CN100401656C - Method for transmitting code channel information using training sequence in time-division duplex multi-antenna system - Google Patents

Abstract

The present invention relates to a method for transmitting code channel information in time-division duplex multi-antenna systems, which can simultaneously support various spread spectrum coefficients and even under the condition of various antenna configurations, a training sequence is used for transmitting code channel information. A sending end is used for carrying the code channel information in the added training sequence. The present invention comprises that mapping relationships are established among a training sequence deflection amount, layer sequence numbers and the number of communication channel simplified codes; the mapping relationships are irrelevant to the spread spectrum coefficients; the training sequence of one training sequence deflection value is added in each layer of data according to the mapping relationships; different deflection values of the training sequence are added in different layers; the sizes of the deflection values correspond to the sizes of the layer sequence numbers and the sizes of scrambling sequence numbers according to the mapping relationships and correspond to the number of the communication channel simplified codes. A receiving end obtains the layer sequence numbers of code channel layers, the scrambling sequence numbers, and the number of the communication channel simplified codes used at present by obtained each layer of data deflection value of the training sequence according to the mapping relationships identical with the sending end, and can obtain scrambling codes and the communication channel simplified codes.

Description

Time-division duplex multi-antenna system utilizes training sequence to transmit the method for code channel information

Technical field

The present invention relates to mobile communication technology, relate in particular to and utilize training sequence to transmit the method for code channel information in the 3-G (Generation Three mobile communication system) under the TDD mode, be when under TDD mode, using multi-antenna technology to carry out high speed data transfer, to utilize training sequence to transmit the method for code channel information.

Background technology

Code channel information comprises the sequence number information (layer sequence number) of the affiliated layer of channelization code information, scrambling code information and this code channel that code channel adopts.

In time-division duplex (CDMA) mobile communication system, Data Detection adopts Multiuser Detection (MUD) technology usually, in order to detect the data bit on all activated code channel simultaneously, receiver need be known the spreading code of current all activated code channel.Wherein, spreading code is made of jointly channel code and scrambler.

For traditional single antenna mobile communication system, a transmitter is assigned with a scrambler, simultaneously receiver is also preserved this scrambler accordingly, has just known the spreading code of each code channel so receiver only need be known channel code on each code channel, just can carry out Multiuser Detection.

CDMA mobile communication system for time division duplex, utilizing its distinctive frame structure to carry out data sends, specifically, the unoccupied place burst sent when the data flow under the TDD mode was one of a time slot, two parts transmission data and training sequence placed in the middle before and after wherein each burst structure comprises, the main application of training sequence is to carry out channel estimating.A transmitter distributes a basic training sequence usually, and the training sequence that is inserted in the burst structure produces (by the decision of training sequence deviation value) by this basic sequence by skew.Channelization code information can be utilized training sequence transmission generally speaking, and transfer mode can have general and default two kinds.Wherein general mode is meant when base station down sends, all code channels use identical training sequence, and inform that with the deviant between this training sequence and the basic training sequences how many channel codes are terminal have in use, such as when the total K=16 of training sequence, the training sequence after each skew can be used symbol m _i(i=1 ... 16) expression is m at the training sequence that distributes so ₁₆The time, promptly the number of channelization codes in the use of announcement is 16; Default mode is the one-to-one relationship of setting up between training sequence deviation value and the channel code, each code channel uses the training sequence of different deviants, terminal is by this one-to-one relationship, obtains the channel code of all code channels by training sequence, such as training sequence m _iCorresponding with i channel code.

Along with the appearance of space multiplexing technique, the method for utilizing a plurality of transmitting antennas, a plurality of reception antenna to carry out high speed data transfer is just becoming a kind of trend of following wireless communication technology development.This technology is applicable to the wireless environment of multiple scattering body, and feasible signal from each transmitting antenna is highly incoherent in each reception antenna, and utilizes this irrelevance that the data of a plurality of antenna transmission are distinguished and detected at receiver end.If the data difference that sends on a plurality of antennas, be called spatial reuse, as the data flow of distributing to all burst structure of a time slot is S, then earlier data flow S is resolved into the X layer data, each layer data is decomposed into Y substream of data according to the channelizing number again, and (X is the transmitting antenna sum, Y is the code channel number corresponding to every antenna), every Y substream of data is launched corresponding to a transmit antennas, and this Y substream of data is called as a layer data.

For the multiaerial system under the TDD mode, receiver need know that equally the spreading code information of each activated code channel comes the data of all code channels are detected, the special order that also needs to know each layer data is so that can recover the transmission data according to correct order (layer sequence number) after detecting each layer data.

When utilizing many antennas to carry out high speed data transfer, in order to obtain the channel impulse response from different transmit antennas at receiver end, every layer needs to distribute different training sequences so that channel estimating to be provided.In the ordinary course of things, the data that send on every antenna are all inequality, and this moment, layer was one to one with transmitting antenna, was assigned to X the situation on the transmitting antenna as above-mentioned X * Y substream of data.But in some cases, number of transmit antennas can be greater than sending the number of plies, have many antennas and send one deck data conditions simultaneously this moment, and for receiver, owing to only need know that this layer data arrives the channel estimation value on certain root reception antenna, distribute different training sequences for every layer so only require assurance.

When utilizing training sequence transfer layer sequence number indication information, scrambling code information, channelization code information, can distribute training sequence according to following principle at transmitter terminal: the training sequence deviation value satisfies the first predetermined mapping relations with layer and (characterizes a layer sequence number as the size of utilizing every layer of training sequence deviation value, be deviant arrange from small to large represent that respectively ground floor, the second layer are up to the M layer), so that receiver end is known the order of layer by these mapping relations, thereby merges multi-layer data exactly; Use the different scrambling codes transmission at different layers, at least distribute two different scramblers, and when requiring in a scrambler group of determining, to select scrambler, the sequence number of training sequence deviation value and scrambler should satisfy the second predetermined mapping relations, and (ground floor is corresponding to first scrambler, the second layer is corresponding to second scrambler, by that analogy), so that receiver end is known the scrambling code information of different layers by these mapping relations; The selective channel sign indicating number also divides timing regularly from a channelizing code character of determining requiring, the quantity of training sequence deviation value and channel code should satisfy the 3rd predetermined mapping relations (in the same channelizing code character sequence number come the front by priority allocation) so that receiver end is known channelization code information by these mapping relations.

Particularly, set up first mapping relations with the size of training sequence deviation value and the size of layer sequence number, promptly deviant arrange from small to large represent ground floor, the second layer respectively, up to all layers, vice versa; When different layers used different scrambler transmission, the information of scrambler also can make in this way and transmit, and promptly ground floor is corresponding to first scrambler, and the second layer is corresponding to second scrambler, and the rest may be inferred.Therefore for second mapping relations, it and first mapping relations are equal to sometimes, and then not needing to set up second mapping relations, the 3rd mapping relations sometimes is to set up satisfying under the condition of first, second mapping relations.

Carry out training sequence according to above-mentioned basic principle and divide timing, can reflect above-mentioned three kinds of mapping relations simultaneously by setting up a training sequence mapping form, as shown in table 1 below.What this table was corresponding is that a transmission number of data layers is 2 (M=2), and spread spectrum coefficient is the mapping form of 16 (SF=16).Wherein, a kind of code channel numerical value of line display situation, totally 16 row represent that respectively 16 kinds take the code channel situation; Preceding 8 row in the delegation, represent that 8 kinds of different deviants and the ratio between the minimum offset values are (because minimum offset values is fixed, known that the ratio between deviant and the minimum offset values just known deviant), if " X " symbol is arranged in certain cell, represent that this ratio is selected, the training sequence deviation value of this layer data is promptly determined so.In the 2nd row, ratio 1,2 is selected, has then determined the 1st layer, the training sequence m of the 2nd layer data ₁, m ₂And for example in the 4th row, ratio 3,4 is selected, has then determined the 1st layer, the training sequence m of the 2nd layer data ₃, m ₄Last tabulation in the form is shown in preceding 8 row code channel number corresponding under " X " symbol distribution situation, and above-mentioned two the code channel numbers that take in for example are respectively 2,4.Receiver end also needs to preserve such mapping form, so that carry out the inverse mapping operation.

Table 1

The training sequence of setting up according to the basic principle of above-mentioned training sequence distribution shines upon form, utilize training sequence to transmit layer sequence number information, scrambling code information and channelization code information, promptly known training sequence, just know the training sequence deviation value, thereby known a layer sequence number, scrambler and channel code.

Yet, when spread spectrum coefficient changes, it is different that training sequence mapping form has institute, need the mobile communication system of adaptively changing this moment for a spread spectrum coefficient, receiver just needs to preserve many such training sequence mapping forms corresponding with spread spectrum coefficient, thereby causes the waste of memory space.

In addition, when the transmission number of data layers is inequality, its training sequence mapping form also can be different, for a system that needs adaptively changing to send number of data layers, receiver also needs to preserve many such training sequence mapping forms corresponding with sending number of data layers, also causes the waste of memory space.

Though by adopting the structure of mapping form 1, can utilize training sequence to transmit code channel information, but in use there is defective, promptly need the system of adaptively changing, the system that the transmission number of data layers needs adaptively changing at spread spectrum coefficient, then need preserve and each spread spectrum coefficient, the corresponding training sequence mapping form of each transmission number of data layers, increase the memory data output of receiver end at receiver end.

Summary of the invention

The objective of the invention is to design a kind of time-division duplex multi-antenna system utilizes training sequence to transmit the method for code channel information, promptly provide and to support to utilize under multiple spread spectrum coefficient even the multiple antenna configurations situation training sequence to transmit the method for code channel information simultaneously, to reduce the memory cell of receiver.

The technical scheme that realizes the object of the invention is such: a kind of time-division duplex multi-antenna system utilizes training sequence to transmit the method for code channel information, transmitting terminal resolves into layer data with the data flow of each time slot, each layer data is decomposed into a plurality of substream of data according to number of channelization codes, each substream of data in each layer data is added training sequence form burst structure, in the training sequence that adds, carry code channel information; Receiving terminal carries out channel estimating to the data that receive, and obtains the training sequence of each layer, obtains code channel information according to training sequence, it is characterized in that:

Described transmitting terminal adds training sequence and forms burst structure, carries code channel information in the training sequence that adds, and comprising:

A. in the training sequence deviation value, set up mapping relations between layer sequence number and number of channelization codes, these mapping relations comprise the mapping relations between training sequence deviation value size and layer sequence number size, mapping relations in layer sequence number size and the scrambler group between scrambler sequence number size, and training sequence deviation value of distributing on the different layers and the mapping relations between number of channelization codes, when setting up mapping relations, described mapping relations when at first setting up a certain spread spectrum coefficient, when spread spectrum coefficient increases, the new mapping relations that keep the constant and correspondence expansion spread spectrum coefficient that increases when the expansion spread spectrum coefficient increases on this mapping relations basis of aforementioned mapping relations, when the new mapping relations of the described corresponding expansion spread spectrum coefficient of expansion, guarantee that the employed training sequence deviation value of each mapping relations is no more than the deviant that quantity limited of its minimum training sequence deviation value, and guarantee that the training sequence deviation value between each mapping relations does not repeat;

B. in each substream of data of each layer data, add a training sequence that has deviant with respect to basic training sequences by above-mentioned mapping relations;

Described receiving terminal obtains code channel information according to training sequence, comprising:

C. obtain the training sequence deviation value of each layer by the training sequence of each layer, the code channel information of layer sequence number, scrambler and the channel code that comprises under the code channel layer according to the training sequence deviation value of distributing on mapping relations between scrambler sequence number size in the mapping relations between the training sequence deviation value identical size and layer sequence number size, layer sequence number size and the scrambler group and the different layers and the acquisition of the mapping relations between number of channelization codes with transmitting terminal.

Described steps A is to set up mapping relations between training sequence deviation value, layer sequence number, scrambler sequence number and number of channelization codes according to the more little principle of the more little employed training sequence deviation value of number of channelization codes, comprising:

The principle of A1. distributing different deviant training sequences by different layers is set up the first predetermined mapping relations between training sequence deviation value and the layer sequence number;

A2. by different layers distribute different scrambling codes, requirement scrambler in a scrambler group of determining, select and this scrambler group when receiver end is known, set up second mapping relations of being scheduled between the sequence number of scrambler in the scrambler group and the layer sequence number;

A3. require in the channelizing code character that determines the selective channel sign indicating number and distribute and this channelizing code character and distribute rule when receiver end is known, set up the 3rd preset mapping relation between the quantity information of the training sequence deviation value of distributing on the different layers and channel code according to the more little principle of the more little employed training sequence deviation value of number of channelization codes by rule.

In the described steps A 3, set up the 3rd preset mapping relation between the quantity information of the training sequence deviation value of distributing on the different layers and channel code according to the more little principle of the more little employed training sequence deviation value of number of channelization codes, comprising:

A311. all number of channelization codes are divided into groups according to the needed minimum spread spectrum coefficient Qmin of each number of channelization codes N;

A312. according to formula $C_K^M\≥Q\min ,$ M is the desired actual number of data layers that sends, calculate the divide into groups minimum number Kmin of desired training sequence deviation value of every minimum spread spectrum coefficient, according to each Kmin, on steps A 311 packet by packet basis each number of channelization codes N is once divided into groups again, K is the quantity of training sequence deviation value;

A313. according to the minimum number Kmin order from small to large of training sequence deviation value, distribute training sequence by the desired actual number of data layers that sends in each Qmin grouping in each Kmin grouping, the deviant of the training sequence that each Qmin grouping of branch timing is used must not surpass the peak excursion value that Kmin limited of its grouping, and the training sequence deviation value of distribution between each Qmin grouping does not repeat.

In the described steps A 3, set up the 3rd preset mapping relation between the quantity information of the training sequence deviation value of distributing on the different layers and channel code according to the more little principle of the more little employed training sequence deviation value of number of channelization codes, comprising:

A321. all number of channelization codes are divided into groups according to the needed minimum spread spectrum coefficient Qmin of each number of channelization codes N;

A322. to send number of data layers M=2, according to formula $C_K^M\≥Q\min ,$ Calculate the divide into groups minimum number Kmin of desired training sequence deviation value of every minimum spread spectrum coefficient Qmin, each number of channelization codes N is once divided into groups on steps A 321 packet by packet basis according to each Kmin again, K is the quantity of training sequence deviation value;

A323. according to the minimum number Kmin order from small to large of training sequence deviation value, distribute training sequence by sending number of data layers M=2 in each Qmin grouping in each Kmin grouping, it is less that the branch timing at first travels through required training sequence deviation value minimum number Kmin, the mapping situation of all number of channelization codes N and training sequence deviation value, and then it is bigger to travel through required training sequence deviation value minimum number Kmin, the mapping situation of all number of channelization codes N and training sequence deviation value, when the desired actual number of data layers M that sends greater than 2 the time, further execution in step A324;

A324. on the basis that steps A 323 is distributed, according to the desired actual number of data layers M that sends, be the 3rd, 4 ... the M layer distributes training sequence, promptly for some number of channelization codes N, select the training sequence that deviant is positioned at after the training sequence deviation value two-layer before distributing to is distributed to the 3rd, 4 ... the M layer.

When needs guarantee to satisfy W 〉=Wmin, according to formula

Obtain required maximum training sequence deviation value number K max, allocation result to steps A 323 or A324 is revised, keep and satisfy the training sequence deviation value of formula K≤Kmax and the mapping relations between number of channelization codes N, give up the training sequence deviation value that do not satisfy formula K≤Kmax and the mapping relations between number of channelization codes N, and each number of channelization codes N that will give up part sets up mapping relations with the training sequence deviation value that satisfies formula K≤Kmax, be about to one and take the less training sequence method of salary distribution of training sequence deviation value and a plurality of number of channelization codes N sets up mapping relations, P is the length of training sequence in the formula, and W is that the window of channel estimating is long.

In the described steps A 3, set up the 3rd preset mapping relation between the quantity information of the training sequence deviation value of distributing on the different layers and channel code according to the more little principle of the more little employed training sequence deviation value of number of channelization codes, comprising:

A331. all number of channelization codes are divided into groups according to the needed minimum spread spectrum coefficient Qmin of each number of channelization codes N;

A332. to send number of data layers M=2, according to formula $C_K^M\≥Q\min ,$ Calculate the divide into groups minimum number Kmin of desired training sequence deviation value of every minimum spread spectrum coefficient Qmin, each number of channelization codes N is once divided into groups on steps A 331 packet by packet basis according to each Kmin again, K is the quantity of training sequence deviation value;

A333. at the different training sequence m of deviant _i, i=1 ... 16, and set up mapping relationship f (m between some numeral _i), according to the minimum number Kmin order from small to large of training sequence deviation value, distribute training sequence by sending number of data layers M=2 in each Qmin grouping in each Kmin grouping, divide timing to guarantee f (m ₍₁₎)+f (m ₍₂₎)=N, mapping relations form, wherein a m when obtaining M=2 ₍₁₎Represent the training sequence that the 1st data Layer distributes, m ₍₂₎Represent the training sequence that the 2nd data Layer distributes, the training sequence deviation value that each Qmin grouping is used must not surpass the peak excursion value that Kmin limited of its grouping, and the training sequence deviation value of distributing between each Qmin grouping does not repeat, when the desired actual number of data layers M that sends greater than 2 the time, execution in step A334;

A334. on the basis that steps A 333 is distributed, according to the desired actual number of data layers M that sends, be the 3rd, 4 ... the M layer distributes training sequence, promptly for some number of channelization codes N, select the training sequence deviation value that deviant is positioned at after the training sequence deviation value two-layer before distributing to is distributed to the 3rd, 4 ... the M layer.

When needs guarantee to satisfy W 〉=Wmin, according to Obtain required maximum training sequence deviation value number K max, allocation result to steps A 333 or A334 is revised, keep and satisfy the training sequence deviation value of formula K≤Kmax and the mapping relations between number of channelization codes N, give up the training sequence deviation value that do not satisfy formula K≤Kmax and the mapping relations between number of channelization codes N, and each number of channelization codes N that will give up part sets up mapping relations with the training sequence deviation value that satisfies formula K≤Kmax, be about to one and take the less training sequence method of salary distribution of training sequence deviation value and a plurality of number of channelization codes N sets up mapping relations, P is the length of training sequence in the formula, and W is that the window of channel estimating is long.

Described step C obtains layer sequence number, scrambler and channel code according to training sequence deviation value and mapping relations, comprising:

C1. according to training sequence deviation value and the first predetermined mapping relations that layer sequence number satisfies, obtain the sequence number of the affiliated layer of this code channel;

C2. when different layers distributes different scramblers, the layer sequence number information under the code channel that obtains according to C1, and scrambler sequence number and the second predetermined mapping relations that layer sequence number satisfies are read scrambler according to the scrambler sequence number, the acquisition scrambler that code channel adopted from the scrambler group;

C3. according to the 3rd satisfied preset mapping relation of the quantity information of training sequence deviation value of distributing on the different layers and channel code, obtain the current number of channelization codes N that is using, and from the channelizing code character, read out N channel code according to pre-determined channelization code allocation rule, obtain the channel code that code channel adopted.

Method of the present invention, at first adopt the mapping form that a kind of exercise sequence distributing method when usage space multiplexing (MIMO) technology is carried out high speed data transfer is provided, enable to be suitable for the situation of multiple spread spectrum coefficient and multiple transmission number of data layers, to reduce the memory cell of receiver; And further use mapping function replaces the mapping form to reduce memory cell, and (mapping form and mapping function all are meant the mapping between distribution of the training sequence on all layers and number of channelization codes, just the former stores with form, and the latter stores with functional form).The inventive method has also provided the revision method simultaneously, makes it to guarantee the estimated performance of training sequence information under the situation of sacrificing some code channel information.

Utilize training sequence to transmit the method for code channel information under the TDD mode of the present invention in the multiaerial system, comprise the scheme one of setting up and using the training sequence mapping form that to support multiple spread spectrum coefficient, the scheme two of setting up and using the mapping form that to support multiple spread spectrum coefficient and multiple transmission number of data layers, with the scheme three of setting up and using the mapping function of simplifying that can support multiple spread spectrum coefficient and multiple transmission number of data layers, and the revision scheme of scheme two mapping forms and scheme three mapping functions.

The inventive method can realize supporting multiple spread spectrum coefficient and multiple transmission number of data layers simultaneously with a kind of training sequence method of salary distribution by such scheme, with existing, send number of data layers method corresponding, that utilize training sequence to transmit code channel information with every kind of spread spectrum coefficient and every kind and compare, can effectively save the memory cell of receiver.

Description of drawings

Embodiment

The inventive method has embodied in the multiaerial system under the TDD mode, can utilize training sequence to transmit code channel information.In order accurately to transmit code channel information, when transmitter terminal inserts training sequence formation burst structure, need follow following basic skills, and guarantee that following preset mapping relation is known at receiver end.Comprise:

(1) the different data Layers that send distribute the training sequence of different deviants, and the training sequence deviation value satisfies the first predetermined mapping relations with layer sequence number;

(2) distribute different scrambling codes at different layers, require scrambler in a scrambler group of determining, to select, and this scrambler group is known at receiver end, the sequence number of scrambler in this scrambler group satisfies the second predetermined mapping relations (if different layers distributes identical scrambler with layer sequence number so, have only a scrambler in the scrambler group this moment, then need not this second mapping relations, in other words these second mapping relations are exactly the fixedly mapping relations between layer sequence number and this scrambler);

(3) need in a channelizing code character of determining, also distribute by the selective channel sign indicating number by rule, known this channelizing code character of receiver end and distribution rule, the training sequence deviation value of distributing on the different layers so and the quantity of channel code, according to the more little principle of the more little employed training sequence deviation value of number of channelization codes, set up the 3rd predetermined mapped relation that is satisfied between the training sequence deviation value of distributing on the different layers and channel code quantity.

For receiver end, obtain just can obtain code channel information by the following method after the training sequence deviation value on each layer by channel estimation methods, comprising:

(1), obtains the serial number information of the affiliated layer of this code channel according to the first predetermined mapping relations that satisfy between training sequence deviation value and layer sequence number;

(2) when different layers distributes different scramblers, according to the layer sequence number information under the code channel of (1) acquisition, and the second predetermined mapping relations that satisfy between scrambler sequence number and layer sequence number, from the scrambler group, read scrambler according to the scrambler sequence number, thereby obtain the scrambling code information that code channel adopted (when different layers distributed identical scrambler, the scrambler that code channel adopted only needed directly to read) from the scrambler group;

(3) the 3rd predetermined mapping relations that satisfied between quantity information according to training sequence deviation value of distributing on the different layers and channel code, obtain the current number of channelization codes N that is using, and from the channelizing code character, read out N channel code, thereby obtain the channelization code information that code channel adopted according to pre-determined channelization code allocation rule.

The scheme one that the inventive method embodiment provides is unified in the mapping relations between training sequence under the different spread spectrum coefficient situations and layer sequence number, scrambling code information, channelization code information in the form, thus the memory cell of saving receiver; The scheme two that the inventive method embodiment provides is on the basis of scheme one, the relation that again difference is sent between training sequence under the number of plies situation and layer sequence number, scrambling code information, channelization code information is unified in the mapping form, thus the memory cell of further saving receiver; The scheme three that the inventive method embodiment provides is transmission methods of a kind of code channel information based on mapping function, it is the further simplification of scheme two, the mapping form of a large amount of memory cell of needs is reduced to a mapping function, thus the memory cell of further saving receiver end.

The inventive method, set up a training sequence mapping form according to predetermined principle, needing adaptively changing for a spread spectrum coefficient is the different system of spread spectrum coefficient Q, need adaptively changing promptly to send the different system of number of data layers M for a transmitting antenna configuration, transmitting terminal relies on such mapping form, utilizes training sequence to transmit code channel information; Receiver relies on such mapping form of being preserved, just can from the training sequence that transmits, obtain code channel information, save the memory space of receiving terminal effectively, avoided receiving terminal to take the generation of the problem of too much memory space because of the different spread spectrum coefficient Q of corresponding stored, the different training sequence mapping forms that send under the number of data layers M.

Embodiment scheme one proposes a kind of method of setting up training sequence mapping form, in the pairing mapping form of different spread spectrum coefficients unification to a form.

To send number of data layers be M, when spread spectrum coefficient is Q, desired training sequence deviation value number K must satisfy $C_K^M\≥Q,$ To guarantee the having training sequence of abundant different deviants to be assigned on the different layers.

In the standard document of 3GPP, for TDD mode, regulation K value can only be selected from following several numerical value: 1,2,4,6,8,10,12,14,16; Spread spectrum coefficient Q can only select from following several numerical value: 1,2,4,8,16.

So, when sending number of plies M=2, during Q=1, require K 〉=2;

Work as M=2, Q=2 or 4 o'clock require K 〉=4;

Work as M=2, during Q=8, require K 〉=6;

Work as M=2, during Q=16, require K 〉=8.

And different spread spectrum coefficients, its corresponding maximum channel number of codes is always identical with spread spectrum coefficient.Therefore, training sequence deviation value in the time of can at first setting up less spread spectrum coefficient and the mapping relations between the number of channelization codes N, training sequence deviation value when setting up big spread spectrum coefficient then and the mapping relations between the number of channelization codes, and when setting up the latter's mapping relations, keep the former mapping relations constant, so just can be in the pairing mapping form of different spread spectrum coefficients unification to a form.This is according to the more little principle of the more little employed training sequence deviation value of number of channelization codes, and the concrete steps of setting up mapping form (this mapping form has comprised described whole three kinds of mapping relations) are as follows:

Step (1) is divided into groups number of channelization codes N according to the minimum spread spectrum coefficient Qmin that it needs;

Step (2) is according to formula $C_K^M\≥Q\min$ Calculate the minimum number Kmin of every group of desired training sequence deviation value, number of channelization codes is once divided into groups again according to Kmin;

Step (3), according to the basic skills of inserting training sequence, in the scope of each Kmin, Qmin grouping, carry out the distribution of training sequence according to Kmin order from small to large, the training sequence deviation value that each grouping is used must not surpass the peak excursion value that Kmin limited, and guarantees not have between the different grouping distribution of repetition.Training sequence deviation value that comprises that training sequence deviation value size and mapping relations between layer sequence number size do not repeat, distributes on the different layers and the mapping relations between number of channelization codes do not repeat.

How to set up training sequence mapping relations form below in conjunction with the example explanation according to above-mentioned steps (1) to (3).With the actual transmission data Layer number M=2 that requires is example, to the number of channelization codes grouping, as shown in table 2 according to step (1) and (2):

Table 2

The grouping first time of step (1) is that number of channelization codes N1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16 is divided into groups according to its minimum spread spectrum coefficient Qmin 1,2,4,8,16 that needs, and is (1), (2), (3,4), (5,6,7,8), (9,10,11,12,13,14,15,16); Step (2) is according to formula $C_K^M\≥Q\min ,$ Calculate the quantity Kmin 2,4,6,8 of every group of desired minimum training sequence deviation value, number of channelization codes N is once divided into groups on step (1) packet by packet basis according to Kmin 2,4,6,8 again, be (1), (2,3,4), (5,6,7,8), (9,10,11,12,13,14,15,16).

According to above-mentioned grouping, and step (3), in the scope of each group, carry out the distribution of training sequence according to Kmin 2,4,6,8 order from small to large, the training sequence deviation value that each group is used must not surpass corresponding Kmin 2,4,6,8 determined peak excursion values, during such as Kmin=2, training sequence can not select deviant to surpass training sequence m ₂The training sequence of deviant, and not have the distribution of repetition between guaranteeing not on the same group, can obtain training sequence mapping relations form as following table 3:

Table 3

In the above table, m ₁To m ₈The numeral 1 and the 2 presentation layer sequence numbers of each row below, m _iExpression deviant different training sequence because the maximum occurrences of K is 16, the training sequence number that so Dui Ying deviant is different also be 16 (i=1 ..., 16).If certain offset m _iInsert certain concrete natural number n in the cell of below, represent that then this training sequence mi is assigned to the layer that sequence number is n; Insert 0, represent that this training sequence is not assigned with (this description is applicable to following table 5,6,7,8).

As can be seen in the such scheme, the selection of training sequence is to select sequence number less (being that deviant is less) training sequence earlier, select the big training sequence of sequence number then, but the generation of training sequence produces a basic training sequences cocycle skew, therefore can be defined as m1 to the training sequence of any one deviant, the sequence number of other training sequences increases successively then, is expressed as m2, m3 up to m16.Especially, we can come the reversed order of the training sequence of m1, m2 to m16 in the above table fully, be expressed as m16, m15 up to m1, and other parts in the form are constant.Be that training sequence divides timing, at first select the bigger training sequence of training sequence sequence number, select the less training sequence of training sequence sequence number then.This moment the training sequence maximum corresponding to ground floor, and the training sequence minimum corresponding to the M layer.

So, obtaining training sequence on each layer by channel estimating after, receiver just can obtain code channel information.With the foregoing description form 3 is example, when the training sequence that obtains is m ₃And m ₇The time, m ₃Pairing layer is a ground floor, m ₇Pairing layer is the second layer; If the scrambler difference of each layer, the scrambler that distributes on the ground floor is first scrambler in the scrambler group so, and the scrambler that distributes on the second layer is second scrambler in the scrambler group; Number of channelization codes in the current use is 11.See above table 3 the 12nd row.

Equally, the training sequence mapping relations table 4 that can set up when sending data Layer number M=4 is as follows:

Table 4

The grouping first time of step (1) is that number of channelization codes N1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16 is divided into groups according to its minimum spread spectrum coefficient Qmin 1,2,4,8,16 that needs, and is (1), (2), (3,4), (5,6,7,8), (9,10,11,12,13,14,15,16); Step (2) is according to formula $C_K^M\≥Q\min ,$ Calculate the quantity Kmin 4,6,8 of every group of desired minimum training sequence deviation value, number of channelization codes N is once divided into groups on step (1) packet by packet basis according to Kmin 4,6,8 again, be (1), (2,3,4,5,6,7,8), (9,10,11,12,13,14,15,16).

According to above-mentioned grouping, and step (3), in the scope of each group, carry out the distribution of training sequence according to Kmin 4,6,8 order from small to large, the training sequence deviation value that each group is used must not surpass Kmin 4,6,8 determined peak excursion values, during such as Kmin=4, training sequence can not select deviant to surpass the training sequence of the deviant of training sequence m4, and not have the distribution of repetition between guaranteeing not on the same group.If this moment, receiver obtained training sequence on each layer by channel estimating, just can obtain code channel information according to this table 4.As the training sequence when acquisition is m ₁, m ₃, m ₇, m ₈The time, m ₁Pairing layer is a ground floor, m ₃Pairing layer is the second layer, m ₇Pairing layer is the 3rd layer, m ₈Pairing layer is the 4th layer, if the scrambler difference of each layer, the scrambler that distributes on the ground floor is first scrambler in the scrambler group so, the scrambler that distributes on the second layer is second scrambler in the scrambler group, the scrambler of the 3rd layer of last distribution is the 3rd scrambler in the scrambler group, the scrambler of the 4th layer of last distribution is the 4th scrambler in the scrambler group, and the number of channelization codes in the current use is 14 (the 15th row in the table).

Such scheme one can be unified the situation of a plurality of spread spectrum coefficients.It is 14 that the maximum that the training sequence mapping form that uses such method to set up is supported sends number of data layers.But need to send data Layer number M, set up such training sequence mapping table, still need take certain receiver memory space at each.

For a transmitter with many antennas, tend to according to the difference of channel circumstance or the difference of business demand, adjust the number of plies that it sends data; In addition, for the mobile communication system of honeycomb, the number of antennas of the transmitter of its each cellular configuration is also different probably, thereby its data Layer number that can send also can be different.Like this for same receiver, the number of plies of the data of facing when it that transmitter sends just needs to preserve the form that the different mappings corresponding with the data Layer number concerns, thereby causes the waste of resource not simultaneously.The present invention program two proposes on the basis of scheme one at this problem, with same mapping form when distributing different spread spectrum coefficients, different pieces of information number of layers training sequence and utilize this training sequence to send code channel information.

When transmission number of data layers M is 2 (ground floor and the second layers), utilize scheme once the mapping relations that can set up fully between training sequence deviation value and the different channels number of codes N, support more multiple when sending number of plies M, only need to get final product according to the layer that the basic skills of inserting training sequence is selected the training sequence deviation value to distribute to arbitrarily to increase newly, this moment, receiver end only need judge that number of channelization codes gets final product according to the training sequence deviation value of the ground floor and the second layer, thereby many forms that difference sends number of plies M correspondence are unified into a form.The concrete steps of setting up mapping form (this mapping form has comprised described whole three kinds of mapping relations) are as follows:

Step (1), with number of channelization codes N according to its minimum spread spectrum coefficient Qmin that needs divide into groups (with aforementioned form 2);

Step (2) is for sending number of plies M=2, according to formula $C_K^M\≥Q\min$ Calculate the minimum number Kmin of every group of desired training sequence deviation value, according to Kmin number of channelization codes N is once divided into groups again (with aforementioned form 2);

Step (3), according to the basic skills of inserting training sequence, in the scope of each Kmin group, carry out the distribution of training sequence according to Kmin order from small to large, promptly divide timing at first to travel through all situations of minimum Kmin value, and then travel through all situations of each Kmin value in proper order by the increase relation of Kmin, finish a mapping relations form when obtaining M=2 up to setting up with the mapping relations of all number of channelization codes N;

Step (4), on the basis of the mapping relations form that step (3) is obtained, send number of plies M according to reality, according to the basic skills of inserting training sequence, for sending the number of plies is 3,4 ... the layer of M distributes the training sequence deviation value, promptly for some number of channelization codes N, selection distributes to the 3rd, 4 with the training sequence that deviant is positioned at after the training sequence deviation value two-layer before distributing to ... the M layer, and guarantee that the magnitude relationship of all layers sequence number is consistent with the magnitude relationship of the training sequence deviation value of its distribution.

Following table 5 is embodiment of the mapping relations form of an operational version two foundation, as space is limited, has omitted corresponding Kmin and Qmin grouping in the form.

Table 5

m 1	m 2	m 3	m 4	m 5	m 6	m 7	m 8	m 9	m 10	m 11	m 12	m 13	m 14	m 15	m 16	N
																	1	2	X	X	X	X	X	X	X	X	X	X	X	X	X	X	1
1	0	2	X	X	X	X	X	X	X	X	X	X	X	X	X	2
																	0	1	2	X	X	X	X	X	X	X	X	X	X	X	X	X	3
1	0	0	2	X	X	X	X	X	X	X	X	X	X	X	X	4
																	0	1	0	2	X	X	X	X	X	X	X	X	X	X	X	X	5
0	0	1	2	X	X	X	X	X	X	X	X	X	X	X	X	6
																	1	0	0	0	2	X	X	X	X	X	X	X	X	X	X	X	7
0	1	0	0	2	X	X	X	X	X	X	X	X	X	X	X	8
																	0	0	1	0	2	X	X	X	X	X	X	X	X	X	X	X	9
0	0	0	1	2	X	X	X	X	X	X	X	X	X	X	X	10
																	1	0	0	0	0	2	X	X	X	X	X	X	X	X	X	X	11
0	1	0	0	0	2	X	X	X	X	X	X	X	X	X	X	12
																	0	0	1	0	0	2	X	X	X	X	X	X	X	X	X	X	13
0	0	0	1	0	2	X	X	X	X	X	X	X	X	X	X	14
																	0	0	0	0	1	2	X	X	X	X	X	X	X	X	X	X	15
1	0	0	0	0	0	2	X	X	X	X	X	X	X	X	X	16

In the table, " X " represents layer sequence number or the expression numeral 0 more than or equal to 3.Concrete tag value is layer sequence number or the mark 0 more than or equal to 3 (M), is determined by the method in scheme two steps (4).But the magnitude relationship of layer sequence number must be consistent with the magnitude relationship of the training sequence deviation value of its distribution, as under the situation of M=4, with m ₁Distribute to ground floor, with m ₆When distributing to the second layer, can not be with m ₇Distribute to the 4th layer, with m ₈Distribute to the 3rd layer, must be with m ₇Distribute to the 3rd layer, with m ₈Distribute to the 4th layer.

At this moment, obtaining training sequence on each layer by channel estimating after, receiver just can obtain code channel information.With the foregoing description form 5 is example, when send the number of plies be 2 and the training sequence that obtains be m ₃And m ₆The time, m ₃Pairing layer is a ground floor, m ₆Pairing layer is the second layer; If the scrambler difference of each layer, the scrambler that distributes on the ground floor is first scrambler in the scrambler group so, and the scrambler that distributes on the second layer is second scrambler in the scrambler group; Scrambler number N in the current use is 13 (seeing above table the 14th row).When send the number of plies be 4 and the training sequence that obtains be m ₄, m ₆, m ₇And m ₈The time, m ₄Pairing layer is a ground floor, m ₆Pairing layer is the second layer, m ₇Pairing layer is the 3rd layer, m ₈Pairing layer is the 4th layer; If the scrambler difference of each layer, the scrambler that distributes on the ground floor is first scrambler in the scrambler group so, and the scrambler that distributes on the second layer is second scrambler in the scrambler group; The scrambler of the 3rd layer of last distribution is the 3rd scrambler in the scrambler group; The scrambler of the 4th layer of last distribution is the 4th scrambler in the scrambler group; Number of channelization codes in the current use is 14 (seeing above table the 15th row).

Operational version two, though can will unite with different transmission number of plies corresponding mapping table lattice, in some cases, the minimum K value of its requirement becomes big, such as when the Q=16, for M=4, according to formula $C_K^M\≥Q,$ Its desired minimum K is 8, and in the foregoing description form 5, its desired K value minimum is 10; And the maximum number of plies of supporting descends to some extent, and the maximum number of plies of operational version one its support is 14, and the maximum number of plies of operational version two its supports is 11.Therefore scheme two is with respect to scheme 1, and its K value wants changes persuing big, channel estimating performance meeting variation, and the transmission number of plies of support diminishes, and the scope of application diminishes.

Can find that the step (3) in the scheme two is also inconsistent with the step (3) in the scheme one, step in the scheme two (3) at first travels through all situations of less K value, traversal is than all situations of big K value then, be for guarantee at first to set up corresponding in the mapping relations under the M=2 situation, the training sequence deviation value of distributing is less, thereby guarantee that other layers mix bigger right to choose at the branch of training sequence deviation value, and guarantee to support bigger transmission number of data layers.Contrast form 3 and form 5 are not difficult to find, uses the mapping relations of the corresponding ground floor of two kinds of methods foundation and the second layer different, and in last column of form 3, the training sequence of the second layer have been assigned to m ₈, remaining 8 training sequence deviation values in addition can distribute, and adds and distributes the preceding two-layer of training sequence, and the maximum number of plies of support only is 10; But in form 5 last columns, the training sequence of the second layer only is assigned to m ₇, remaining 9 training sequence deviation values in addition can distribute, and adds and distributes the preceding two-layer of training sequence, and its maximum number of layers that can support is 11.

In order further to save the storage resources of receiver, can on the basis of scheme one and scheme two,, the form mapping relations be come out with certain simple Function Mapping relationship expression by some modification, be called scheme three.The concrete steps of setting up mapping relations (these mapping relations have comprised described whole three mapping relations) are as follows:

Step (1), with number of channelization codes N according to its minimum spread spectrum coefficient Qmin that needs divide into groups (with aforementioned form 2);

Step (2) is for sending number of plies M=2, according to formula $C_K^M\≥Q\min$ Calculate the minimum number Kmin of every group of desired training sequence deviation value, according to Kmin to number of channelization codes N once divide into groups again (with aforementioned form 2);

Step (3) is set up the training sequence m of different deviants _i(i=1 ... 16) and certain mapping relationship f (m between some numeral _i);

Step (4) according to the basic skills of inserting training sequence, is set up the mapping relations between different training sequence deviation values and the different channels number of codes N, and guarantees f (m ₍₁₎)+f (m ₍₂₎)=N, mapping relations form, wherein a m when obtaining M=2 _(m)Represent the training sequence that the m layer distributes, m=1,2, the training sequence deviation value that each Qmin grouping is used must not surpass the peak excursion value that Kmin limited of its grouping, and the training sequence deviation value of distributing between each Qmin grouping does not repeat;

Step (5), on the basis of the mapping relations form that step (4) is obtained, send number of plies M according to reality, according to the basic skills of inserting training sequence, for sending the number of plies is 3,4 ... the layer of M distributes the training sequence deviation value, promptly for some number of channelization codes N, selection distributes to the 3rd, 4 with the training sequence that deviant is positioned at after the training sequence deviation value two-layer before distributing to ... the M layer, and guarantee that the magnitude relationship of all layers sequence number is consistent with the magnitude relationship of the training sequence deviation value of its distribution.

So just set up one with different spread spectrum coefficient Q and different send that training sequence that number of plies M unites distributes and number of channelization codes N between the mapping form.Use such mapping form, just need not to preserve this form again at receiver end, and only need know mapping relationship f (m _i) and mapping function f (m ₍₁₎)+f (m ₍₂₎)=N gets final product.

Provide the embodiment of the mapping relations form of operational version three foundation below.

Mapping relations between training sequence and the optional network specific digit are (being documented in the row of the 2nd in the table 6):

$\left\{\begin{array}{cc}f\left(m_i\right)=i-1& i=1\·\·\·5\\ f\left(m_i\right)=f\left(m_5\right)+2\×(i-5)& i=\mathrm{6,7}\\ f\left(m_i\right)=12& i=8\\ f\left(m_i\right)=0& i=9\·\·\·16\end{array}\right.$

The mapping form is following table 6 (as space is limited, having omitted corresponding Kmin and Qmin grouping in the form):

Table 6

Receiver is according to the training sequence m of ground floor and second layer correspondence ₍₁₎And m ₍₂₎, can calculate corresponding number of channelization codes N by mapping relations.With the foregoing description form is example, when send the number of plies be 2 and the training sequence that obtains be m ₃And m ₆The time, m ₃Pairing layer is a ground floor, m ₆Pairing layer is the second layer; If the scrambler difference of each layer, the scrambler that distributes on the ground floor is first scrambler in the scrambler group so, and the scrambler that distributes on the second layer is second scrambler in the scrambler group; Use mapping function f (m _i), f (m ₃) corresponding numeral is 2, f (m ₆) corresponding numeral is 6, the number of channelization codes in the so current use only needs two digital additions, i.e. 2+6=8.When send the number of plies be 2 and the training sequence that obtains be m ₂And m ₇The time, m ₂Pairing layer is a ground floor, m ₇Pairing layer is the second layer; If the scrambler difference of each layer, the scrambler that distributes on the ground floor is first scrambler in the scrambler group so, and the scrambler that distributes on the second layer is second scrambler in the scrambler group; Use mapping function f (m _i), f (m ₃) corresponding numeral is 1, f (m ₇) corresponding numeral is 8, the number of channelization codes in the so current use only needs two digital additions, i.e. 1+8=9.

Operational version three further makes the minimum K value that needs become greatly, and the maximum number of layers of its support also reduces to some extent, is example with the foregoing description, and in form 6 last columns, the training sequence of the second layer is assigned to m ₈, remaining 8 training sequence deviation values in addition can distribute, and adds and distributes the preceding two-layer of training sequence, and its maximum data number of layers that can support is 10.

In the mimo system under the TDD mode, the basic demand of inserting training sequence is K 〉=M (guaranteeing every layer training sequence deviation value difference), and in three kinds of embodiments that the inventive method proposes, all improved this requirement and caused K to become big, made the long W of window of channel estimating shorten according to following relational expression:

P is the length of training sequence in the formula, is a fixed numbers, oeprator

The maximum integer smaller or equal to x is got in expression.

Raising is to the requirement of K value, this is because in the training sequence that distributes, because the information of having carried channel code, training sequence has become well-regulated distribution from random distribution, so sacrificed some training sequences, thereby make the long W of window of channel estimating shorten, influenced performance for estimating channel.

In order to save training sequence, increase the long W of window of channel estimating, can use to take between the less training sequence deviation value of training sequence deviation value and a plurality of number of channelization codes N as far as possible and set up mapping relations.Use a kind of like this revision method, can set up a kind of mapping relations of one-to-many, yet cause receiver can't be unique to determine its number of channelization codes, so receiver need use extra method to carry out the judgement of channel code, such as activating detection technique.

Particularly, make and satisfy W 〉=Wmin, according to formula Obtain required maximum training sequence deviation value number K max, step (3) or (4) to scheme two, the allocation result of the step of scheme three (4) or (5) is revised, keep and satisfy the training sequence deviation value of formula K≤Kmax and the mapping relations between number of channelization codes N, give up the training sequence deviation value that do not satisfy formula K≤Kmax and the mapping relations between number of channelization codes N, and each number of channelization codes N that will give up part sets up mapping relations with the training sequence deviation value that satisfies formula K≤Kmax, is about to one and takies the less training sequence method of salary distribution of training sequence deviation value and a plurality of number of channelization codes N sets up mapping relations.

Providing specific embodiment below describes.

For mimo system, common configuration is to support four layer data to send, and generally requires the long W=16 of estimating window; For the TD-SCDMA mobile communication system, therefore the length P=128 of training sequence expects that its K value that can support is reduced to 8, thereby can effectively increase the long W of window of channel estimating.As target scheme two, three is revised below.

For the embodiment shown in the form 5, the mapping form that it is to use scheme two to set up.When supporting that the M=4 layer sends data, the minimum K value of its requirement is 10, at this moment the long W=12 of window of its channel estimating.Use the revision method, keep the training sequence deviation value of K≤8 and the mapping relations between the channel code N, the pairing N of the mapping relations that are rejected is set up mapping relations with the training sequence deviation value of K≤8 again.As shown in table 7 below.Can find only last column mapping relations in the form 5 need be given up at this moment to get final product, and the one-to-many mapping of the first pairing method of salary distribution foundation of row and N=1 and N=16 is got final product.Only just there is the mapping of one-to-many in a kind of revision so when Q=16.

The mapping form 7 that following embodiment provides, the minimum K value that requires when M=2 is reduced to 6, and the minimum K value that requires when M=4 is reduced to 8, and the supported maximum number of layers that sends increases to 12.

Table 7

m 1	m 2	m 3	m 4	m 5	m 6	m 7	m 8	m 9	m 10	m 11	m 12	m 13	m 14	m 15	m 16	N
																	1	2	X	X	X	X	X	X	X	X	X	X	X	X	X	X	1 or 16
1	0	2	X	X	X	X	X	X	X	X	X	X	X	X	X	2
																	0	1	2	X	X	X	X	X	X	X	X	X	X	X	X	X	3
1	0	0	2	X	X	X	X	X	X	X	X	X	X	X	X	4
																	0	1	0	2	X	X	X	X	X	X	X	X	X	X	X	X	5
0	0	1	2	X	X	X	X	X	X	X	X	X	X	X	X	6
																	1	0	0	0	2	X	X	X	X	X	X	X	X	X	X	X	7
0	1	0	0	2	X	X	X	X	X	X	X	X	X	X	X	8
																	0	0	1	0	2	X	X	X	X	X	X	X	X	X	X	X	9
0	0	0	1	2	X	X	X	X	X	X	X	X	X	X	X	10
																	1	0	0	0	0	2	X	X	X	X	X	X	X	X	X	X	11
0	1	0	0	0	2	X	X	X	X	X	X	X	X	X	X	12
																	0	0	1	0	0	2	X	X	X	X	X	X	X	X	X	X	13
0	0	0	1	0	2	X	X	X	X	X	X	X	X	X	X	14
																	0	0	0	0	1	2	X	X	X	X	X	X	X	X	X	X	15

For the embodiment shown in the form 6, when supporting that M=4 sends data Layer, the minimum K value of its requirement is 10, at this moment the long W=12 of window of its channel estimating.Use the revision method, keep the training sequence deviation value of K≤8 and the mapping relations between the channel code N, the pairing N of the mapping relations that are rejected again with the training sequence deviation value opening relationships of K≤8.Because the mapping relations that scheme three is to use mapping function to set up, same number of channelization codes N, exist the multiple training sequence deviation value situation corresponding with N, can utilize this redundancy during revision, change the many-one mapping relations between training sequence deviation value and the channel code N into mapping one to one, after redundancy is finished using, set up the mapping relations of one-to-many again.

Following table 8 promptly be one to the mapping relations form of form 6 after revising.Can find wherein to exist the one-to-many mapping (the third line) between training sequence deviation value and the number of channelization codes N, though and remaining mapping relations in form 8, show as mapping one by one, the functional relation f (m that operational version three provides ₍₁₎)+f (m ₍₂₎When)=N carries out computing, there is the one-to-many mapping, though promptly by functional relation f (m ₍₁₎)+f (m ₍₂₎)=N can calculate some number of channelization codes N, but its number of channelization codes that may represent is N or is another one value N '.Such as, though f (m in the form 8 ₂)+f (m ₃)=f (m ₁)+f (m ₄)=3, but in form 8 with m ₂And m ₃When two-layer, its corresponding number of channelization codes N is 3 before distributing to, and m ₁And m ₄When two-layer, its corresponding number of channelization codes N is 11 before distributing to.Only just there is the mapping of one-to-many in a kind of revision so when Q=16.

The minimum K value that the mapping form 8 that embodiment provides requires when M=2 is reduced to 6, and the minimum K value that requires when M=4 is reduced to 8, and the supported maximum number of layers that sends all increases to 12.

Table 8

The table of comparisons 6 and table 8 are as can be seen, use the revision method, the training sequence deviation value of the row of the 1st, 2,3,5,7,9,11,14,15,18 in the reservation table 6 and the mapping relations between the channel code N (1,2,3,4,5,6,7,8,9,10), leave out the 19th to the 26th row, rebulid the training sequence deviation value of the 1st, 4,6,8,10,12 row and the mapping relations between the channel code N (16,11,12,13,14,15), and set up table 8.

The training sequence that utilizes that uses the present invention to propose transmits the method for code channel information, can support the multi-antenna technology under the TDD mode.Operational version one can unify to set up the mapping relations form under the different spread spectrum coefficient situations; Operational version two can unify to set up different spread spectrum coefficients and the different mapping relations forms that send under the number of layers situations; The mapping relations tableau reduction that operational version three can be set up scheme two is certain functional relation, thereby saves the storage resources of receiver.The inventive method has also provided when requiring the long W of channel estimation window higher (it is long to sacrifice channel estimation window) revision method by the training sequence allocative decision of sacrificing some code channel information and realizing simultaneously.

Claims (12)

1. a time-division duplex multi-antenna system utilizes training sequence to transmit the method for code channel information, transmitting terminal resolves into layer data with the data flow of each time slot, each layer data is decomposed into a plurality of substream of data according to number of channelization codes, each substream of data in each layer data is added training sequence form burst structure, in the training sequence that adds, carry code channel information; Receiving terminal carries out channel estimating to the data that receive, and obtains the training sequence of each layer, obtains code channel information according to training sequence, it is characterized in that:

Described transmitting terminal adds training sequence and forms burst structure, carries code channel information in the training sequence that adds, and comprising:

A. in the training sequence deviation value, set up mapping relations between layer sequence number and number of channelization codes, these mapping relations comprise the mapping relations between training sequence deviation value size and layer sequence number size, mapping relations in layer sequence number size and the scrambler group between scrambler sequence number size, and training sequence deviation value of distributing on the different layers and the mapping relations between number of channelization codes, when setting up mapping relations, described mapping relations when at first setting up a certain spread spectrum coefficient, when spread spectrum coefficient increases, the new mapping relations that keep the constant and correspondence expansion spread spectrum coefficient that increases when the expansion spread spectrum coefficient increases on this mapping relations basis of aforementioned mapping relations, when the new mapping relations of the described corresponding expansion spread spectrum coefficient of expansion, guarantee that the employed training sequence deviation value of each mapping relations is no more than the deviant that quantity limited of its minimum training sequence deviation value, and guarantee that the training sequence deviation value between each mapping relations does not repeat;

B. in each substream of data of each layer data, add a training sequence that has deviant with respect to basic training sequences by above-mentioned mapping relations;

Described receiving terminal obtains code channel information according to training sequence, comprising:

C. obtain the training sequence deviation value of each layer by the training sequence of each layer, the code channel information of layer sequence number, scrambler and the channel code that comprises under the code channel layer according to the training sequence deviation value of distributing on mapping relations between scrambler sequence number size in the mapping relations between the training sequence deviation value identical size and layer sequence number size, layer sequence number size and the scrambler group and the different layers and the acquisition of the mapping relations between number of channelization codes with transmitting terminal.

2. method according to claim 1, it is characterized in that: described steps A, be to set up mapping relations between training sequence deviation value, layer sequence number, scrambler sequence number and number of channelization codes, comprise according to the more little principle of the more little employed training sequence deviation value of number of channelization codes:

The principle of A1. distributing different deviant training sequences by different layers is set up the first predetermined mapping relations between training sequence deviation value and the layer sequence number;

A2. by different layers distribute different scrambling codes, requirement scrambler in a scrambler group of determining, select and this scrambler group when receiver end is known, set up second mapping relations of being scheduled between the sequence number of scrambler in the scrambler group and the layer sequence number;

A3. require in the channelizing code character that determines the selective channel sign indicating number and distribute and this channelizing code character and distribute rule when receiver end is known, set up the 3rd preset mapping relation between the quantity information of the training sequence deviation value of distributing on the different layers and channel code according to the more little principle of the more little employed training sequence deviation value of number of channelization codes by rule.

3. method according to claim 2 is characterized in that: in the described steps A 2, also be included in different layers when distributing identical scrambler, have only a scrambler in the scrambler group, these second predetermined mapping relations are exactly the fixedly mapping relations between layer sequence number and this scrambler.

4. method according to claim 2, it is characterized in that: in the described steps A 3, set up the 3rd preset mapping relation between the quantity information of the training sequence deviation value of distributing on the different layers and channel code according to the more little principle of the more little employed training sequence deviation value of number of channelization codes, comprising:

A311. all number of channelization codes are divided into groups according to the needed minimum spread spectrum coefficient Qmin of each number of channelization codes N;

A312. according to formula $C_K^M\≥Q\min$ M is the desired actual number of data layers that sends, calculate the divide into groups minimum number Kmin of desired training sequence deviation value of every minimum spread spectrum coefficient, according to each Kmin, on steps A 311 packet by packet basis each number of channelization codes N is once divided into groups again, K is the quantity of training sequence deviation value;

A313. according to the minimum number Kmin order from small to large of training sequence deviation value, distribute training sequence by the desired actual number of data layers that sends in each Qmin grouping in each Kmin grouping, the deviant of the training sequence that each Qmin grouping of branch timing is used must not surpass the peak excursion value that Kmin limited of its grouping, and the training sequence deviation value of distribution between each Qmin grouping does not repeat.

5. method according to claim 2, it is characterized in that: in the described steps A 3, set up the 3rd preset mapping relation between the quantity information of the training sequence deviation value of distributing on the different layers and channel code according to the more little principle of the more little employed training sequence deviation value of number of channelization codes, comprising:

A321. all number of channelization codes are divided into groups according to the needed minimum spread spectrum coefficient Qmin of each number of channelization codes N;

A322. to send number of data layers M=2, according to formula $C_K^M\≥Q\min$ , calculate the divide into groups minimum number Kmin of desired training sequence deviation value of every minimum spread spectrum coefficient Qmin, each number of channelization codes N is once divided into groups again on steps A 321 packet by packet basis according to each Kmin, K is the quantity of training sequence deviation value;

A323. according to the minimum number Kmin order from small to large of training sequence deviation value, distribute training sequence by sending number of data layers M=2 in each Qmin grouping in each Kmin grouping, it is less that the branch timing at first travels through required training sequence deviation value minimum number Kmin, the mapping situation of all number of channelization codes N and training sequence deviation value, and then it is bigger to travel through required training sequence deviation value minimum number Kmin, the mapping situation of all number of channelization codes N and training sequence deviation value, when the desired actual number of data layers M that sends greater than 2 the time, further execution in step A324;

A324. on the basis that steps A 323 is distributed, according to the desired actual number of data layers M that sends, be the 3rd, 4 ... the M layer distributes training sequence, promptly for some number of channelization codes N, select the training sequence that deviant is positioned at after the training sequence deviation value two-layer before distributing to is distributed to the 3rd, 4 ... the M layer.

6. method according to claim 5 is characterized in that further comprising: when needs guarantee to satisfy W 〉=Wmin, according to formula Obtain required maximum training sequence deviation value number K max, allocation result to steps A 323 or A324 is revised, keep and satisfy the training sequence deviation value of formula K≤Kmax and the mapping relations between number of channelization codes N, give up the training sequence deviation value that do not satisfy formula K≤Kmax and the mapping relations between number of channelization codes N, and each number of channelization codes N that will give up part sets up mapping relations with the training sequence deviation value that satisfies formula K≤Kmax, be about to one and take the less training sequence method of salary distribution of training sequence deviation value and a plurality of number of channelization codes N sets up mapping relations, P is the length of training sequence in the formula, and W is that the window of channel estimating is long.

7. method according to claim 2, it is characterized in that: in the described steps A 3, set up the 3rd preset mapping relation between the quantity information of the training sequence deviation value of distributing on the different layers and channel code according to the more little principle of the more little employed training sequence deviation value of number of channelization codes, comprising:

A331. all number of channelization codes are divided into groups according to the needed minimum spread spectrum coefficient Qmin of each number of channelization codes N;

A332. to send number of data layers M=2, according to formula $C_K^M\≥Q\min$ , calculate the divide into groups minimum number Kmin of desired training sequence deviation value of every minimum spread spectrum coefficient Qmin, each number of channelization codes N is once divided into groups again on steps A 331 packet by packet basis according to each Kmin, K is the quantity of training sequence deviation value;

A333. at the different training sequence m of deviant _i, i=1 ... 16, and set up mapping relationship f (m between some numeral _i), according to the minimum number Kmin order from small to large of training sequence deviation value, distribute training sequence by sending number of data layers M=2 in each Qmin grouping in each Kmin grouping, divide timing to guarantee f (m ₍₁₎)+f (m ₍₂₎)=N, mapping relations form, wherein a m when obtaining M=2 ₍₁₎Represent the training sequence that the 1st data Layer distributes, m ₍₂₎Represent the training sequence that the 2nd data Layer distributes, the training sequence deviation value that each Qmin grouping is used must not surpass the peak excursion value that Kmin limited of its grouping, and the training sequence deviation value of distributing between each Qmin grouping does not repeat, when the desired actual number of data layers M that sends greater than 2 the time, execution in step A334;

A334. on the basis that steps A 333 is distributed, according to the desired actual number of data layers M that sends, be the 3rd, 4 ... the M layer distributes training sequence, promptly for some number of channelization codes N, select the training sequence deviation value that deviant is positioned at after the training sequence deviation value two-layer before distributing to is distributed to the 3rd, 4 ... the M layer.

8. method according to claim 7 is characterized in that: in the described steps A 333, and training sequence m _iAnd the mapping relationship f (m that sets up between some numeral _i) be:

$\left\{\begin{array}{cc}f\left(m_i\right)=i-1& i=1...5\\ f\left(m_i\right)=f\left(m_5\right)+2\×(i-5)& i=\mathrm{6,7}\\ f\left(m_i\right)=12& i=8\\ f\left(m_i\right)=0& i=9...16\end{array}\right..$

9. according to claim 7 or 8 described methods, it is characterized in that further comprising: when needs guarantee to satisfy W 〉=Wmin, according to Obtain required maximum training sequence deviation value number K max, allocation result to steps A 333 or A334 is revised, keep and satisfy the training sequence deviation value of formula K≤Kmax and the mapping relations between number of channelization codes N, give up the training sequence deviation value that do not satisfy formula K≤Kmax and the mapping relations between number of channelization codes N, and each number of channelization codes N that will give up part sets up mapping relations with the training sequence deviation value that satisfies formula K≤Kmax, be about to one and take the less training sequence method of salary distribution of training sequence deviation value and a plurality of number of channelization codes N sets up mapping relations, P is the length of training sequence in the formula, and W is that the window of channel estimating is long.

10. method according to claim 1 is characterized in that: described step C, obtain layer sequence number, scrambler and channel code according to training sequence deviation value and mapping relations, and comprising:

C1. according to training sequence deviation value and the first predetermined mapping relations that layer sequence number satisfies, obtain the sequence number of the affiliated layer of this code channel;

C2. when different layers distributes different scramblers, the layer sequence number information under the code channel that obtains according to C1, and scrambler sequence number and the second predetermined mapping relations that layer sequence number satisfies are read scrambler according to the scrambler sequence number, the acquisition scrambler that code channel adopted from the scrambler group;

C3. according to the 3rd satisfied preset mapping relation of the quantity information of training sequence deviation value of distributing on the different layers and channel code, obtain the current number of channelization codes N that is using, and from the channelizing code character, read out N channel code according to pre-determined channelization code allocation rule, obtain the channel code that code channel adopted.

11. method according to claim 10 is characterized in that described step C2 also comprises: when different layers distributed identical scrambler, the scrambler that code channel adopted directly read from the scrambler group of having only a scrambler.

12. method according to claim 1, it is characterized in that: described steps A, be to set up mapping relations between training sequence deviation value, layer sequence number, scrambler sequence number and number of channelization codes, comprise according to the big more principle of the more little employed training sequence deviation value of number of channelization codes:

The principle of A1. distributing different deviant training sequences by different layers is set up the first predetermined mapping relations between training sequence deviation value and the layer sequence number;

A2. by different layers distribute different scrambling codes, requirement scrambler in a scrambler group of determining, select and this scrambler group when receiver end is known, set up second mapping relations of being scheduled between the sequence number of scrambler in the scrambler group and the layer sequence number;

A3. require in the channelizing code character that determines the selective channel sign indicating number and distribute and this channelizing code character and distribute rule when receiver end is known, set up the 3rd preset mapping relation between the quantity information of the training sequence deviation value of distributing on the different layers and channel code according to the big more principle of the more little employed training sequence deviation value of number of channelization codes by rule.