WO2004028036A1 - The coding method, application method and coding device of extended frequency multiaddress - Google Patents

Abstract

The invention provides a coding method, application method and coding device of extended frequency multiaddress, it relates to the wireless communications fields of extended frequency and code division multiple address (CDMA); repeats the elements on each of the sequence bits in the initial orthogonal sequence group by means of selecting the any orthogonal sequence group as the initial orthogonal sequence group, and then obtains another sequence group, obtains a subcode portion of the DOC code by means of multiplying the each subcode in another subcode portion by each bit of the switching sequence Q one by one, obtains another subcode portion of the DOC code by means of multiplying the each subcode in the subcode portion by each bit of the switching sequence I one by one; obtains the DOC code group by means of combining the subcode portion; since the invention can obtain the enough code groups, it's possible to average the disturbs that go through the different users by means of applying the form of skip code or disturb code; the said coding device of extended frequency multiaddress of the invention can produce the DOC code.

Description

 Spread-spectrum multiple access code encoding method, application method and encoding device

 The present invention relates to the technical field of spread spectrum and code division multiple access (CDMA) wireless communications, and in particular to a spread spectrum multiple access code encoding method, application method, and encoding device.

Background technique

 LS codes are a type of "Complementary Orthogonal Codes" (Complementary Orthogonal Code), published in the PCT, with the application number PCT / CN00 / 00028. The inventors are: Li Daoben, and the invention name is: In the patent application document of the multiple access coding method, a codeword is divided into two parts, C and S, and a tree structure is used to generate a codeword having an "Interference Free Window" (IFW).

 Similar to the Golay Pair, the LS code is generally divided into two sub-code parts: a C-code part and an S-code part. In an LS code group, the autocorrelation function of each codeword is ideal, and the correlation function has a zero correlation window near the origin. In fact, more generally, cross-correlation functions may have non-zero values only at certain locations ("windows"). In the LAS-CDMA initial system solution using the LS code, the C part and the S part of the LS code are transmitted on two time slots with a width of 4 chips (chips), and different arrangements of the LA code are used. ) To distinguish between different cells or sectors.

Due to the small number of available LS code groups, the application of LS codes is limited. For example, even if all the codewords in the code group are reversed, it is regarded as obtaining another code group (without allowing non-180-degree rotation). For a case where the spreading factor is equal to 128 and the width of a single-sided IFW is equal to 3 At most, only a total of 8 LS code groups can be constructed. This is one of the reasons why LA codes (different permutations) are used in LAS-CDMA systems to distinguish between different cells / sectors. On the other hand, the correlation characteristics between different LS codeword pairs are different, which also means that if a fixed LS code is assigned to each user, when a traditional single-user detector is used, each user experiences Interference The specific codeword used is related, that is, the performance among all users is non-average. Because the system design must provide a certain communication quality for each user, the system design must meet the relevant requirements when using codewords with poor related characteristics, which is a waste of resources for those codewords with good related characteristics. .

 If in the CDMA system, each user is assigned a fixed spreading codeword with a length equal to the spreading factor (Spreading Factor), this is the case of "short code CDMA"-a shortcode CDMA system. Many multiuser detection technologies are designed for the "short code CDMA" situation (especially in the uplink). In contrast to "short code CDMA", a "code hopping" method can be used to solve the problem of unbalanced interference between users. At this time, the spreading codeword used by each user is constantly changed in a certain way, so that the interference is equalized among the users (although the total interference value has not changed). If the traditional single-user detection method is adopted, this "code hopping" method is adopted to design a better and more robust system than the aforementioned "short code CDMA" method. Long pseudo-random (PN) is adopted. Sequence scrambling of short spreading codes

(scrambling), can be regarded as a very convenient and efficient code hopping method, and the corresponding system is called

"Longcode CDMA" system.

 In existing commercial CDMA systems, including IS-95, IS-2000, and WCDMA, this long PN sequence scrambling method is used. Therefore, these commercial CDMA systems are "long code CDMA" systems. In these systems, Walsh orthogonal sequences are used as channelization codes, and long PN sequences (such as m-sequences, Gold sequences, etc.) are used for scrambling. In these systems, the function of the scrambling code mainly includes: a) distinguishing different cells / sectors; b) providing a means for data privacy; c) making the relative offset of Walsh sequences in the same cell / sector non-zero The autocorrelation value and cross-correlation value are randomized at the time; note that the non-zero offset correlation value of the Walsh sequence itself may be large; d) randomize the cross-correlation value of Walsh sequences in different cells / sectors.

For LS codes, because the number of code groups is too small, it is difficult to use some form of code hopping or scrambling codes to average the interference experienced by different users. Meanwhile, if the LS code is directly scrambled Code, its original Zero Correlation Window (IFW) and related slatted structures will be destroyed.

Summary of the Invention

 An object of the present invention is to provide a spread spectrum multiple access code encoding method, an application method, and an encoding device. The spread-spectrum multiple access code according to the present invention may be referred to as a DOC (Distinctive Orthogonal Code) code. The correlation function of the DOC code only has large non-zero values at certain specific positions. These specific positions are called window stiles (Window Frame \ DOC codes can focus on the interference introduced by the non-ideal correlation characteristics between code words. These slat positions. If interference paths (including ISI, MAI, and ACI) exist only at those non-slat positions, they can be completely eliminated by the DOC code. If the interference path exists at the slat position, it can also In some simple way (for example, equalization, Pre-RAKE, and interference cancellation), the interference located on the slat position is eliminated or moved to other non-slat positions in advance, so that the purpose of eliminating interference can also be achieved. The DOC code and other related technologies can completely eliminate or partially eliminate the interference in the corresponding two-way synchronous code division multiple access (CDMA) system, making it possible to establish a large-capacity digital mobile communication system. Frequency multiple access codes (DOC codes) have very similar characteristics to the LS codes, such as the relevant IFW properties and the properties of the window grille. LS codes can be regarded as New Class of a spread spectrum code exception, however, in addition to the LS code exceptions, proposed by the present invention a new class of spreading codes is generally not address orthogonal complementary codes.

 The spread spectrum multiple access code encoding method according to the present invention can obtain a sufficient number of code groups. Therefore, in the application of the spread spectrum multiple access code (DOC code), a form of hopping or scrambling can be used to make different users The interference experienced is averaged.

 In the application of the spread spectrum multiple access code (DOC code), the present invention also provides a method for performing DOC code group allocation in a cellular communication system network to facilitate the application of the DOC code.

 The spread spectrum multiple access code encoding device of the present invention can generate the DOC code.

 The technical scheme of the present invention is as follows:

A method for encoding a spread spectrum multiple access code, which is characterized by including the following steps: Select any orthogonal sequence group as the initial orthogonal sequence group;

 Repeating the elements on each sequence bit in the initial orthogonal sequence group sequentially to obtain another sequence group;

 Generate transformation sequence Q;

 Multiplying each sequence in the another sequence group by the transformed sequence Q bit by bit to obtain a first subcode part of the DOC code;

 Generate a series of transformation sequences I;

 Multiplying each subcode in the first subcode portion by the transform sequence I bit-by-bit to obtain another one or more subcode portions of the DOC code;

 The subcode parts are combined to obtain a DOC code group.

The selection of an arbitrary orthogonal sequence group as the initial orthogonal sequence group means that: a group of sequence groups W including N orthogonal sequences of length N can be arbitrarily selected as the initial orthogonal sequence group; where: W = {Wp w ₂ ,…, w ^}.

 The repeating the elements of each sequence bit in the initial orthogonal sequence group sequentially to obtain another sequence group refers to:

A₂、 …、 A ; When the initial orthogonal sequence group is W = {W _n W ₂ ,… ,, each sequence w, W in the W (. = (1,2, ~, Λ), the elements on each bit are repeated respectively in turn Times to get another sequence group containing N

A ₂ ,…, A;

 Among them, you can set:

= (ww, ₂ … _W! ., w) i = l, 2, "; N and:

A, = (a _{u ¾2} …) (i = 1,2, ~, N) then:

aiJR + l: ^a i R + 2 =… = ^ i _t Q + i R ⁼ W, y ₊₁

 a = l, 2, '", N; j = 0, l,"', N-l)

 The step of generating a transformation sequence Q includes:

The transformation sequence Q may be obtained by repeating the basic periodic sequence P; Corresponding to the case where each sequence in the initial orthogonal sequence group is repeated R-1 times bit by bit, P can be set to a basic period of length 2, where is a positive integer, which is composed of two lengths before and after. Both are composed of R subsequences P, and P and P are each composed of a small sequence segment having a length of 1, that is:

 P = {P, P} and

…， ^c

…, ^C

P ² jai ⁼ Wji, s ² x,…, ^s

 Among them, ^ and ^ a = ι, 2 ··, TO are small sequence segments of length R, and (and s satisfy the following orthogonal relationship:

 (^) (^ = (ζ · = 1, ··, 7θ

Among them, T represents a transposition of a row vector; the sequence P _2BY having the structure is a “complementary sequence”; the basic period of the transformation sequence Q must be a “complementary sequence”.

Further, the two subsequences and P of the basic period P _{2 ¾} can be constructed so that they are also a certain "complementary sequence", and the transformed sequence Q is constructed from the basic period after their combination.

The above construction process can be further generalized, and a "basic period" of the transformation sequence Q can be obtained based on a spanning tree structure. Each node in the spanning tree contains a "basic period", and the corresponding transformation sequence Q This "basic period" is obtained by repeating; the r-th column of the spanning tree (r = 0,1,2,…) contains a "basic period" with a length of 2 ^ = 2?, And corresponds to the initial A case where each bit of each sequence in the orthogonal sequence group is repeated Rl = 2 ^r -l times; assuming that the initial orthogonal sequence group contains N orthogonal sequences of length N in total, the transformation sequence Q is to convert the length to The 2R "basic period" is obtained by repeating N / 2 times, and the length of the obtained transformation sequence Q is NR.

的一个 "基本周期" P°, 它由两段长各为 2=R的子序列 a°和 b^Q组成， 即： One way to generate a "basic period" spanning tree of the transformation sequence Q can be described as follows: On the spanning tree, the root period contained in the root node on the 0th column is (+1 +1); the spanning tree can be Constructed in the following way: Let a node in the r-th column contain a length of

A "basic cycle" P °, which consists of two subsequences a ° and b ^Q , each 2 = R, that is:

P ° = (a ⁰ , b °)

Then on the spanning tree, the two nodes derived from this node and located in the r + 1 column contain two "basic periods" P ¹ and P ^{2 of} length 2 ² = 4i ?, respectively:

P ^J = (a ¹ , b ¹ )

= (a ⁰ , b °, a ⁰ , -b °)

P ² = (a ² , b ² )

= (a ⁰ , -b °, a ⁰ , b °)

 According to the recurrence relationship, a generator structure of the basic period of the transformation sequence Q can be obtained.

The multiplying each sequence in the other sequence group by the transform sequence Q bit by bit to obtain the first subcode part of the DOC code means that: A sequence group W of orthogonal sequences of length N is the initial orthogonal sequence group, and \ ¥ = {\ ^, W ₂ ,…, W ^}; and each sequence W (i = l, 2 in W , -, N on the element you Rl views are sequentially repeated to obtain another set of a set of N sequence groups comprising a length of NR _{-, a 2, ..., a} ^};

Each sequence A, in the another sequence group A is multiplied with the transformation sequence Q bit by bit to obtain the first subcode part of the DOC code. ¹ ^^ ¹ ^…, ^), that is:

 C ^ A ^ Q (ζ · = 1,2 ··, Λ

 Among them, the symbol "·" represents the bitwise multiplication of two sequences.

Said generating step comprises converting the sequence I: by substantially the length of period 2 (Q transform sequence substantially different from the cycle) _X-2 through repetition NR / 2Z times after generating the transformed sequence I, and the The transformation sequence I may include I 1 ₂ , ...; where:

The basic period X ₂ may consist of the sequence 8 _{2 £} , or may consist of the sequence-B ₂ ; wherein: the sequence B ₂ consists of consecutive L "+1" followed by consecutive L "-and the sequence-consists of consecutive L "—1" is composed of "+" which are consecutive Z; When different transformation sequences ft, 1 ₂ ,…, I, J are used to perform the co-transformation, the basic period of each transformation bank I can be used in the set {X _2i , = 1,2, ···} Selection, the basic period of each length can only be selected once;

Since there are two choices for each of these transformation sequences, namely B _2i or-B _{2 £} , there are 2 "'possible choices of transformation sequence series, and the codewords they generate will also have some IFW characteristics and Slot position characteristics.

The step of multiplying each subcode in the first subcode part by the transformation sequence I bit-by-bit to obtain another subcode part or parts of the DOC code means that: The first subcode part is sequentially multiplied bit by bit with a series of transformation sequences I to obtain one or more other subcode parts of the DOC code. The transformation sequence I may include 1 ₂ , ...;

 Wherein, when each subcode in the first subcode part is multiplied by a transform sequence ^ having a length of NR bit by bit to obtain the other subcode part, it may be set as follows:

The first subcode portion is the first subcode portion C ¹ , and the other subcode portion is the second subcode portion C ² ,:

The first two sub codes of the other length portions are converted to a sequence of NR 1 ₂ bitwise multiplication, and respectively 3, 4 subcode C ³ and C ⁴ portions, i.e.:

C ^ C ¹ ,-^ (= 1, 2, ·, Λ0

C ⁴ !. = CI ₂ G = 1,2 ", N)

 The above operations can be continued. Generally, when a subcode part is obtained:

CK C ² ,…, C ^L , where: L = l ^j , _ / is a non-negative integer,

Then CC \…, (^ can be multiplied bit-by-bit with a transformation sequence I of length NR, respectively, to obtain another sub-code part C ^{i + 1} , C ^{L +} \…, C ^2L , that is:

(i = \, 2,-, N, j = \, 2, "', L) When the above operation is performed a total of times, and the transformation sequence 1 ₂ ,…, l _m is used in sequence, a total of M = 2 ′ ”subcode parts ( ¹ , C…, C ^M , where each subcode part contains N Subcode codewords of length NR.

The combination of the subcode parts to obtain a DOC code group means that: M subcode parts can be combined to obtain a DOC code group containing N codewords with a total length of NRM {DOC DOC ₂ ,…, DOC ^}, Where:

DOQ = (C, C, ···, C ^M _{ (z '= l, ", N)

 This generates a DOC code group.

 In the DOC code group with a total length of NRM code words, the parameters used are N, R, and M. Different parameter combinations can be used to generate a series of DOC code groups that include different numbers of code words, code word lengths, and related properties.

 The transformation sequence Q and the transformation sequence I may use a variety of combinations to perform a variety of transformations, so that the generated DOC codeword satisfies a non-zero correlation value only at the window position, and exists near the origin. A zero correlation window,

 The above process requires at least one transformation sequence Q or transformation sequence I to be transformed. Using different said initial orthogonal sequence groups, and the same said transformation sequence Q and transformation sequence I, a batch of DOC code groups can be constructed, which can be said to form a code group set; The set of code groups described above corresponds to a set of transformation sequences Q and I.

 The M sub-code parts can be transmitted on different mutually orthogonal synchronous fading channels; that is,

 The M subcode portions may be sequentially allocated to M consecutive time slots; in order to avoid overlap between the subcode portions, a protection period composed of all zero elements may be added at the end of each of them.

However, the above conditions are not necessary because the extra correlation value due to the overlapping of the subcodes is small; especially when the overlapping portion is small, the extra non-zero correlation value introduced can generally be ignored. The invention also includes a method for applying a spread spectrum multiple access code, which is characterized in that the code words in the DOC code group are scrambled, that is, each code word in the DOC code group is respectively a sequence of the same length. Bitwise multiplication.

 The scrambling code includes the following steps:

 One scrambling code segment of length N can be intercepted in sequence from the long PN sequence;

 Repeating each bit element in each said scrambling code segment R-1 times in order to obtain a sequence of length NR;

 The entire sequence of length NR is repeated M-1 times as a whole to obtain a sequence of length NRM, that is, the scrambling code sequence; the scrambling code sequence is composed of M identical sequences, that is, the scrambling code segment. Composed of

 The M-segment scrambling code segments in the scrambling code sequence are respectively multiplied by the M sub-code partial codewords in the DOC codeword bit by bit, thereby completing the scrambling process.

 In the construction process of the DOC code, a Walsh orthogonal sequence group may be selected as an initial orthogonal sequence group. Such a DOC code group generated based on the Walsh orthogonal sequence group may be referred to as a "basic DOC code group";

 The length of the actual scrambling code segment applied to each symbol is smaller than the code length of the codeword in the basic DOC code group; where: the length of the basic DOC code is N ^, and the length of the scrambling code segment For N, N can be called "the actual effective length of the scrambling code segment", and its value is equal to the length of each sequence in the initial orthogonal sequence group.

 In a code group set, all code groups can be regarded as obtained by one basic DOC code group after being scrambled with different scrambling code segments, that is, the basic DOC code group after the scrambling is still maintained as it is. The IFW characteristics and window position characteristics of the code group.

 When the DOC code and the long scrambling code are used for spreading operation, a "multiple spreading" (Multiple Spreading) method can be adopted.

The multi-spreading means that: it can be divided into two stages for spreading: the first stage is to use the basic DOC The code group is spread, where each basic DOC code defines a code channel; the second stage is to scramble the basic DOC code by using the scrambling code sequence obtained according to the scrambling step.

 In the application method of the present invention, a cascaded "multiple spread spectrum" structure of "channelization code + (long ·) PN scrambling code" can be established by using the DOC code, and the original basic DOC code group is still maintained. IFW characteristics and related sash position characteristics. The long PN scrambling code can be used to distinguish different cells or sectors, and to whiten the interference caused by the non-zero correlation value at the window position (Whitening uses this spread-spectrum multiple access code, which can partially or even completely eliminate synchronization code division. In CDMA and spread-spectrum systems, various interference caused by non-ideal correlation characteristics may greatly increase system capacity and simplify other related interference cancellation techniques.

 In the application of the present invention, a DOC code is provided in a cellular communication system network.

The DOC code group allocation method is characterized in that: a DOC code group is allocated to each cell or sector.

 The method for allocating a DOC code group is characterized in that: each cell or sector is assigned the same basic DOC code group; at the same time, each cell or sector uses a different long PN code as a scrambling code.

 The DOC code group allocation method is characterized in that: different basic DOC code groups are allocated to adjacent cells or sectors; wherein these basic DOC code groups belong to different code group sets; at the same time, each Each cell or sector uses a different long PN code as the scrambling code.

 The invention also provides a spread spectrum multiple access code encoding device, which includes: a PN sequence generator, a Walsh code generator, a transform sequence generator, a frequency divider, and a correlator; wherein:

 The output data of the PN sequence generator is correlated with the output data of the Walsh code generator and then correlated with the output data of the transformation sequence generator to obtain output data.

The device is characterized in that the PN sequence generator refers to: there can be N PN sequence generators, and the Walsh code generator refers to: there can be N Walsh code generators, the The transformation sequence generator refers to: there can be N transformation sequence generators; wherein: The output data of the 0th PN sequence generator is correlated with the output data of the 0th Walsh code generator and then correlated with the output data of the 0th transformation sequence generator to obtain the 0th output data; the Nth PN sequence is generated The output data of the generator is correlated with the output data of the N-th Walsh code generator and then correlated with the output data of the N-th transform sequence generator to obtain the N-th output data; the 0-th output data is related to the The Nth output data is correlated to obtain the output data.

 The device of the present invention can generate a DOC (Distinctive Orthogonal Code) spread-spectrum multiple access code. The code words in the code group have the characteristics of zero interference window (IFW), and are only in certain specific positions (called It is called "window" and may have a large non-zero correlation value. BRIEF DESCRIPTION OF THE DRAWINGS

 FIG. 1 is a flowchart of a DOC code generation process in the present invention;

 FIG. 2 is a characteristic diagram of a typical cross-correlation function between two DOC codewords in the present invention; FIG. 3 is a schematic diagram of a spanning tree structure of a "basic period" of a transform sequence Q in encoding according to the present invention;

 4 is a schematic diagram of different options that transform sequence I may adopt in encoding according to the present invention; FIG. 5 is a schematic diagram of a multiple spread spectrum structure implemented by applying basic DOC code for spreading and long PN sequence for scrambling according to the present invention;

FIG. 6 is a mean square value diagram of correlation functions of two basic DOC codewords in the DOC code group (N = 32, R = 2, = 2) after scrambling (Computer simulation results, 10 ⁵ Samples); FIG. 7 is a mean square value diagram of the correlation function values of two basic DOC codewords in the DOC code group (N = 32, R = 4, M = 2) in the present invention (computer the simulation result, ¹⁰⁵ samples); FIG. 8 shows two different codes belonging to group codes DOC set of groups (N = 32 invention, R = 4, M = two substantially between the 2 codewords DOC) through scrambling correlation function values of the mean square value after the code images (simulation results, ¹⁰⁵ samples);

FIG. 9 is a schematic diagram of system networking using a DOC code as a spread spectrum multiple access code in the present invention FIG. 10 is a diagram of another ^ J¾ DOC 5 in the Ming, which is a ^ 1 ^ line system with multiple spreading lines; and FIG. 11a is a schematic structural diagram of the device according to the present invention;

 FIG. 11b is a schematic diagram of another structure of the device according to the present invention.

detailed description

 The present invention provides a new DOC (Distinctive Orthogonal Code) coding method, application method and coding device for a new spread spectrum multi-address code. See Figure 2. The related functions of the DOC code only have large non-zero values at certain specific positions, these specific positions are called "window frame". The DOC code can concentrate the interference introduced by the non-ideal correlation characteristics between the code words on these window positions. If the interference paths (including ISI, MAI, and ACI) exist only in those non-window positions, they can be completely eliminated by the DOC code. Conversely, if the interference paths exist at the window positions, you can also use some simple method. (For example, equalization, Pre-RAKE, interference cancellation, etc.) eliminates or moves the interference located on the slat location in advance to other non-slat locations, so as to achieve the purpose of eliminating interference. The application of DOC codes and other related technologies can completely or partially eliminate the above-mentioned interference in the corresponding two-way synchronous code division multiple access (CDMA) system, making it possible to establish a large-capacity digital mobile communication system.

 (1) DOC code encoding steps

 In order to achieve the foregoing object of the present invention, the method for constructing a DOC code designed by the present invention includes the following steps (the general process of generating a DOC code is shown in FIG. 1):

Step I: Arbitrarily select a sequence group containing N orthogonal sequences of length N W = {W _r

W ₂ W ^}; When N is a certain power of 2, the orthogonal sequence group must exist, such as Walsh sequence group; for convenience of description, this orthogonal sequence group is called "initial orthogonal sequence group" later Step II: Each sequence W _z in the initial orthogonal sequence group W. (i = 1, 2, ~, N) The elements on each bit are sequentially repeated R-1 times, respectively, to obtain a group containing N lengths of NR's sequence group A = {A ,, A ₂ ,…, A ^}; where R is a positive integer and can generally be chosen to be a power of 2; that is, if we assume: W _f = (w _u w… w _{! Yv} ) "· = 1,2. ， Λ

as well as,

A _z = (a _¾2 … a,) (ζ · = 1,2., Λ

Then there are:

aiJR + l ~ ^a iJR + 2 ^{= = a} W + l) R ^{= W} ij + \

 (z '= l, 2, ~, N; j = 0, l, ~, N-1)

Step III: Generate a transform sequence Q of length NR (the construction method of Q is given separately later). Then each sequence in the sequence group A is entered into _2. The first subcode part of the DOC code is multiplied by bit by bit with the transformation sequence Q ¹ = {< ¹ "= 1,2, ~, ^}, that is:

 C ^ A ^ Q (i = l, 2 .., N)

Among them, the symbol "·" represents the bitwise multiplication of two sequences.

Step IV: Multiply each subcode in the first subcode part C ¹ with a transform sequence NR length ^ bit by bit to obtain the second subcode part ( ² = { ² = 1,2, ~, 7 \ ^, i.e.: then, may be 1, 2, respectively subcode portion other long sequence of transformations is NR 1 ₂ bit-wise multiplication, and respectively 3, 4 subcode C ³ and C ⁴ portions, that is:

，Λ C ³

, Λ

C = C ² I ₂ = 1,2 .., N)

The above operations can be continued. Generally, if L = 2 j is a non-negative integer) sub-code parts C ¹ C ² C ^L , they can be respectively transformed with a NR transformation sequence 1 _£ bit by bit. Multiplying to obtain another sub-code part C ^{i + 1} , C ^{L + 2} C ^2L , that is:

0 ^ = 0 · Ι '= 1,2, ~, N, _ / = 1,2, —,) If the above operations are performed a total of times (using the transformation sequence II ₂ l _{m in turn} ), a total Af = 2 '"subcode sections ( ¹ , C ² C ^M , where each subcode section contains N subcode codewords with a length of NR. The construction method of the transformation sequence I (including I 1 ₂ ,…) will also be given separately later.

Step V: Combine the M sub-code parts in a certain way to obtain a DOC code group containing N codewords with a total length of NRM {DOC ^ DOC ₂ DOC, where:

DOC, = (C ¹ ,., C, C ^.) (Z '= l, 2 .., N)

 This generates a DOC code group, where the parameters used are (N, R, M). Using different parameter combinations can generate a series of DOC code groups containing different numbers of codewords, codeword length, and related properties. In steps III and IV, various flexible combination of the transformation sequence Q and the transformation sequence I can be used to perform various transformations, so that the generated DOC codeword satisfies a non-zero correlation value only at the window position, and There is a zero correlation window near the origin (here the correlation function is defined as the sum of the correlation functions of each subcode part). Note that in the above process, the transformation must be performed at least once (via a certain transformation sequence Q or I).

 不同 A group of DOC code groups can be constructed by using different initial orthogonal sequence groups and the same transformation sequences Q and I. For the convenience of description, the batch code groups are said to form a "code group set". Each "code set" corresponds to a set of transformation sequences Q and I.

 In practical applications, these M subcode parts should be transmitted on different mutually orthogonal synchronous fading channels. For example, the M subcode portions may be sequentially allocated to M consecutive time slots. At this time, in order to avoid overlap between the subcode parts, a protection period composed of all zero elements may be added at the end of each of them. However, this is not necessary. In fact, although the overlap of the sub-code parts will destroy the all-zero correlation value at the original non-window, only small correlation values will be entered at these positions. In general, You can ignore their effects.

 The construction of the transform sequences Q and I is one of the cores of the DOC code. The methods for generating them are given below.

 Generation method of transformation sequence Q:

The transformation sequence Q is obtained by repeating some so-called "basic period" sequence P. Corresponding to the case where each sequence in the initial orthogonal sequence group is repeated R-1 times bit by bit, let P _{2 ¾} be 2KR in length A "basic period" (where K is a positive integer), which is composed of two subsequences P, both before and after the length of KR, and P, P ² ^ each by the length of R The sequence segment consists of:

以及

as well as

^ KR ⁼ C ² R, ■ ·,, ^C

s² ?， ···， ^s

s ² ?, ···, ^s

Among them, c and s (i = \, 2, ..., K) are both small sequence segments of length R, and there is sufficient orthogonal relationship between c and s:

(c ^) (s ^) ^T = 0 "· = 1,2.,)

Among them, T represents the transposition of the row vector. For ease of description, a sequence with this structure is called? Is a "complementary sequence". The "fundamental period" of the transformation sequence Q must be a "complementary sequence". It can be explained that based on the above-mentioned form 妁 "basic period" P _2ffli 's transformation sequence Q, the _{value of the auto-} / cross-correlation function of the generated sub-code part codeword at the position of (2j l) XR (where j is an integer) Is zero.

(其中 为正整数） 点位置上的自 /互相关函数值为零。 Further, it is possible to construct two subsequences P before and after the "basic period" P _{2 ¾} and make them also a certain "complementary sequence". At this time, the transformation sequence Q constructed by the "basic period" P after combining them is generated Subcode part of the codeword,

(Where is a positive integer) The value of the auto / cross correlation function at the point position is zero.

By extending the above-mentioned construction process, a "basic period" of the transformation sequence Q can be obtained based on a spanning tree structure (see FIG. 3). Each node in the spanning tree contains a "basic period", and the corresponding transformation sequence Q is obtained by repeating the "basic period". The r-th column of the spanning tree (r = 0,1,2,…) contains a "basic period" with a length of 2 ^ = 2, and corresponds to repeating each bit of each sequence in the initial orthogonal sequence group by R- The case when l = 2 ^r -l times. Assuming that the initial orthogonal sequence group contains N orthogonal sequences of length N, the transformation sequence Q is obtained by repeating the length 2R "basic period" N / 2 times, and the length of the obtained transformation sequence Q is for? . Correspondingly, it can be shown that in the subcode part generated by this transformation sequence Q, except for the odd-numbered points, The value of the autocorrelation function of each codeword may have a nonzero correlation value only at the point of J2R (j is a non-zero integer), and the value of the cross-correlation function between any two codewords is only at the point of j2R (j is an integer) There may be non-zero correlation values.

The method of generating a "basic period" of a transformation sequence Q based on a spanning tree can be described as follows. In the spanning tree, the basic period contained in the root node on the 0th column is (+ 1 + 1). The spanning tree can be constructed as follows: Suppose that a node in the r-th column contains a "basic period" P ⁰ of length 2 2R, which consists of two subsequences ^ and 2 of 2 =? Each! ^ Composition, ie

Ρ ⁰ = (a ⁰ , b °)

Then on the spanning tree, the two nodes derived from this node and located in the r + 1 column contain two "basic periods" P ¹ and P ^{2 of} length 2 ² = 4 ?, respectively:

P ^J = (a ¹ , b ¹ )

= (a ⁰ , b ⁰ , a ⁰ , -b °)

P ² = (a ² , b ² )

= (a ⁰ , -b °, a ⁰ , b °)

 According to the recursive relationship, we can get a kind of bowl cycle of changing column Q.

 On the other hand, when different "basic periods" P of the same length are used to construct the transformation sequence Q, the obtained different subcode codewords may also have certain IFW characteristics and window position characteristics.

 Generation method of transformation sequence I:

The transformation sequence I (including I 1 ₂ ,…) is also obtained by repeating a certain "basic period" (see Figure 4). The NR transformation sequence I is obtained by repeating NR / 2 times for a "basic period" X _{2L of} length 2L (= 1,2, ...). Among them, the "basic period" X _{2i is} composed of consecutive L's (set as the sequence _2L ), or composed of consecutive L '-' + '(ie, the sequence-B ₂ ). E.g:

Basic period X _{2 of} length: B ₂ = (+1 -1) or —B ₂ = (-1 +1)

Basic period X _{4 with} degree ₄ : B ₄ = (+ 1 + 1— 1— 1) or -B ₄ = (-1 -1 +1 +1)

Basic period of length X ₈ : B ₈ = (+1 + 1 + 1 +1 1 -1-1 -1) or

-B ₈ = (-1 -1 one 1 one 1 +1 + 1 + 1 +1)

It can be illustrated that, based on the transformation sequence I constructed based on the "basic period" ₂ , the sum of the values of the auto / cross-correlation functions at the position of (2 Η) £ (where positive integers) of the two subcode codewords is zero.

When m different transformation sequences ft, 1 ₂ 1, „} are used in turn for a total of m transformations, the basic period of each transformation sequence I may be in the set {X _{2 £} , = 1,2, ...} Selection, but the "basic period" of each length can only be selected once. Since there are two choices for each transformation sequence (ie B ₂ or -B _2i ), there are a total of 2 '"possible transformation sequence series. When selected, there will also be some IFW characteristics and sash position characteristics between the codewords they generate.

 In the aforementioned generation process of the DOC code, in steps III and IV, various transformations made by using the transformation sequences Q and I can adopt various flexible combinations. The reasonable selection of various transformation sequences Q and I will make the value of the auto / cross-correlation function at some positions zero (if the correlation of each subcode part is performed separately, and the total correlation function is the correlation function of each subcode part with). The correlation values of the remaining positions are generally non-zero values-these positions are called "window" positions. In particular, if the correlation function values within +/- W near the origin are all zero, the resulting code group has a zero interference window (IFW) with a single-sided width of W.

 For example, ¾ 殳 wants to construct such a DOC code group, which contains N = 16 DOC codewords with a length of 128. Its correlation function value is only an integer. The point position may have a non-zero correlation value, and the width of a single IFW Is equal to 7. With the related encoding method described in the present invention, different DOC code groups can be generated in the following ways, which respectively use different combinations of the transformation sequences Q and I:

Method 1: N = 16, R = 4, M = 2, where a transformation sequence Q (basic period P ₈ , generated by the spanning tree) and a transformation sequence I (basic period X ₂ ) are used;

Method 2: N = 16, R = 2, M = 4, where a transformation sequence Q (basic period P ₈ , Complementary sequence is sufficient) and two transformation sequences I (the basic period is ₂ and ₂ respectively); Mode 3: N = 16. R = 2, M = 4, where one transformation sequence Q (basic period P ₄ ) and two Secondary transformation sequence I (basic periods are ₂ and ₈ respectively);

 Method 4: N = \ 6, R = l, M = 8, where three transformation sequences I are used (the basic periods are

X ₂ , X ₄ and X ₈ );

 (2) Scramble process of DOC code

Based on any initial orthogonal sequence group, a DOC code group can be generated according to the related generation method described in the present invention; and there are 1 ^N orthogonal sequence groups containing N sequences of length ^N (if not considered Equivalence of code words in the code group in order or inverse transformation), so the number of DOC code groups is very large. In an actual cellular CDMA communication system, a DOC code group can be allocated for each cell or sector. This method corresponds to the "short code CDMA" method.

 On the other hand, if the code hopping method is used, the performance among users can be averaged. Using long PN code for scrambling is a very convenient and efficient code hopping method, which corresponds to the "long code CDMA" method.

 In the step of generating the DOC code, a Walsh orthogonal sequence group may be selected first, and each codeword in the sequence group is multiplied by a sequence of the same length bit by bit (the sequence is taken from a section in a long PN code, which The length is equal to the length of each sequence in the Walsh sequence group), thus generating an initial orthogonal sequence group. Then, construct the DOC code group according to other relevant steps.

Since the construction of the DOC code group involves repetition and bit-by-bit multiplication operations, the scrambling step can also be delayed to the last step. At this time, the DOC code group is generated by the Walsh sequence group as the initial orthogonal sequence group, and the corresponding DOC code group is referred to as a "basic DOC code group". Then, the original scrambled code segment is repeatedly transformed with the basic DOC code. Each subcode part of the group is multiplied bit by bit to complete the scrambling operation. In a code group set, all code groups can be regarded as obtained by scrambling different scrambling code segments of a basic DOC code group. After scrambling, the basic DOC code group still maintains the original IFW characteristics and the characteristics of the sash position. The codewords in the "basic DOC code group" can be scrambled, that is, each codeword is multiplied bit by bit with a sequence of the same length. For a DOC code group with parameters N, R, and M), the scrambling process is divided into the following steps:

 Step A: Truncate one "N" scrambling code segment of length N from the long PN sequence, D;

 Step B: Repeat each bit element in each "scrambling code segment" R-1 times in order to obtain a sequence E of length NR;

 Step C: The sequence E is repeated M-1 times as a whole to obtain a sequence F of JVRM, that is, a scrambled sequence; the scrambled sequence is composed of the same sequence of M segments (that is, "scrambled segments");

 Step D: Multiply the M-segment "scrambled-segment" in the scrambling sequence E with the M sub-code part codewords in the DOC codeword, respectively, to multiply the bits one by one, thereby completing the scrambling process;

 It is noted that, unlike the existing CDMA system (for example, IS-95 uses the Walsh sequence to lengthen the PN code scrambling method), the length of the actual scrambling code segment applied to each symbol is longer than the codeword of the basic DOC code group That is, the processing gain or spreading factor) should be small. Here, the length of the basic DOC code is NRM, and the length of the scrambling code segment is therefore. In the following description of the scrambling process, for convenience of description, N is called the "actual effective length" of the scrambling code segment, and its value is Equal to the length of each sequence in the initial set of orthogonal sequences.

When the DOC code and the long PN sequence scrambling code are used for the spreading operation, a so-called "multiple spreading" (Multiple Spreading) method may be adopted, as shown in FIG. 5. Multiple spread-spectrum methods have been widely adopted in existing commercial CDMA systems. Multiple spreading for DOC codes can be divided into two levels: The first level is spreading with basic DOC codes, where each basic DOC code defines a code channel (Code Channel). The codeword is a "channelization code" (Channelization Code); the second level is Scrambling using a (repeatedly transformed) long PN code. The purpose of the scrambling code is to whiten the interference caused by the non-ideal correlation characteristics of codewords in the same cell and between cells, and can be used to distinguish the code groups used in different cells or sectors to achieve the same basic DOC. "Reuse" of code groups in different cells or sectors (reuse).

 (3) Examples of encoding methods:

 The present invention is described in detail below through examples and drawings. In the foregoing construction process, the selection of different R and M parameters will generate DOC code groups with different characteristics. Here are some typical examples to explain the construction method of DOC codes in detail, and briefly give the corresponding DOC code groups. Related characteristics.

 Example 1: DOC code group (N = 4, R = 2, M = 2)

 First, the coding steps of the present invention will be explained with a simple example. Suppose that the initial orthogonal code group is selected as a Walsh sequence group containing 4 sequences of length 4 (corresponding to the encoding step 1). They are:

 (+1 +1 +1 +1)

 (+1 -1 +1-1)

 (+1 +1 -1-1)

 (+1 -1 -1 +1)

 Then, each bit of each of the above codewords is repeated one by one (corresponding to the encoding step II), and a sequence group A of 4 sequences of length 8 is obtained:

 (+1 +1 +1 +1 +1 +1 +1 +1 +1)

 (+1 +1 one 1 -1 +1 +1 -1 -1)

 (+1 +1 +1 +1 -1 -1 -1 -1)

 (+1 +1 -1 -1 -1 -1 +1 +1)

 Next, the basic period is selected on the spanning tree of the basic period of the transformation sequence Q. There are two basic periods in the first column of the spanning tree, which are (+1 +1 + 1 -1) and (+1 -1 +1 +1). Repeat these two basic periods twice to get two An optional transformation sequence Q of length 8, let them be Q1 and Q2:

Q1 = (+1 +1 +1 one 1 +1 +1 +1 one 1) Q2 = (+1-1 +1 +1 +1 -1 +1 +1) Select Q1 as the transformation sequence, multiply each sequence in the aforementioned sequence group A bit by sequence Q1 (corresponding to encoding step III) ), Get 4 sequences of length 8:

 c = (+1 +1 +1 -1 +1 +1 +1 -1)

 c = (+1 +1 -1 +1 +1 +1 -1 +1)

 c = (+1 +1 +1 -1 -1 -1 -1 +1)

 c = (+1 +1 -1 +1 -1 -1 +1 one 1)

They form the first subcode part of the DOC code.

 In order to generate another subcode part, the basic period of the transformation sequence can be (+1-1) or (-1 + 1). At this time, the corresponding transformation sequences are:

I _t = (+1 -1 +1 -1 +1 —1 +1 — 1) or

I _t = (-1 +1-1 +1 -1 +1 -1 +1)

 Select the previous ^ as the transformation sequence, and multiply each sequence in the first subcode portion obtained by the sequence ^ bit by bit (corresponding to the encoding step IV) to obtain another 4 sequences of length 8:

 c \ = (+1 -1 +1 +1 +1 -1 +1 +1)

 c \ = (+1 -1 -1 -1 +1 -1 -1-1)

C ² ₃ = (+1 -1 +1 +1 -1 +1 -1-1)

 \ = (+1 -1 -1 -1 -1 +1 +1 +1)

They form the second subcode part of the DOC code.

 Combining the first subcode part and the second subcode part (corresponding to the encoding step V), a DOC code group containing 4 codewords with a total length of 16 is obtained:

DOC1 = (C ¹ !, C ² !) = (+1 +1 +1 -1 +1 +1 +1 -1, + 1—1 +1 +1 +1 -1 +1 +1) DOC2 = ( C ¹ ,, C ² ₂ ) = (+ 1 + 1- 1 + 1 + 1 + 1- 1 + 1, + 1-1-1-1 + 1 1-1-1) DOC3 = (C ^, C ² ₃ ) = (+ 1 + 1 + 1-1-1 -1-1 + 1, + 1-1 + 1 + 1-1 + 1-1-1) DOC4 = (C ^! ₄ , C ² ₄ ) = (+ 1 + 1—1 + 1— 1—1 + 1—1, + 1—1-1—1—1 + 1 + 1 + 1) The auto / cross-correlation function values between the four codewords are given in Table 1 (the correlation function here is the sum of the respective correlation functions of the sub-code parts without considering the overlap between the sub-code parts), where R ^ represents the correlation function between coders DOCX and DOCY (autocorrelation function when X = Y, cross-correlation function when X ≠ Υ). It can be seen from the table that all related functions may have non-zero values only at +/- 4 points, which are the "window frame" positions. In particular, in the interval (-3, +3), except for the maximum value of the auto-correlation function at the origin, the values of the auto / cross-correlation functions at all other points are zero, so the DOC code group has a zero interference window ( IFW), the width of one side is equal to 3. Table 1 DOC sequence (N = 4, R = 2, M = 2) related mechanical in-line Walsh sequence, the bowl period of the change column Q is (+1 +1 + 1-1), and the bowl of the change column ¾ The period is (+1 -1)

Since the above code group is generated using the Walsh sequence group as the initial orthogonal sequence group, it is a "basic DOC code group" as defined above. The basic DOC code group is scrambled. Since the "actual effective length" of the scrambled code segment is 4, an initial scrambled code segment of length 4 can be selected. For example, the scrambled code segment can be selected as (+1 +1 -1 +1). First, each bit of the scrambling code segment is repeated once to obtain a sequence: (+1 +1 +1 + 1-1 -1 +1 +1)

Then repeat the whole sequence once again to obtain a 4 e-code sequence with a length of 16:

 (+1 +1 +1 +1-1-1 1 +1 +1, +1 +1 +1 +1-1-1 1 +1 +1)

 After multiplying each scrambling code sequence with each codeword in the basic DOC code group obtained previously, the following DOC code group containing 4 scrambled codewords is obtained:

 DOC1 = (+1 +1 + 1—1 -1-1 + 1-1, + 1-1 1 +1 +1 -1 +1 +1 +1) DOC2 = (+ 1 + 1-1 +1 -1 -1 -1 +1, +1 -1 -1 -1 -1 +1 -1 -1) DOC3 = (+1 + 1 + 1-1 + 1 + 1-1 + 1, + 1-1 + 1 + 1 + 1-1-1-1) DOC4 = (+1 + 1-1 + 1 + 1 + 1 + 1-1, + 1-1-1- 1 + 1-1 + 1 + 1) Table 2 The auto / cross-correlation functions between the four scrambled codewords are given in. It can be seen from the table that the correlation characteristics between the scrambled code group and the basic code group are exactly the same, that is, there may be non-zero correlation values only at the "window" at +/- 4 points. The difference between the two The correlation value is different at the window position. In fact, after long PN code scrambling, the correlation function value of the DOC code group at the window position can be whitened and Gaussianized.

(N=4、 R = 2. M=2)的相关) ¾¾, ¼½正 列组为 Walsh序 列、 变辦列 Q的鉢周期为（+1+1 + 1-1)、变辦列 ¾的鉢周期为（+1-1), 扰 (+1+1—1+1) Table 2

(N = 4, R = 2, M = 2) Correlation) ¾¾, ¼½ is the Walsh sequence, the bowl period of the change column Q is (+ 1 + 1 + 1-1), and the change column is ¾ The bowl period is (+ 1-1), and the disturbance is (+ 1 + 1—1 + 1)

Relative shift

 —7 —6 —5 -4 -3 -2 -1 0 1 2 3 4 5 6 7-bit:

 0 0 0 0 0 0 0 16 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 16 0 0 0 0 0 0 0 i? ₃₃ (r) 0 0 0 0 0 0 0 16 0 0 0 0 0 0 0

^ 44 (r) 0 0 0 0 0 0 0 16 0 0 0 0 0 0 0 0

 0 0 0 -8 0 0 0 0 0 0 0 0 -8 0 0 0

 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 -8 0 0 0 0 0 0 0 0 8 0 0 0

 In addition, during the above-mentioned construction process, other optional basic periods may also be used to construct the transformation sequences Q and I, respectively, to generate code groups in other DOC code group sets. You can verify that they have the same IFW width and properties related to the position of the sash as in the previous example. No longer praise here.

 Example 2: DOC code group (N = 4, R = 1, M = 4)

 First, the initial orthogonal code group is a Walsh sequence group containing 4 sequences of length 4 (corresponding to the encoding step I). They are:

 (+1 +1 +1 +1)

 (+1 -1 +1-1)

 (+1 +1 -1-1)

 (+1 -1 -1 +1)

 Since R = 1 here, it is equivalent to not using the transformation sequence Q for corresponding transformation (that is, skip encoding steps II and III). Therefore, the above Walsh sequence group constitutes the first subcode part of the DOC code, that is:

 c = (+1 +1 +1 +1)

c ⁼ (+1 -1 +1-1)

 ^ 3 = (+1 +1 -1-1)

^ ₄ = (+1 -1 -1 +1)

 In order to generate the second subcode part, the basic period of the transformation sequence I, can be selected as (+ 1-1) or (-1 +1), and the former basic period is selected. At this time, the corresponding transformation sequences are:

1] = (+1 1 1 +1 1 1) multiply each sequence in the first subcode part by the sequence ^ bit by bit (corresponding to the encoding step IV), to obtain another 4 sequences of length 4, which constitute the second subcode part of the DOC code:

 C = (+1 -1 +1 -1)

c ² ₂ = (+ i +1 +1 +1)

 -1 -1 +1)

C ² ₄ = (+1 +1 -1 one 1)

To generate the 3, 4 subcode portion, converting the basic cycle sequence of ₁₂ selected to (+1 +1 -1--1) or (-1 -1 +1 +1), before selecting a basic period II, The corresponding transformation sequences are now

 +

For:

1 ₂ = (+1 +1 -1 -1)

Each sequence of the first sub-code section, respectively, multiplied by the sequence bit by bit ₁₂ (corresponding to the encoding step IV), to give the other four sequences of length 4, which constitutes part of the third sub-code code DOC: c ³ ^ (+1 +1 -1 -1)

c ³ ₂ = (+1 -1 -1 +1)

 (+1 +1 +1 +1)

c ³ ₄ = (+1 -1 +1 -1)

The sequence of each of the second sub-code section, respectively, multiplied by the bit-wise sequence ₂ (corresponding to the encoding step IV), to give the other four sequences of length 4, which constitutes a part of the fourth sub-code code DOC: c \ = (+1 -1 -1 +1)

 c \ = (+1 +1 -1 -1)

 c \ = (+1 -1 +1-1)

C ⁴ ₄ = (+1 +1 +1 +1)

 Combining the above 4 subcode parts (corresponding to the encoding step V), a DOC code group containing 4 codewords with a total length of 16 is obtained:

(+1+1-1-1, +1-1-1+1, +1+1+1+1, +1-1+1-1) DOC4=(C¹ ₄, (¾， C?₄, (¾)= (+1-1-1+1， +1+1-1-1, +1-1+1-1, +1+1+1+1) 可以验证， 该组 DOC码的自 /互相关函数均是理想的， 即各码字的自相 关函数仅在原点取到非零值（等于 16)、 任意两个码字之间的互相关函数为 全零（这里的相关函数为各子码部分各自相关函数的和， 而不考虑子码部 分之间的重叠）。 ' DOC CC ² !, C ³ !, C ⁴ ₁ ) = (+ 1 + 1 + 1 + 1, + 1-1 + 1-1, + 1 + 1-1-1, + 1-1-1 + 1 ) DOC2 = (C ¹ _{? 5} C ² ,, ₂ , C ₂ ) = (+ 1-1 + 1-1, + 1 + 1 + 1 +1, + 1-1-1 + 1, + 1 + 1 -1-1)

(+ 1 + 1-1-1, + 1-1-1 + 1, + 1 + 1 + 1 + 1, + 1-1 + 1-1) DOC4 = (C ¹ ₄ , (¾, C? ₄ , (¾) = (+ 1-1-1 + 1, + 1 + 1-1-1, + 1-1 + 1-1, + 1 + 1 + 1 + 1) It can be verified that the The auto / cross-correlation functions are ideal, that is, the auto-correlation function of each codeword only takes a non-zero value (equal to 16) at the origin, and the cross-correlation function between any two codewords is all zeros (the correlation function here) Is the sum of the respective correlation functions of each subcode part, regardless of the overlap between the subcode parts. '

 The above basic DOC code group can be scrambled. Since the "actual effective length of the scrambled code segment is 4", the initial scrambled code segment with a length of 4 can be selected. For example, the scrambled code segment can be selected as (+1 +1 -1 +1). Repeat the entire scrambling code segment Ml = 3 times to obtain a scrambling code sequence with a total length of 16: (+ 1 + 1—1 +1, +1 +1-1 + 1, +1 +1-1 + 1, + 1 + 1-1 + 1)

 After multiplying each scrambled code sequence with each codeword in the basic D0C code group obtained previously, the following D0C code group containing 4 scrambled codewords is obtained:

 D0C1 = (+1 + 1—1 +1, + 1-1-1—1, + 1 + 1 + 1-1, + 1—1 + 1 + 1)

 D0C2 = (+ 1-1-1-1, + 1 + 1-1 + 1, + 1-1 + 1 + 1, + 1 + 1 + 1-1)

 D0C3 = (+1 +1 + 1-1, + 1-1 + 1 + 1, + 1 + 1—1 + 1, + 1-1-1—1)

 D0C4 = (+1 -1 + 1 + 1, + 1 + 1 + 1—1, + 1-1 -1 -1, + 1 + 1—1 + 1) It can also be verified that this group is scrambled D0C The auto / cross-correlation function of the code group is still ideal,

 In addition, during the above-mentioned construction process, other optional basic periods may also be used to construct the transformation sequences ^ and ^, respectively, to generate code groups in other DOC code group sets. You can verify that they have the same IFW width and properties related to the position of the sash as in the previous example. I won't repeat them here.

 Example 3: D0C code group (N = 32, R = 2, M = 2)

As a more complicated example, a Walsh sequence group containing 32 sequences of length 32 is selected as an initial orthogonal sequence group to generate a basic DOC code group containing 32 codewords of length 128. Here, the parameter is selected as R = 2, that is, each sequence in the initial orthogonal sequence group is repeated bit by bit in turn. Repeat R-1 = 1 times. The basic period of the transformation sequence Q is selected in the first column of its spanning tree. Here, the basic period of the transformation sequence Q is selected as (+1 +1 + 1 -1), and it is repeated 16 times to obtain the transformation sequence Q. The basic period of the transformation sequence I is selected as (+1 -1) and repeated 32 times to obtain the transformation sequence. According to the above-mentioned DOC code generation method, a DOC code group containing 32 codewords can be obtained as follows:

 Basic DOC code group (iV = 32, R = 2, M = 2), the basic period of the transformation sequence Q is (+1 +1 + 1-1), and the basic period of the transformation sequence is (+ 1-1)

Codeword 1

 (+++ — +++ — +++ — +++ — +++ — +++ — +++ — +++ — +++ — +++ — +++ — +++ — + ++ _ +++ — +++ — +++ —, + — +++ _ +++ — ++-+++ — +++ — +++ _ +++ — +++ — + ++ — +++ _ +++ — +++ — +++ _ +++ — +++ — ++) Codeword 2

 (++ — +++ — +++ — +++ — +++ — +++ — +++ — +++ — +++ — +++ — +++ — +++ — ++ + — +++ — +++ — +++ — +,

+ — + — + + — + — + — + — + — + — + — + — + — + — + — + — + — + —) Codeword 3

(+++ ++++ ++++ ++++ ++++ ++++ ++++ ++++ + ₅

+-++-+ 一一 +-++-+ — +-++-+ 一一 +-++-+ 一一 +-++-+ 一一 +-++-+ 一一 +-+ +-+ 一一 +-++-+ ——) codeword 4

 (++-+ one one +-++-+-one +-++-+ one one +-++-+ one one +-++-+ one one +-++-+ one one +-+ +-+ — 一 +-++-+ 一一 + —, + ++++ ++++ ++++ ++++ ++++ ++++ ++++ +++) code Word 5

(+++-+++ + ++++-+++ + ++++-+++ + ++++-+++ + + ₅

+-+++-++-+ + One-One +-+++-++-+ + One-One +-+++-++-+ H—— +-+++-++-+ + One one) code word 6

(+ H h + H 1 1 1 h + H 1 1 1 h ~ l hH 1 1 1 hH— h + H 1 1 1—, + + ++ a + + + + + + +-+ + + + +++ — +++ + +++-++) ί) Ηi + — + hτ +++++ ί ++ +++ +++++ Γί ~

 (++ + +++ — + T ++ + + ++ Ϊ + † ++ i + ———. Ι1 ————— lΙi) ++ + — + + — + + ++ 卞 + + + — +++ 卞 + I | ——I ———— |

(+-+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + I I

 6) † + + — + ++ ++ + + + + + + + + + + + + + + + + + + + + 1 |

 (+ + + + — +++ — ++ — ++ + + +++ | II II |

+ + + + + + + + + + + + — + + + + Γ— II |

 + ΐ + + + ++ ± ΐ ++ ί + ten + + !. ιι} ί + ΐί ίί + + +++++ ΐ + — + ++++ +

 (++ † — ++++++++++ ++++ + Ten + + † + +-+ T—— II—— ~ —— -II II1. Ί1——Ι} + +++ ++ ++ † + 卞 +++ + T ++++ — +1 ++ + — + 卞 +++ Ι | —— | —— -—— ΙΙ——1——1——

 (— +++ +++ — +++ ++ Ϊ + ++++ ++ ++++ + +++++

 Π) + + + + + + + + + + + + + + + + ΐ + + ί + + + ΐ + + Ι1-!

 (— + +++++ — ++++ +++ † + † +++ + ++ † ——1 II1) ++++ — + +++ — +++++ + +++ + — + | II II || 1

(+ ++++++++++ + ++ 4 † ++ + ++ +11 O / iiAV 9is3 /: / isl £ C99i) +++ ++++++++ ++++ + + + + + + + + + + + + + + + ΙΙ-

(+++++ ++++ +++ ++++++++++ + ++ + +++++ + IΙΙ ίI1 -IΙΙΙ 1-ΙΙΙIIII—— IIΙΙ) +++ Τ + + ++++++++++ _ +++++ + + +++ +++ ++ ΓΙΙ-ΙΙΙ-ΙΙΙΙ --ΙΙ --.-

(+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +! Ι

+++++ +++++ +++ + + +++ —— I——I IIIII-

(++++ ++++ ++ ++++++++ ++++ + + + IIII ϋ ε9υ /: 1 £ 9ii OooiAV) + +++++ ++ +++ + ++++++--

 o

0Z (+++ + ++++ ++++-+++ + ++++-+++ ++++ + ++++ —,

+-++-HH +-++-+ —— +-+++-++-+ + —— +-+++-++-H +-++-+ h one one +-++ ) Codeword 32

 (++-+ 一一 + +-++-+ 一一 +-++-+++-+ 一一 + +-++-+++-+ 一一 +-++-+ _ 一 + +-++-+,

+ +++-++++ ++++ H +++-++++ + ++++ +++-++++) It can be verified that the position of the sash of this group of codewords is relatively biased The shift is equal to 4k points, where k is a non-zero integer.

Scramble the basic D0C code group (where no protection period is added between the first and second subcode parts, and the overlap between the first and second subcode parts is counted). The scrambling code sequence is random by the computer. produced. Different scrambling code segments are used (when the length is N = yi, the correlation values between the obtained codewords are generally different except for the origin. From a statistical point of view, when the long scrambling code used is sufficiently random, The value of the correlation function at each relative offset will be a Gaussian distribution with a mean value of zero. Figure 6 shows that in the above basic DOC code group, after codeword 1 and codeword 10 are scrambled, The mean square value of the correlation function value at the offset point (that is, the average of the squared values of all statistical samples). Assuming that the receiver receives the target code channel corresponding to codeword 1, the mean square value directly corresponds to knowing the expanded codeword 10 The average interference power value caused by codeword 1 at each offset position. As can be seen from Figure 6, at the position of the window (point 4 / _C , where A is a non-zero integer) there is between codeword 1 and codeword 10. The larger the mean square value of the correlation function; while at the non-window position, the mean square value of the correlation function is small. Note that the non-zero correlation value at the non-window position is due to the first and second of the DOC codeword Caused by the overlap between subcode parts, but their values are generally small (especially Near the origin), even considering the situation of 31 other interfering users, within the original correlation window (ie, relative offset positions of + 1 / -1, + 2 / -2, +31-3), at each point due to The total interference caused by the overlap of the two subcodes is relatively small compared to the target signal (below -20dB, assuming 31 interference codewords), so it can generally be ignored in the actual system design.

As mentioned above, other possible basic periods can also be selected to construct the transformation sequences Q and I, to Generate code groups from other DOC code group sets. If it is assumed that the transformation sequence repeated from the basic period (+1 + 1 + 1-1) and (+1-1 +1 +1) is Ql and Q2 respectively; from the basic period (+1-1) and (-1 +1) The transformation sequences obtained repeatedly are II and 12, respectively, then the following four code group sets and corresponding transformation sequence (Q, I) pairs can be defined:

 Code set SETA: (Ql, II)

 Code set SET B: (Q2, II)

 Code set SET C: (Ql, 12)

 Code set SET D: (Q2, 12)

 Then there is some correlation between the four code groups, which are listed in Table 3. Marked in shaded squares): Table 3 Correlation characteristics between code group sets (labeled as unilateral IFW width)

 Example 4: DOC block (N = 16, R = 4, M = 2)

 In this example, a Walsh sequence group containing 16 sequences with a length of 16 is selected as an initial orthogonal sequence group to generate a basic DOC code group containing 16 codewords with a length of 128. Here, the parameter is selected as R = 4, that is, each sequence in the initial orthogonal sequence group is repeated bit by bit in sequence?-1 = 1 times. The basic period of the transformation sequence Q can be selected in the second column of its spanning tree, corresponding to different basic periods. After repeating 8 times, a total of 4 transformation sequences Ql, Q2, Q3 and Q4 can be generated, of which

Basic period (+ 1 + 1 + 1-1 +1 +1 -1 +1) → Transformation sequence Q1; Basic period (+1 + 1—1 +1 +1 +1 +1 +1 -1) → Transformation sequence Q2; Basic period (+ 1-1 + 1 + 1 +1 -1-1-1-1) → transform order 歹 ij Q3; Basic period (+1 -1—1—1 +1 -1 + 1 + 1) → transformation order 歹 'J Q4;

The other transformation sequence I has two options. Let the sequence generated by repeating the basic period (+1 -1) be II, and the sequence generated by repeating the basic period (+1 -1) is 12. Therefore, the transformation sequence pair (Q, I) has a total of 8 possible selection combinations, which respectively correspond to 8 code group sets SET A-SET H. The transformation sequence pairs corresponding to the eight code group sets are:

 Code set SETA: (Ql, 11)

 Code set SET B: (Q2, 11)

 Code set SET C: (Q3, 11)

 Code set SETD: (Q4, 11)

 Code set SET E: (Ql, 12)

 Code set SETF: (Q2, 12)

 Code set SET G: (Q3, 12)

 Code Set SETH: (Q4, 12)

 The code groups in the eight code group sets and some of them have some correlation characteristics, which are listed in Table 4 (see also Figure 3 for related characteristics):

 Table 4 Correlation characteristics between 8 code group sets (labeled as single-sided IFW width)

SETA SETB SETC SETD SETE SETF SETG SETH

IFW = IFW = IFW = IFW =

 SETA

 7 3 1 1 orthogonal orthogonal orthogonal 'orthogonal

IFW = IFW = IFW = IFW =

 SETB

 3 7 1 1 orthogonal orthogonal orthogonal orthogonal

IFW = IFW = IFW = IFW =

 SETC

 1 1 7 3 orthogonal orthogonal orthogonal orthogonal

IFW = IFW = IFW = IFW =

 SETD

 1 1 3 7 orthogonal orthogonal orthogonal orthogonal

 IFW = IFW = IFW = IFW =

SETE Orthogonal Orthogonal Orthogonal Orthogonal 7 3 1 1

 IFW = IFW = IFW = IFW =

SETF orthogonal orthogonal orthogonal orthogonal 3 7 1 1

 IFW = IFW = IFW = IFW =

SETG orthogonal orthogonal orthogonal orthogonal 1 1 7 3

 IFW = IFW = IFW = IFW =

SETH orthogonal orthogonal orthogonal orthogonal 1 1 3 7 Two basic DOC code groups in the above code group sets SET A and SET B are given below, where each code group contains 16 code words with a length of 128.

 The basic DOC code group (N = 16, R = 4, M = 2) in the set of DOC code groups is set, and the bowl period of the transformation sequence Q is (+ 1 + 1 + 1-1 +1 +1 +1 +1 ). Change the bowl period of ^^ column ^ to (+1-1). Codeword 1

 (+++ — ++ — ++++ — ++ — ++++ _ ++ — ++++ _ ++ — ++++ — ++ — ++++ — ++ — ++ ++ — ++ — ++++ — ++ — +,

+-+++ — +-+++ +-+++ +-+++ — +-+++ +-+++ +-+++ — +-+++) Codeword 2

 (+++ +-++ HH · — ++ H +-+++ +-+++ +-+++ +-+++ +-+++ + —, + — ++ _ +++ + — ++-++++ — ++ — ++++ — ++ — ++++ — ++ — ++++ — ++ — ++++ — ++ — ++++ _ ++-+++) Codeword 3

 (+++ — + — + — 一 H "一 +++ — ++ — + + —— + — +++ — + _ + + — 一 H" ― +++ — ++ — + + — 一+ —,

+ — +++ + ++++ — ++ + ++++ — +++ + ++++ — +++ + +++) Codeword 4

 (+++ + +++ — ++++ + +++ — +++ + +++ — +++ + +++ — +,

+ — ++ — +++ — + —— + + — ++ — +++ — + —— + ++++++++ — + — One + + — ++ — +++ ~ + — — +) Codeword 5

 (+++-++-++++-++-+ + one one + + _ one +-+++-++-++++-++-+ + one one + + one one

+-+++ — +-+++ + +++-+ ++++-+++ +-+++ + +++-+ +++) Codeword 6

 (+++ +-+++ + +++-++++ +-+++ h +++-+ +++ — +,

+-++-++++-++-+++-+ one one H + _ one + +-++-++++-++-+++-+ one one + +-+) Codeword 7

(+++-++-+ + one one + +-one +-+ + + + + + + + + + + + + + + one one + + one one + + + + + + +- + ₅ H "一" ++ 1 h + H "一" I · 一 "(-+ H 1" 一 "h + H 1 h + H · ―" I h +++-++ H) /: /: / O ε99000ίοοίζ> 1 £ 9SS0S0iAV) + + ++ ++ + + + + ++ ++ ++ +++++++ ++ + + + + ++++ ΙΙΙΙΙΙ

 (++++++ 卞 ++++++ +++ ^ + + Τ ++++ + Τ +++++ + | Ι IIΙΙΙΙΙΙΙΙΙ

 6

 ) + + + + + + + + + _ ++++++ ++++++++++++++++ -I -I ————.--- || ---

(+++++ ++ +++++ ++++ ++ + ++ +++++++ ++++ ΙΙΙΙΙΙΙ) ++ +++++++ + +++++ ++ ++ + IIII

 (+ + + + + T + + + + + + + + + + + + + + + + + +-II II) + + + + + + + + + + + + + + + + + +--III II ---

(+ T ++ ++ +++++ _ + 卞 + +++, I ++ +++++++ +++++++++++ + ++ Τ + ΙΙΙ.

 (++++++ — 1 +++++++++++++ + + ΙΙΙΙΙ ———-

(++ +++++++++ ++ +++++ ++++++ ++ +++ ++ + ΙΙΙΙΙΙΙΙ

 $ Inch ΐ

 } + ++++ +++++++ ++++++ +++ ++++ +++ ++++ ONΙΙΙΙΙ-Ι—- ■

^ + ί ++++++++++ + ++++++ + +++++ ΙΙΙΙ——ΙΙII 1-ΙΙΙ} ++++++++ ++++ ++ + ++++++ + +++ ++ ++++ IIII -II—II—— 1-I-- (+++ + +++-+ +++-++++ +++-++++ +-+++ + +++-+,

+ _ ++ _ +++ — H + + one one + +-++-+++-+ one one + +-++-++++-++-+++-+ one

The basic D0C code group (N = 16, R = 4, = 2) in the D0C code set set B, and the basic period of the transformation sequence Q is (+1 +1-1 +1 +1 +1 + 1 -1) The basic period of the transformation sequence I _t is (+1 -1)

Codeword 1

 (++ — ++++ — ++ — ++++ — ++-++++ — ++ — ++++ — ++ — ++++ — ++ — ++++ — + + — ++++ — ++ — ++++-, + +++ + + + + + + + + + + + + + + + + + + + + + + + + ++ h-++) codeword 2

(++-+ — +++-+ — +++-+ — +++-+ — +++-+ — +++-+ — +++-+ — +++-+ — + ₅

Codeword 3

 (++-++++ — +++++++++ — +++-++++ l · +++-++++ — +

 + — +-++-+++-+ 一一 + — +-++-+++-+ 一一 + — +-++-+++-+ 一一 + — +-++-+ ++-+ one one) codeword 4

 (++-+ + one one + + + + + + + + + + + one one + + + + + + + + + + one one + + + + + + + + + + + one one + + + ++-,

+ + h +++-+++ + ++++-+++ + ++++-+++ + ++++-++) Codeword 5

 (++-++++-++-++++ μ + —— + + + + + + + + + + + + + — + + one + +, H 1 — h + H 1 H h + — I h + H—— I 1 1 h + H 1 l · ^ — l · +-1 h + H— I) code word 6

(++-+ — +++-+ — + 一一 +-+++ — +-+++-++-+ — +++-+ — + 一一 +-+++ — +-+ ++- ₅

-1 1 1 h ++ H HH— — I- + H 1 1 1 H- + H— H— I h + H— h +) Code word 7

(++-++++ h + one one + + + + + + + + + + + + + + + + one + + + + + + + + +- ) + + + + + + + + + + + + + + + + + +--I. iIII III

 (++ +++ + ++++++++++++ ++++ + ++++++++ — II 1-II

 ) ί + — + + + ++ τΐ + +++ — ++! ——- 1

 "++ 4 + + +++ 卞 ++ 卞 + ++ +++ ++ ΐττ + —— 1——I ~ ι!

) — +++ † + + —4 + — +++ Ϊ +++ + + .. II ——— Ι Codeword 16

 (++-+ — + ― +-+++ — +-+++-++-+ — + 一一 +-+++-++-+ — +++-+ — + 一一 + — +++ —, + + — ++++-++-++++-+++ + — ++++-+++ + —— h + — ++++-++) can be verified The relative offset of the correlation function in the above two code groups is at a relative offset equal to k points, where k is a non-zero integer; and the relative position of the correlation function between the two code groups is relative offset. Equal to 8 +4 points, where is a non-zero integer.

 The two basic DOC code groups (where no protection period is added between the first and second subcode parts and the overlap between the first and second subcode parts are counted) are used to scramble using different long PN sequences Code (where the length of the scrambling code segment is N = 16), the scrambling code sequence is randomly generated by the computer. Figure 7 shows the basic DOC code group of the code group set SET A. After scrambling the codeword 1 and codeword 10, the mean square value of the correlation function values at each relative offset point. It can be seen from FIG. 7 that at the position of the sash (at the point ^, where ^: is a non-zero integer) there is a large correlation function between the codeword 1 and the codeword 10; at the position of the non-window, The mean square value of the correlation function is generally small (especially near the origin) and can generally be ignored in actual system design. Note that, compared with the previous example, although the stiles only appear once every 8 points, that is, the number of stiles is half less than in the previous example, the mean square value of the correlation function value at each stile is also the previous one. The corresponding point in the example is about doubled. This also reflects the real shield of the DOC code from one side-the DOC code does not reduce the total interference power at all possible offset points, but concentrates the interference power on the corresponding sash position. '

Figure 8 shows codeword 1 in the basic DOC code group of the code group set SET A and codeword 1 in the basic DOC code group of the code group set SET B. After scrambling with different PN sequences, The mean square value of the correlation function values at each relative offset. It can be seen from FIG. 8 that at the point of the scoring position, where k is a non-zero integer) there is a large correlation function mean value between the two codewords; while at the non-scoring position, the correlation function means square Generally small (especially near the origin), it can generally be ignored in actual system design. (4) Relevant considerations for designing a DOC code group in an actual system:

 The system's spreading factor, the width of a single-sided IFW (or the spacing between window sills), and the number of available DOC codewords are mutually restricted. Assuming that the total length of the codeword is NRM, and the width of the single-sided IFW (also the interval between the stiles) is RM-1, a DOC code group can be constructed according to the method described in the present invention. The codewords contained in the code group The number is N. When designing the system, the above-mentioned factors should be considered to select an appropriate DOC code group. For example, if the spreading factor is 128 and the required single-sided IFW width is 3, a DOC code group can be designed with the parameters (N, R, M) = (32, 2, 2), which is the included codeword The quantity is 32. For some communication occasions, such as: in a cellular (Microcell), the multipath delay is usually small, so as long as the DOC code group is designed so that the width of the unilateral IFW is greater than the maximum multipath time delay, the multiple access interference can be completely eliminated. .

 On the other hand, in some communication occasions (such as in a large area cellular communication system) and some bad propagation conditions, the multipath delay may exceed the single-sided IFW width of the spreading code group. For example, assuming that the chip rate of a CDMA system is 1.28Mcps, the width of each slice is 0.78125 microseconds, and the width of three slices is about 2.344 microseconds. However, in some harsh wireless propagation environments (such as mountainous terrain) ), The maximum multipath delay can reach about 10-20 microseconds. In this case, it is not feasible to eliminate the interference by increasing the width of the IFW, because in this way, the number of available codewords will be greatly reduced, which will cause the system capacity to decrease. At this time, when designing a DOC spread spectrum multiple access code, generally, the number of code words should first meet the basic capacity requirement of the system, and then a compromise is made between a certain spreading factor and the IFW width. As mentioned earlier, since the width of the IFW cannot be very large, the multipath component located just at the window sill may cause greater interference. At this time, these interferences can be eliminated by other methods, such as equalization technology, Pre-RAKE technology, or interference cancellation technology in multi-user detection, and so on. Since it is only necessary to eliminate these multipath interferences at the position of the window slab, but not all multipath interferences, the related interference cancellation technology can be simplified to a certain degree.

(5) DOC code group allocation method during system networking: Because the number of DOC code groups is very large, the system networking has a more flexible way when allocating code groups for each cell / sector. Two possible code group allocation methods are described below.

 Generally, the same basic DOC code group can be assigned to each cell / sector. The code group has certain IFW characteristics and related window position characteristics, which can be used to define different code channels. Each cell / sector is then assigned a long PN sequence as the scrambling code sequence for that cell / sector. Referring to FIG. 9, in which all cells use the same code group set (labeled with SET A), and are composed of the basics in the code group set. The DOC code group and the cell-specific long PN scrambling code (labeled with PNx) Multiple spreading structure to determine the spreading multiple access code used for each symbol. Each cell is distinguished by a long PN scrambling code. The long PN code here can be different offsets of a certain PN sequence (such as the m sequence), or different sequences with good correlation characteristics, such as the Gold sequence. The role of the PN sequence In addition to distinguishing the code groups in each cell, various interference (including ISI, MAI, and ACI) components located at the window are whitened.

 Another code group allocation method is shown in FIG. 10, in which three different code group sets are used (labeled with SET A, SET B, and SET C respectively). The code group sets here are mutually orthogonal. Adjacent cells use different sets of code groups; and each cell uses a cell-specific long PN sequence (marked with PNx) for scrambling. Among them, there may be some IFW relationship between some code group sets (for example, there is an IFW with a one-sided width between code group sets A and B). The orthogonality and even IFW characteristics of the code group between neighboring cells can eliminate some ACI interference to a certain extent.

 As shown in FIG. 11a and FIG. Lib, a spread spectrum multiple access code encoding device according to the present invention includes: a PN sequence generator, a Walsh code generator, a transform sequence generator, a frequency divider, and a correlator; wherein:

 The output data of the PN sequence generator is correlated with the output data of the Walsh code generator and then correlated with the output data of the transformation sequence generator to obtain output data.

The device is characterized in that the PN sequence generator refers to: there can be N PN sequence generators, and the Walsh code generator refers to: there can be N Walsh code generators, The transformation sequence generator refers to: there can be N transformation sequence generators; wherein:

 The output data of the 0th PN sequence generator is correlated with the output data of the 0th Walsh code generator and then correlated with the output data of the 0th transformation sequence generator to obtain the 0th output data; the Nth PN sequence is generated The output data of the generator is correlated with the output data of the N-th Walsh code generator and then correlated with the output data of the N-th transform sequence generator to obtain the N-th output data; the 0-th output data is related to the The Nth output data is correlated to obtain the output data.

 The device of the present invention can generate a DOC (Distinctive Orthogonal Code) spread-spectrum multiple access code. The code words in the code group have the characteristics of zero interference window (IFW), and are only in certain specific positions (referred to as It is called "window" and may have a large non-zero correlation value.

 So far, a coding method and a device for spread spectrum multiple access codes capable of focusing interference in a CDMA system at certain specific positions have been completed. Those of ordinary skill in the art should recognize that the various logic units, modules, circuits, and algorithm steps used to describe the present invention may be electronic hardware, computer software, or a combination thereof. Come to fruition. The various components, units, modules, circuits, and steps are usually described in terms of their functions. Whether hardware or software is used for implementation is determined by the specific application and design constraints of the entire system. Those of ordinary skill in the art should be able to recognize the interchangeability of hardware and software in specific situations, and can use the best way to implement the type of encoding method described in the present invention for specific applications.

For example, various logic units, modules, circuits, and algorithm steps used to describe the present invention may be implemented in the following manners or groups thereof, including: digital signal processor (DSP), special-purpose integrated circuit (ASIC) , Field Programmable Gate Array (FPGA) or other programmable logic device, discrete logic gate (gate) or transistor logic, separate hardware components (such as registers and FIFO), execution of a series of firmware ( firmware) instructions processor, traditional programming software (programmable software) and related processors (rocessor). The processor may be a microprocessor (microprocessor), or Are traditional processors, controllers, microcontrollers, or state machines; software modules can exist in RAM memory, flash memory, ROM memory, EPROM memory, EEPROM memory, Register, hard disk, removable disk, CD-ROM, or any existing known storage medium.

 It is obvious and understood by those skilled in the art that the preferred embodiments of the present invention are only used to illustrate the present invention, and are not intended to limit the present invention. The technical features in the embodiments of the present invention may be arbitrarily combined. Without departing from the idea of the present invention. According to a method for encoding a spread spectrum multiple access code capable of concentrating interference in a CDMA system at certain specific positions and a device for generating the same, there are many ways to modify the disclosed invention, and in addition to the specific In addition to the preferred modes, the present invention may have many other embodiments. Therefore, any method or improvement that can be obtained according to the concept of the present invention should be included in the scope of rights of the present invention.

Claims

Rights request  A method for encoding a spread spectrum multiple access code, which comprises the following steps:  Select any orthogonal sequence group as the initial orthogonal sequence group;  Repeating the elements on each sequence bit in the initial orthogonal sequence group sequentially to obtain another sequence group;  Generate transformation sequence Q;  Multiplying each sequence in the another sequence group with the transformation sequence Q bit by bit to obtain a subcode part of the DOC code;  Generating a transformation sequence I;  Multiplying each subcode in the one subcode portion with the transform sequence I bit by bit, to obtain another subcode portion of the DOC code;  The subcode parts are combined to obtain a DOC code group. 2. The method according to claim 1, wherein the selection of an arbitrary orthogonal sequence group as an initial orthogonal sequence group refers to: arbitrarily selecting a sequence of N orthogonal sequences of length N Group W is the initial orthogonal sequence group; where: W = {Wp W ₂ ,…, D.  3. The method according to claim 1, wherein the repeating the elements of each sequence in the initial orthogonal sequence group in sequence to obtain another sequence group refers to: When the initial orthogonal sequence group is W = {WW ₂ ,… ,, each sequence in W may be wa = i, 2, ~, N) The elements on each bit are sequentially repeated Ri times respectively to obtain N other sequence groups A = {A _P A ₂ ,…, A;  Among them, you can set:

as well as: Α _; · = (aa _{!) 2} … Ά _ίΜ ) (i = 1,2, ~, N) then: (z '= l, 2 ", N; j = 0, l,"', N- 1).  The method according to claim 1, wherein the transformation sequence Q is obtained by repeating the basic period having a length of 1R twice NI, and the length of the obtained transformation sequence Q is NR. 5. The method according to claim 1, wherein each of said sequences in said another sequence group is multiplied with said transformation sequence Q bit by bit to obtain a sub-group of a DOC code. The code part refers to: A group of sequence W containing N orthogonal sequences of length N can be selected arbitrarily as the initial orthogonal sequence group, and W = {W _n W ₂ ,…, W ^} _; Each element of each sequence (i = 1,2, ~, N) is repeated R-1 times in turn to get a set of sequence groups containing N NRs = {, A ₂ ,…, Each sequence in the other sequence group A described above is multiplied with the transformation sequence Q bit by bit to obtain the first subcode portion of the DOC code C ¹ = (C ¹ ;, i = 1,2,- ^ i}, that is: C \ = A _; «QG '= 1,", N)  Among them, the symbol "·" represents the bitwise multiplication of two sequences. 6. The method according to claim 1, wherein the step of generating a transformation sequence I comprises: generating the transformation sequence I after a certain basic cycle is repeated, and the transformation sequence I is Including multiple sequences: ι _Ρ 1 ₂ ,….  7. The method according to claim 1, wherein, each of the subcodes in the one subcode portion is multiplied bit by bit with the transform sequence I to obtain another DOC code. The subcode part refers to: one subcode part may be sequentially multiplied bit by bit with a series of transformation sequences I to obtain other subcode parts of the DOC code, and the transformation sequence I may include multiple sequences: I 1 ₂ ,…;  Wherein, when each subcode in the one subcode part is multiplied bit by bit with a transform sequence having a length of NR, to obtain the other subcode part, it may be set as follows: The one sub-code part is the first sub-code part C ¹ , and the other sub-code part is the second Subcode part C ² ,

Then: the first 1, 2 and the other portions are subcode length sequence of transformations NR ₁ bitwise multiplication, and respectively 3, 4 subcode C ³ and C ⁴ portions, i.e.: C = C * l ₂ ϋ = 1,2, ..., Α0

 The above operations can be continued. Generally, when a subcode part is obtained: C ¹ , C ² ,…, C ^L , where: L = ll, J 'is a non-negative integer, Then CC ² ,…, (^ can be multiplied bit by bit with a transformation sequence of NR 1 _£ , respectively, to obtain another subcode part (^ ⁺¹ , C ^{L +} \…, C ^2L , that is: i = Y, 2 ", N, j = l, 2,-, L) When the above operations are performed a total of times, and the transformation sequences 1 ₂ ,…, 1,… are sequentially used, a total of M = 2 ^M subcode portions CC ² ,…, C ^M can be generated, where each subcode portion contains N A subcode codeword that is NR long. 8. The method according to claim 1, wherein the combination of the subcode parts to obtain a D0C code group means that: M subcode parts can be combined to obtain N total lengths of NR. D0C code group of the codeword {DOCp DOC ₂ ,…, DOC, where: DOQ = (C C, ···, C ^.) G = 1, 2 ", N)  This generates a D0C code group.  9. The method according to claim 4, characterized in that: the basic period of the transformation sequence Q is obtained based on a spanning tree; wherein: Each node in the spanning tree contains a "basic period", and the corresponding transformation sequence Q is obtained by repeating the "basic period"; in the r-th column of the spanning tree = 0,1,2,…) Contains as

"Basic period" and corresponds to each sequence in the initial orthogonal sequence group When each bit of R is repeated R-1 = 2 ''-1 times; assuming that the initial orthogonal sequence group contains N orthogonal sequences of length N, then the transformation sequence Q is 2R "basic period" Obtained after repeating M2 times, the length of the obtained transformation sequence Q is NR A generation method based on a spanning tree that generates the "basic period" of the transformation sequence Q can be described as follows: On the spanning tree, the root period contained in the root node on the 0th column is (+ 1 + 1); The tree can be constructed as follows: Suppose that a node in the r-th column contains a "basic period" P ^Q of length 2 2R. It consists of two subsequences a ^Q and b ^Q of length 2 = ^ , which is: P ° = (a ⁰ , b °) Then in the spanning tree, the two nodes derived from the node and located in the r + 1 column contain two "basic periods" P ¹ and P ^{2 of} length 2 ² = 4, respectively: (a ¹ , b ¹ ) = (a ⁰ , b °, a ⁰ , -b °) P ² = (a ² , b ² ) = (a ⁰ , -b °, a ⁰ , b °)  According to the recurrence relationship, a spanning tree structure of the basic period of the transformation sequence Q can be obtained.  10. The method according to claim 4, characterized in that:  The transformation sequence Q can be obtained by repeating the basic periodic sequence P; Wherein, in the case where each sequence in the initial orthogonal sequence group described by the sub-item is repeated R-1 times bit by bit, it is possible to set P _2Ώ? As a basic period with a length of IfiTR, where is a positive integer, which is defined by The length of the front and back segments is composed of KR subsequences P and P ² ^, and P and each are composed of small sequence segments of length R, that is:  P2O? = {P] AS, ^ KR] and pijo? = { ^C c ² / ?,…, ^c P ² ¾ = { ^{S S2} R, .., ^S Where and s (i = l, 2,-', K are small sequence segments of length R, and the following orthogonal relationship is satisfied between s and s: (c ^) (s ^) ^T = 0 (ί = 1,2-,, Κ)  Wherein, T represents a transposition of a row vector; the sequence having the structure is a complementary sequence; the basic period of the transformed sequence Q must be a complementary sequence.  11. The method according to claim 6, characterized in that: Each transformation sequence I having a length of NR can be obtained by repeating NR / 2 times for a basic period X _{2i having} a length of 2L, here = 1,2,…; the basic period X _2i can be consecutively “+1” tight Consecutive L consecutive "-1", or may be composed of consecutive L "-1" followed by consecutive "+".  12. The method according to claim 1, wherein the DOC code group refers to: a DOC code group having a total length of NRM codewords; and in the DOC code group having a total length of NRM codewords, used The parameters are N, R, and M. Using different parameter combinations can generate a series of DOC code groups containing different numbers of codewords, codeword lengths, and correlation shields.  13. The method according to claim 1, wherein the transformation sequence Q and the transformation sequence I can perform multiple transformations by using various combinations, so that the generated DOC codeword satisfies only There is a non-zero correlation value at the position of the sash, and there is a zero correlation window near the origin; the above process requires at least one transformation sequence Q or transformation sequence I to perform a transformation.  14. The method according to claim 1, wherein, using different sets of the initial orthogonal sequences, and the same transform sequence Q and transform sequence I, a batch of DOC code groups can be constructed, The batch of code groups is said to form a code group set; each said code group set corresponds to a set of transformation sequences Q and I.  15. The method according to claim 1, wherein the M sub-code parts should be transmitted on different mutually orthogonal synchronous fading channels; that is: The M sub-code parts may be sequentially allocated to M consecutive time slots; in order to avoid each sub-slot Overlapping between code parts can add a protection period consisting of all-zero elements at the end of each of them.  16. A method for applying a DOC code based on the method of claim 1, characterized in that the codewords in the DOC code group are scrambled, that is, each codeword in the DOC code group is the same as one Sequences of length are multiplied bit by bit.  17. Only the method described in 16 is required. However, the following steps can be used to sequentially intercept a scrambling code segment of length N from the long PN sequence;  Repeating each bit element in each said scrambling code segment R-1 times in order to obtain a sequence of length NR;  重复 Repeat the entire sequence of length NR M-1 times to obtain a sequence of length N belly, that is, the scrambling code sequence; the scrambling code sequence is composed of the same sequence of M segments, that is, the scrambling code Composed of paragraphs;  The M-segment scrambling code segments in the scrambling code sequence are respectively multiplied by the M sub-code partial codewords in the DOC codeword bit by bit, thereby completing the scrambling process.  18. The method according to claim 17, wherein, in the process of constructing the DOC code, a Walsh orthogonal sequence group can be selected as an initial orthogonal sequence group, which is generated based on the Walsh orthogonal sequence group. The DOC code group can be called a basic DOC code group.  19. The method according to claim 18, wherein in a code group set, all code groups can be regarded as being obtained after one basic DOC code group is scrambled by different scrambling code segments, that is, : After scrambling, the basic DOC code group still retains the IFW characteristics and window position characteristics of the original code group.  20. The method according to claim 18, characterized in that, when spreading operation is performed by using a DOC code and a long scrambling code, multiple spreading modes can be adopted. The method according to claim 20, wherein the multiple spreading means: can be divided into two levels to perform spreading: the first level is to perform basic frequency using a basic DOC code group, wherein each basic The DOC code defines a code channel; the second level is to use the scrambling obtained according to the scrambling steps. The code sequence scrambles the basic DOC code.  22. A method for performing D0C code group allocation in a cellular communication system network based on the method of claim 1, wherein a DOC code group is allocated to each cell or sector.  23. The method according to claim 22, characterized in that: each cell or sector is assigned the same basic DOC code group; at the same time, each cell or sector uses a different long PN code as a scrambling code.  24. The method according to claim 22, wherein: different basic DOC code groups are allocated to adjacent cells or sectors; wherein these basic DOC code groups belong to different code group sets; and Each cell or sector uses a different long PN code as the scrambling code.  25. A spread spectrum multiple access code encoding device, comprising: a PN sequence generator, a Walsh code generator, a transform sequence generator, a frequency divider, and a correlator; wherein:  The output data of the PN sequence generator is correlated with the output data of the Walsh code generator and then correlated with the output data of the transformation sequence generator to obtain output data.  26. The device according to claim 25, wherein the PN sequence generator refers to: there can be N PN sequence generators, and the Walsh code generator refers to: there can be N Walsh codes The generator, the transformation sequence generator refers to: there can be N transformation sequence generators; wherein:  The output data of the 0th PN sequence generator is correlated with the output data of the 0th Walsh code generator and then correlated with the output data of the 0th transformation sequence generator to obtain the 0th output data; 笫 N PN sequence generation The output data of the generator is correlated with the output data of the N-th Walsh code generator and then correlated with the output data of the N-th transform sequence generator to obtain the N-th output data; the 0-th output data is related to the The Nth output data is correlated to obtain the output data.