrong bit is relevant with its position at planisphere.

Images