Summary of the invention
Though provided by the present invention also is a kind of time-division multiplex technology, its purpose is not for multi-path digital communication, and only is in order to improve the spectrum efficiency of system.For multiplex techniques such as conventional time division multiplexing tdm, conventional frequency division multiplexing FDM, orthogonal frequency division multiplex OFDMs, the multiplexing spectrum efficiency of system that can't improve itself, and the present invention leans on the multiplexing spectrum efficiency that increases substantially system.Not only do not need mutual isolation between its each symbol, and can have very strong overlappedly, therefore be referred to as overlapping time division multiplexing (Overlapped Time Division Multiplexing).But is-symbol overlaps not to be regarded as the phase mutual interference among the present invention but to be utilized energetically and forms a kind of new coding bound relation, and overlapping many more, encoding constraint length is long more, and coding gain is high more, and spectrum efficiency is also high more.Under same threshold wanted to interfering signal ratio condition, its available spectrum efficiency will be far above prior aries such as higher-dimension modulation; Otherwise when equal spectrum efficiency, its required thresholding wanted to interfering signal ratio is more much lower than technology such as higher-dimension modulation, and particularly its thresholding wanted to interfering signal ratio improves more in random time-varying channel.This is because the signal code in each time slot of the present invention can be a broadband signal, allows to appear the selectivity decline, so itself can have stronger anti-fading ability.Never the people proposed initiatively to utilize the coding bound that interference caused that occurs between the symbol to concern and improved system spectral efficiency.
One object of the present invention is to provide a kind of time division multiplexing method, through the multiplexing spectrum efficiency of system that makes raising is significantly arranged; And make the required number of levels that can distinguish of system not present the relation of exponential increase with the raising of system spectral efficiency, and just present the relation that algebraically increases, thus reduce requirement greatly to the system linear degree; Described overlapping time, division multiplexing did not have any specific (special) requirements to the transfer function of system etc., thereby avoided in system, using complicated technology such as adaptive channel equalizer; With other compared with techniques, in same spectrum efficiency, equal condition of work following overlapping time of division multiplexing has lower thresholding wanted to interfering signal ratio, thereby practices thrift transmitting power or increase service radius; Make it to the shape, bandwidth etc. of multiplexing signal spectrum do not do any specific (special) requirements; When in random time-varying channel, working especially; Owing to can adopt the signal spectrum that is re-used (rate of delivering a letter that comprises the raising system) of broad; This moment, the change at random of channel can produce the implied diversity effect automatically, improved the transmission reliability of system.The signal spectrum that is re-used is wide more, and diversity gain is high more, and transmission reliability is also high more.
Another object of the present invention is to provide a kind of time division multiplexing system, have significantly through the multiplexing spectrum efficiency that makes system and improve; And make the required number of levels that can distinguish of system not present the relation of exponential increase with the raising of system spectral efficiency, and just present the relation that algebraically increases, thus reduce requirement greatly to the system linear degree; Described overlapping time, division multiplexing did not have any specific (special) requirements to the transfer function of system etc., thereby avoided in system, using complicated technology such as adaptive channel equalizer; With other compared with techniques, in same spectrum efficiency, equal condition of work following overlapping time of division multiplexing has lower thresholding wanted to interfering signal ratio, thereby practices thrift transmitting power or increase service radius; Make it to the shape, bandwidth etc. of multiplexing signal spectrum do not do any specific (special) requirements; When in random time-varying channel, working especially; Owing to can adopt the signal spectrum that is re-used (rate of delivering a letter that comprises the raising system) of broad; This moment, the change at random of channel can produce the implied diversity effect automatically, improved the transmission reliability of system.The signal spectrum that is re-used is wide more, and diversity gain is high more, and transmission reliability is also high more.
The invention provides a kind of time division multiplexing method, utilize a plurality of symbols in time-domain parallel transmission data sequence, this method may further comprise the steps: transmitting terminal forms overlapped the transmitting on time-domain of a plurality of symbols; Receiving terminal carries out detecting by data sequence in the time-domain to received signal according to the one-to-one relationship between transmission data sequence and the transmission data sequence time waveform.
Transmitting terminal forms overlapped the transmitting on time-domain of a plurality of symbols according to design parameter.
Determine described design parameter according to given in advance channel parameter and system parameters.Said design parameter comprises that basic modulation level counts M, basic symbol length
Symbol lengths T
s, mark space Δ T
s, the overlapping tuple K of symbol and frame length T.
The overlapping tuple K of said symbol, mark space Δ T
sAnd symbol lengths T
sBetween satisfy following relation: (K-1) Δ T
s<T
s≤K Δ T
s
Said symbol lengths
wherein
is a basic symbol length, and Δ is the maximum time diffusing capacity of channel.
Said basic symbol length
wherein Δ is the maximum time diffusing capacity of channel.
Said basic symbol length
is equal to or less than the maximum time diffusing capacity Δ of channel.
Described mark space Δ T
sLess than the coherence time of channel
Frame size
where
is the channel coherence time.
Through reducing mark space Δ T
sIncrease the overlapping tuple K of symbol.
Said channel parameters include the maximum time spread of the channel or channel volume Δ coherence bandwidth
and the maximum frequency of the channel diffusion capacity
or channel coherence time
Described system parameters comprises system bandwidth B, to the requirement and the linearity of spectrum efficiency.
Through increasing system bandwidth B, or utilize and interweave and Methods for Coding, or the raising system rate of delivering a letter, or the method for spread signal frequency spectrum improves the latent frequency diversity tuple of system.
Described transmitting terminal forms overlapped the transmitting on time-domain of a plurality of symbols and may further comprise the steps: the phase wiggles and the orthogonal waveforms that form l=0 road modulation signal envelope waveform with digital form; Described phase wiggles and orthogonal waveforms behind time shift, are formed the homophase envelope waveform and the quadrature envelope waveform of other each modulation signal; With the homophase envelope waveform of said each modulation signal and quadrature envelope waveform in-phase data signal and orthogonal data signal multiplication with each corresponding symbol, obtain each modulation signal through the filtered modulation signal waveform of data-modulated; With described each modulation signal waveform addition, formation transmits.
Described receiving terminal carries out detecting by data sequence in the time-domain to received signal and may further comprise the steps: the reception signal in each frame is formed the receiving digital signals sequence according to the one-to-one relationship between transmission data sequence and the transmission data sequence time waveform; Described receiving digital signals sequence implementation sequence is detected, to obtain the judgement that is modulated at the modulating data on whole symbols in the said frame length.
Saidly reception signal in each frame is formed the receiving digital signals sequence may further comprise the steps: it is synchronous to form symbol time at receiving terminal to received signal; According to sampling theorem, the reception signal in each frame is carried out digitized processing.
Described digitized processing is carried out at intermediate frequency, or carries out in base band.
Described Sequence Detection is a Maximum likelihood sequence detection when each sequence equiprobability, when each sequence unequal probability, is the maximum a posteriori probability Sequence Detection.
Described described receiving digital signals sequence implementation sequence is detected may further comprise the steps: the plural convolutional encoding model of making overlapping time division multiplex system; List whole states of overlapping time division multiplex system; Make the trellis structure of overlapping time division multiplex system, and list the coding output of each branch road; For each stable state is prepared two memories; In said trellis structure, search out the path with receiving digital signals sequence minimum Eustachian distance or weighting minimum Eustachian distance, the corresponding data sequence in this path is conclusive judgement output.
Described plural convolutional encoding model of making overlapping time division multiplex system may further comprise the steps: actual channel is measured, found out the valuation that receives complex envelope in the distinct symbols time interval; Utilize the valuation of described reception complex envelope to form the tap coefficient in the overlapping time division multiplex system channel model.
Described actual channel is measured, found out the valuation that receives complex envelope in the distinct symbols time interval and use special pilot signal measurement; Or utilize the information of having adjudicated through the mode of received signal computing is calculated valuation; Or get above the two combination; Or find the solution valuation with blind estimating method.
Whole states of said overlapping time division multiplex system comprise initial condition, preceding transition state, stable state, back transition state and end-state.
The described surviving path memory of preparing in two memories for each stable state is used to store the surviving path that arrives said state; Euclidean distance memory or weighted euclidean distance memory are used to store the surviving path of the said state of arrival and the Euclidean distance or the weighted euclidean distance of receiving digital signals sequence.
Described transition state can be used the memory of arbitrary stable state.
The described path that in said trellis structure, searches out with receiving digital signals sequence minimum Eustachian distance or weighting minimum Eustachian distance may further comprise the steps: step 1, the path Euclidean distance that makes start node (l=0) state or weights Euclidean distance are zero; Step 2, calculate l (l=1 ..., L-K+1) all the state S in the individual node calculate its all branch road code signal and branch road Euclidean distance or weighted euclidean distance receiving digital signals of each bar from last state transitions to said state S; Step 3, branch road Euclidean distance or branch road weighted euclidean distance and they that will arrive this state to each state S set out separately the path Euclidean distance or the addition of weights Euclidean distance of state form new one or more paths Euclidean distance or weights Euclidean distance; If a plurality of described path Euclidean distances or weights Euclidean distance are arranged; Then from wherein selecting reckling; As path Euclidean distance or the weights Euclidean distance of l node state S, upgrade and deposit in Euclidean distance memory or the weighted euclidean distance memory of said state S.Step 4, at node l, each state S is found out its path Euclidean distance or the pairing surviving path of weights Euclidean distance, upgrade the surviving path memory deposit this state S in; Step 5, next node is repeated above-mentioned steps one to step 4, until the L+K-2 node, this moment, surviving path was only remaining one, and then the pairing data sequence of this surviving path is conclusive judgement output.
In calculating, check the surviving path memory of each state at any time, in case in the path of finding to keep identical initial part is arranged, then that this is identical initial part is as judgement output.
Still not judgement is then adjudicated by force when described surviving path memory is filled with, and promptly exports the initial bit that minimum range is arranged as judgement.
Still the majority logic judgement is then carried out in not judgement when described surviving path memory is filled with, and promptly exports the majority in each surviving path initial bit as judgement.
Described path Euclidean distance memory or weights Euclidean distance memory are only stored relative distance; Even path Euclidean distance or weights Euclidean distance minimum or the maximum are zero distance; The path Euclidean distance memory of other each state or weights Euclidean distance memory are only stored relative Euclidean distance or relative weighted euclidean distance, and said relative Euclidean distance or relative weighted euclidean distance are meant the difference with said path Euclidean distance or weights Euclidean distance minimum or the maximum.
Described Sequence Detection is a Maximum likelihood sequence detection.
The present invention also provides a kind of time division multiplexing system, and said system comprises transmitter and receiver; Wherein transmitter comprises: overlapping time division multiplexing modulating unit is used to form overlapped the transmitting on time-domain of a plurality of symbols; Transmitter unit is used for said transmitting is transmitted into receiver; Receiver comprises: receiving element is used to receive the signal that said transmitter unit is launched; The Sequence Detection unit is used for carrying out to received signal detecting by data sequence in the time-domain.
Described overlapping time division multiplexing modulating unit comprises: digital waveform generator is used for phase wiggles and orthogonal waveforms with first modulation signal envelope waveform of digital form formation; Shift register, the phase wiggles and the orthogonal waveforms of first modulation signal envelope waveform that is used for digital waveform generator is produced carry out time shift, with homophase envelope waveform and the quadrature envelope waveform that produces other each modulation signal; Serial-parallel converter is used for the data sequence of serial input is converted into the parallel in-phase data signal and the orthogonal data signal of corresponding each modulation signal; Multiplier; The homophase envelope waveform and the quadrature envelope waveform that are used for the modulation signal that each in-phase data signal of said serial-parallel converter output and orthogonal data signal is corresponding with each multiply each other, with obtain each modulation signal through the filtered modulation signal waveform of data-modulated; Adder, be used for each modulation signal of multiplier output through the filtered modulation waveform addition of data-modulated, transmit with formation.
Said transmitter can also comprise spectrum-spreading unit, is used for increasing in case of necessity overall system bandwidth.
Said transmitter can also comprise interleave unit and coding unit, is used for improving in case of necessity the latent frequency diversity or the latent time diversity tuple of system, to improve the transmission reliability of system.
Said receiver also comprises pretreatment unit, is used to form synchronous receiving digital signals sequence complete in each frame.
Said pretreatment unit comprises: lock unit, and it is synchronous to be used in receiver, forming symbol time to received signal; Pilot cell is used for channel parameter is measured; The digitized processing unit is used for the reception signal in each frame is carried out digitized processing.
Described Sequence Detection unit comprises: the analytic unit memory, and be used to make the plural convolutional encoding model and the trellis structure of overlapping time division multiplex system, and list whole states of overlapping time division multiplex system, and storage; Comparator is used for the trellis structure according to the analytic unit memory, searches out the path with receiving digital signals minimum Eustachian distance or weighting minimum Eustachian distance; The surviving path memory of stable state S; Be used to store Euclidean distance memory or the weighted euclidean distance memory of the surviving path stable state S that arrives said stable state S, be used to store the surviving path of the said stable state S of arrival and the Euclidean distance or the weighted euclidean distance of receiving digital signals sequence; Wherein, stable state S is any in said whole stable state.
Each state has a surviving path memory and Euclidean distance memory or weighted euclidean distance memory, and transition state can be used the memory of arbitrary stable state.
Described surviving path memory length be 4 * K (4K) to 5 * K (5K), wherein K is overlapping tuple.
Described surviving path memory length is shorter or longer than 5K than 4K, and wherein, K is overlapping tuple.
Described Euclidean distance memory or weighted euclidean distance memory are only stored relative distance.
Beneficial effect of the present invention is; Provide a kind of novel time division multiplexing that utilizes to increase substantially the brand-new learning concept and the correlation technique of system spectral efficiency; It need not do any preliminary treatment to the transfer function of channel; Not only can not reduce channel capacity, the actual capacity that can make system on the contrary is the channel capacity of approximation theory more.With in short, the invention provides one and increase substantially the communication system spectrum efficiency, effective, reliable, practical, brand-new time-division multiplex technology.
Embodiment
To be described with reference to the accompanying drawings basic principle of the present invention, mathematical description, Maximum likelihood sequence (MLSD) detection below (because the maximum a posteriori probability detection only is that the sequence of different prior probabilities is implemented different weightings; Do not have essential distinction with MLSD, repeat no more) and embodiment.
Basic principle of the present invention at first is described.
For the sake of simplicity, in discussion subsequently, do not comprise space channel.Each signal code that well-known traditional Time Division Multiplexing requires to be re-used is isolated in time-domain fully each other, makes to have no between them to disturb, and each signal that is re-used can use any communication and modulation system independently.
Obviously; If make the signal code that is re-used in time-domain, present mutual covering and overlapping; Just can further improve the spectrum efficiency of system, but people generally believe that this will cause the adjacent serious phase mutual interference of generation between the signal code that is re-used; As shown in Figure 2, there are three symbols in time-domain, to overlap among the figure.Because symbol overlapped, any symbol is wherein carried out the serious interference that demodulation all will receive other adjacent-symbols by conventional method, it is absolutely not therefore carrying out right demodulation by conventional method.But scrutinize Fig. 2 again, suppose that the symbol intervals of three be re-used signal A, B, C among the figure is T
sSecond, the interval between them, promptly the time shift amount does relatively
Second, that is to say that the symbol of these three signals overlaps fully.For the sake of simplicity, suppose that the shape of three signal A that are re-used, B, C symbol is in full accord, phase characteristic is zero entirely, and modulation system all adopts the positive counter modulation of binary, and symbol lengths is T
sSecond, each road modulation signal bandwidth is B
0Conspicuous, the bandwidth after their stacks also is B
0Conspicuous, but the symbol time duration elongatedly do
, three signals are synchronous fully.Because their symbol overlaps, and all will receive the interference of adjacent other symbols, by conventional method, right demodulation is absolutely not.But the present invention does not handle each symbol not isolatedly, but with the unified consideration of these three symbols, at this moment situation is just different fully, because at certain 5T
sIn/3 time periods, the data that symbol A, B, C three are transmitted are nothing but one of eight kinds of situation shown in the table 1:
Table 1
And corresponding reception signal is respectively eight kinds of D, E, F, G, H, I, J, K among Fig. 2 when disregarding noise, and its data and waveform are one to one fully.Equally, for any other overlapping tuple, can check, on mathematics, also can prove fully simultaneously, its data combination (sequence) also necessarily has one-to-one relationship with waveform.In the ordinary course of things, if the symbol intervals of each road signal that is re-used is T
sSecond, T here
sShould take into account the All Time broadening factor (like system's timing wander, multipath broadening etc.) in the system, when making overlapping time division multiplexing, the time shift amount is that each mark space is Δ T each other
sSecond, and satisfy
(K-1)ΔT
s<T
s≤KΔT
s;K=1,2,...
Promptly have K adjacent-symbol together overlapped, that each symbol transmitted all is M=2
QMetamessage, the i.e. equal load Q=log of each symbol
2The information of M bit, if L this overlapped symbol arranged in the system, it finally is 2 that the data that then transmitted possibly make up
QL=M
LKind, also have 2 one to one with it
QL=M
LPlant time waveform.Therefore to belong to any waveform just passable as long as find out the data combination that is transmitted at receiving terminal.Though intersymbol time-interleaving has destroyed the waveform of single data symbol itself; Destroy the one-to-one relationship between single data symbol and its time waveform, but do not destroyed the one-to-one relationship between data symbol sequence and its time waveform.The most important theories basis of institute of the present invention foundation that Here it is.Certainly work as M
LWhen very big, the complexity that how to reduce system is exactly a very important practical problem.The present invention will provide the algorithm of an optimum to address the above problem, and its complexity only is decided by M
KRather than M
L
Be illustrated in figure 3 as ordinary circumstance.For the sake of simplicity, the spectrum width of setting each signal code that is re-used is B
0Conspicuous; Symbol intervals is T
sSecond; Modulation level is counted M=2
Q, i.e. each symbol load Q bit; As comparing, when not adopting overlapping time division multiplexing, its bit rate is: Q/T
sBps; Spectrum efficiency is:
Bps/conspicuous.The bandwidth of system still is B after adopting L overlapping symbol
0Conspicuous, but total symbol lengths but becomes T
s+ (L-1) Δ T
sSecond, L symbol is total to load LQ bit, so always deliver a letter being increased to simultaneously
Bps, so its spectrum efficiency is:
Can find out, overlapped total symbolic number L>>K, when promptly L is enough big, the spectrum efficiency of system will improve K (symbolic number that overlaps simultaneously) times, K is big more, the spectrum efficiency of system is also high more.We find the increase along with the symbolic number K that overlaps simultaneously, the proportional raising of the spectrum efficiency of system, but the level number of system do not resemble and is index law the higher-dimension modulation technique and increases, and increases and just be the algebraically rule.For example, work as Q=1, when promptly each subcarrier adopted binary modulation, the level number of the heavy symbol overlapping system of K was K+1, only is linear growth with K, works as Q=2, and promptly each subcarrier adopts M=2
2During=4 yuan of modulation, the level number of the heavy symbol overlapping system of K homophase I channel is K+1, and the level number of quadrature Q channel also is K+1, and the overall level number of system is (K+1)
2, also just being square-law with K increases.Obviously when differentiable level number one timing of channel, adopt the spectrum efficiency of overlapping time division multiplex system will be higher than higher-dimension (many level) transmission system.For example if it can adopt the 64QAM modulation under high-speed mobile condition to certain wireless communication system; Its level number M=64; The level number of homophase I and quadrature Q channel is
and the overlapping tuple K=7 of the overlapping time-division system of QPSK of same I, Q channel level number is arranged; But its each symbol load 14 bits; And the every symbol of 64QAM system load 6 bits only; Its spectrum efficiency has only overlapping time-division system
in general; If original system can be supported the M-QAM modulation; Then under the same level number; Adopt the overlapping tuple
of the overlapping time-division system of QPSK modulation, its spectrum efficiency than M-QAM modulating system exceeds
doubly.
After having introduced basic principle, carrying out overlapping time-multiplexed mathematical description below.
Being without loss of generality, supposing that information source is that equiprobability is memoryless, is T after its symbol duration channel transmission
sSecond, the message of being transmitted is transmitted in time-domain with parallel mode, total L the overlapped symbol of every frame in the system, and each symbol is all occupied B behind modulated filter and channel broadening
0Conspicuous bandwidth, for the sake of simplicity, suppose the modulation system and the filter complex envelope characteristic of each symbol in full accord, at basic symbol width T
sThere is K symbol overlapped in second.
The complex data sequence that it is transmitted is:
In the formula:
l=0; 1; 2 ..., L-1;
I
l, Q
lBe at t ∈ [lT
s, (l+1) T
s], the data message level symbol that promptly its homophase I and quadrature Q channel are transmitted in l symbol time interval.
Its complex envelope that sends signal (is not promptly remembered complex carrier signal frequency expj2 π f
oT) be:
In the formula:
Be that complex modulated signal envelope, its complex frequency spectrum are sent in normalization
Bandwidth be B
0
f
0Be carrier frequency: f
0>B
0/ 2, while f
0T
s>>1 or be positive integer;
Δ T
sBe relative time shift amount (intersymbol is at interval), it satisfies:
(K-1)ΔT
s<T
s≤KΔT
s;
Symbol duration T
sIn should comprise the time explanation factor of channel;
E
0Be every symbol transmit signal energy.
And the occupied bandwidth of system still is B
0But, frame length, promptly overlapping total symbol lengths really is:
T=T
s+(L-1)ΔT
s
L+K-1 rather than L Δ T are arranged in frame length T
s
We will handle together the influence that is caused by channel time diffusion (multipath broadening) in the present invention, suppose that the symbol duration after the diffusion of channel time is Ts second, and then multiple connection is collected mail and number is:
In the formula:
Be the complex envelope of white Gauss noise, its power spectral density is N
0Watt/conspicuous;
E
sBe the receiving symbol energy, E
s=α E
0, α is a fading channel;
Duration be [l Δ T
s, l Δ T
s+ T
s], promptly
Be illustrated in figure 4 as in the overlapping time division multiplex system of K=3 and receive signal schematic representation.
Because channel possibly be a random time-varying channel; So for different l; The complex envelope
that receives signal maybe be different, so represent this possibility with subscript l here.If channel becomes when being nonrandom, though perhaps be that the slowly variation that becomes at random the time causes at a frame length T and can think that with interior the characteristic of channel is constant basically, then subscript l all can omit in these cases.But generally speaking, the complex envelope of reception signal is not necessarily consistent with the complex envelope that sends signal.Can find out, owing to symbol-interference only appears in the adjacent K symbol, generally speaking, except initial and final (K-1) individual Δ T
sIn addition, other any moment being received signal all is stacks of K symbol.Particularly at t ∈ [l Δ T
s, (l+1) Δ T
s], l=0,1 ...., multiple connection during L-K+1 is collected mail and number is:
Wherein:
● l (t)=● (t) * [u (t-l Δ T
s)-u (t-(l+1) Δ T
s)] l=0,1,2 ..., L+K-1; Wherein, ● be meant that other meet the letter of this formula expression formula.
Wherein:
l=0,1,2,...,L-K+1;k=0,1,...,K-1;(7)
U (t) is the time-domain unit step function.
Before the span of l was different from here, it had gone out K-1 than L is big, and this is because in the frame length T of system, Δ T
sNumber go out K-1 than L is big, but should be noted that when l>L-1
Simultaneously
Be illustrated in figure 5 as overlapping time division multiplex system the time become plural convolutional encoding model.
Remaining problem is used Maximum likelihood sequence detection MLSD algorithm provided by the present invention exactly, finds the solution in t ∈ [0, T] T=(L+K-1) Δ T
s, promptly in a frame time, make following formula minimal data sequence u in T:
In the formula: || ● ||
2The expression ● square mould.
(9) physical meaning of formula be t ∈ [0, (L-1) Δ T
s+ T
s] promptly in a frame time length, seek most probable data sequence U, make its time corresponding waveform
With the reception signal waveform
Near (Euclidean distance minimum), we will introduce best Sequence Detection MLSD algorithm specially in the back the most, and we will be subsequently provide other quick accurate optimal algorithm with the form of other patent.
Then; Analyze the tap coefficient in the shift register channel model of overlapping time division multiplex system: well-known; In random time-varying channel; The impulse Response Function of channel
along with observing time t difference and present change at random, the shape of complex envelope behind time shift that therefore receives signal generally is vicissitudinous (Fig. 6).At t ∈ (l Δ T
s, (l+1) Δ T
s) in the value signal of each channel tap coefficient see shown in Figure 6ly, particularly be longer than symbol lengths, promptly when the coherence time of channel
The time, these tap coefficients will tighten and be some sample values (numerical value), and this situation can bring great convenience to practical applications.
Explain again that then the tree graph of overlapping time division multiplex system representes:
The tree graph of overlapping time division multiplex system representes it is the mode that a kind of very vivid overlapping time division multiplex system input-output of expression concerns.Fig. 7 be exactly the Q=1 of a K=3 be the input-output graph of a relation of the overlapping time division multiplex system of binary, we represent input bit u with branch upwards among the figure
n=1, downward branch is then represented input bit u
n=-1, corresponding codes output then is illustrated in the top of each branch.Among the figure thick line path representation list entries be u=[1 ,-1 ,-1,1 ... ]
T, corresponding plural convolutional encoding output waveform then does
....Scrutinize this figure, we can find that between input and the output sequence be one to one fully.There is not certain list entries corresponding extremely with two or more output sequences, on the contrary also true.So one-to-one relationship between the input and output sequence in the overlapping not time to rupture of the symbol territory.So if in time-domain, detect just the error probability that can not appearance can not subtract again by sequence, traditional certainly just must have been abandoned by quilt by the symbol detection mode.If the length of sequence is fixedly the time, for example its length is L, then Q is tieed up binary source, and possible sequence sum will be 2
QL=M
LSo our problem will be summed up as 2
QL=M
LUnit's problem of signal detection.Because each sequence equiprobability of supposition occurs usually in communication; Therefore should adopt the Maximum Likelihood Detection criterion; When each sequence unequal probability occurs, should adopt maximum posteriori criterion (promptly the big sequence of probability of occurrence multiply by bigger weighting, otherwise then smaller).So, seem the optimum signal of overlapping time division multiplex system to detect problem and solved, the way it goes theoretically, has any problem but implement.Because L is bigger usually, so direct utilization maximum likelihood or maximum posteriori criterion will be very complicated.People are when the list entries equiprobability fortunately; Decoding algorithm to convolution code; After deliberation decades; And overlapping time division multiplex system also can be counted as a plural convolution coder; Many decoding algorithms of therefore relevant convolution code are for example said Fano algorithm and various storehouse (Stack) algorithm and the bcjr algorithm of in tree graph, searching for best (being the maximum likelihood function value) path, can use for reference the detection that is applied to signal in the overlapping time division multiplex system basically through after putting into the melting pot.Because these algorithms are not real best maximum likelihood algorithm, can only be referred to as the maximum likelihood algorithm that is as the criterion, the present invention does not just plan to have introduced.Below we will introduce another kind of algorithm---Maximum likelihood sequence algorithm (MLSD).This is to use for reference Viterbi (Viterbi) algorithm of convolution code and the real maximum likelihood algorithm that comes, need introduce the Trellis figure and state (State) figure of overlapping time division multiplex system earlier for this reason.
Then, the Trellis figure that overlapping time division multiplex system is described again representes with state diagram:
Though tree graph can vividly describe the relation between input and the output very much, this figure at first is bad picture, and particularly along with the increase of L, it will be index expands, and is not easy to use very much, and therefore is necessary it simplification.Let us returns Fig. 7 again, after examining, can find, this tree graph just becomes repetition after the 3rd, because every branch that gives off from the node that is labeled as a all has same output, this conclusion is to node b, and c, d are correct too.They are nothing more than being following several kinds of possibility (see figure 8)s.As can be seen from the figure can only transfer to (through input+1) node a and (through input-1) node b from node a, b can only arrive (input+1) c and (input-1) d simultaneously, and c can only arrive (input+1) a and (input-1) b, and d can only arrive (input+1) c and (input-1) d.The reason that produces this phenomenon is very simple, because have only the individual symbol of adjacent K (is 3 specific to this example) just can form the phase mutual interference.So when the K bit data was input to channel, the 1st bit data had the earliest shifted out a rightmost shift unit.Therefore the output of channel only is decided by the input of preceding K-1 data except depending on the input of data now.In general, for M=2
Q, i.e. Q dimension binary data input, as long as preceding K-1 Q dimension binary data is identical, their corresponding output is just identical.Therefore among Fig. 7 (Q=1) behind the 3rd branch road, the node of every a of being labeled as just may be incorporated in together, same b, each node of c and d also may be incorporated in together, has so just formed a folding tree graph---Trellis schemes.Translator of Chinese has the people to be referred to as hedge figure or trellis structure (seeing Fig. 9 A and Fig. 9 B).Among the figure our regulation be input as+1 branch road representes with solid line, be input as-1 branch road and be represented by dotted lines.This is that downward branch has been imported for-1 because we can not stipulate simply that again branch upwards is+1 input after folding.
Scheme the repetitive structure on time shaft if get rid of Trellis, we can obtain a figure-state diagram (State Diagram) of further simplifying, and are shown in figure 10, for the sake of simplicity, draw the final back that reaches to transition state among the figure.State in the state diagram is to determine according to each node among the Trellis figure, and promptly each state is to be determined by the preceding K-1 position Q dimension binary data bit of remembering at channel.Therefore be the overlapping time division multiplex system of K to memory (constraint) length, its stable state number is 2 when binary is imported
K-1Individual, when the input of Q dimension binary, be 2
Q (K-1)=M
K-1Individual.Initial in addition, final, preceding in addition transition state and back transition state, initial in this example and end-state is (0,0); Preceding transition state is (0 ,-1) and (0,1) totally two; Back transition state is (1,0), and (1,0) also totally two.State transitions relation initial and transition state is very simple; As long as attention is initial and end-state must be complete zero dummy status; And in preceding transition state, the former data that are stored in the channel contain just has Q dimension binary in odd, the new data+or-data; So they can only be come by a state transitions, and can to other 2
Q=M state transitions; For the back transition state, opposite with preceding transition state, former be stored in the channel be Q dimension binary+or-data, just odd in the new input, therefore, they all can be by 2
Q=M state transitions come, but can only be to a state transitions.Please note: in this example, Q=1, K=3, we are writing state a (1; 1), b (1 ,-1), c (1; 1), the relation of each information bit is arranged by routine from left to right during d (1 ,-1), and for example state b (1;-1) bit 1 is to get into channel the earliest in, but the bit that gets into channel in the Q=1 channel model of Fig. 6 the earliest but exists in shift unit of rightmost, and its time relation is from right to left.This point please the reader must not be obscured.
We can find out that overlapping time division multiplex system is a finite state machine, and its oriented state diagram can be described the input/output relation of channel fully.Because preceding K-1 the Q dimension binary information bit that each STA representation is stored by channel, i.e. (K-1) Q information bit, the transfer branch road between state is then represented the information bit that import now.K=3 for example; The data bit of the binary channel input of Q=1 does ... ,-1,1; 1 ..., then be in state diagram to a state transitions from the c state; Because of c=(1,1), import one 1 again after; Originally existed-1 in shift unit of rightmost to shift out channel, and 1 entering channel of new input, state transitions is to a=(1; 1), the output of channel then does ..., it is indicated on
on the transfer branch road from c to a.
In general memory (constraint) length is that the Q of K ties up the binary input channel and has 2
Q (K-1)=M
K-1Kind of stable state, each stable state can to other 2
QIndividual state transitions, also can from other 2
QIndividual stable state shifts.In Trellis figure above-mentioned conclusion then report for: memory span is that the Q dimension binary input channel of K has 2
Q (K-1)=M
K-1Plant different nodes, under stable case, each node outwards sends 2
Q=M branch road has 2 again simultaneously
Q=M branch road is incorporated in this node.
Trellis figure is very useful when research Maximum likelihood sequence MLSD algorithm.
After having introduced overlapping time-multiplexed mathematical description, introduce Maximum likelihood sequence detection below, i.e. the MLSD algorithm.
Maximum likelihood sequence decoding algorithm in the convolution code can come the signal of overlapping time division multiplex system is detected through transforming to transplant.Below we still are that example is specifically introduced the MLSD algorithm with the binary signal.We know for length is the Q dimension binary list entries of L, and its possible output sequence (possible path in Trellis or the state diagram) number is 2
QL=M
L,, directly use Maximum Likelihood Detection and will become very complicated because L is very big usually.The essence of MLSD algorithm is maximum likelihood algorithm, but its complexity only is exponential increase with the memory span K-1 of channel, rather than is exponential increase with L.We suppose that the noise of channel is a white noise for this reason; And the input data sequence that in the white noise channel, has the maximum likelihood function value should be in tree graph or Trellis figure with receive signal the pairing list entries in path of minimum Eustachian distance arranged; Promptly select best
, make and satisfy
In the formula: T all receives signal time;
But because each path periodically merges, we there is no need to calculate the likelihood function or the Euclidean distance of entire path length fully in Trellis figure.Because when the path merges, those paths that before merging, have relatively large Euclidean distance can be removed fully.T=3 Δ T among Fig. 9 for example
sThe time have two paths to overlap for the first time at node a place, they are respectively:
(corresponding list entries is 1; 1,1)
(corresponding list entries is-1 to reach
; 1,1).
We calculate this two paths respectively and receive the Euclidean distance between signal, stay a relative smaller of distance, and we are referred to as surviving path (Survivor Path), and another distance relatively large person then remove.Therefore we write down the surviving path that arrives it earlier to node a; For example say
and also have two paths to overlap for the first time to node b equally with the Euclidean distance
that receives between signal; They are respectively:
(corresponding list entries is 1; 1;-1) (corresponding list entries is-1 to reach
; 1 ,-1).We select one and receive the path that relative minimum range is arranged between signal; And write down this surviving path; For example say
and and the Euclidean distance
that receives between signal node c and d are also done same processing; The result sees Figure 11; Wherein, * path and do not draw the path representation path that is eliminated.Surviving path among the figure all is relative minimum distance path.So we have obtained arrival node a, b, c, the relative optimal path of d and corresponding Euclidean distance:
At the also bad work of our any judgement of this stage.At t=4 Δ T
sThe time, we calculate equally respectively and arrive different paths of each node and the Euclidean distance that receives between signal, and the path of selection with relative minimum range.For example to node a, at t=4 Δ T
sThe time arrive a the path in original Trellis figure, have four, promptly 1,1,1,1; 1 ,-1,1,1;-1,1,1,1;-1 ,-1,1,1.But in the calculating of phase I, first three branch road of second and the 3rd paths is eliminated, so we can only make one's options in first and the 4th liang of paths.Need calculate them for this reason respectively and receive the Euclidean distance between signal.Notice that we need not calculate the Euclidean distance of entire path now, because noise is a white noise, we only need to calculate at t=3 Δ T
sThe time node a to t=4 Δ T
sThe time a between branch road and receive the Euclidean distance between signal, add
It is the Euclidean distance between article one path and reception signal.Equally, we only need to calculate at t=3 Δ T
sThe time node c to t=4 Δ T
sThe time node a between branch road and the Euclidean distance that receives between signal add
It is the Euclidean distance between the 4th paths and reception signal.Between this two paths; We eliminate the relatively large person of distance, and write down less relatively path of distance and Euclidean distance
and
certain
thereof and just can from memory, dispose with
.We also carry out same processing to node b, c and d.So continue; At each the stage l that calculates, to l (l=0,1; 2; ..., L-K+1) nodes in stage separately the state of representative (be Trellis figure in l each node in the stage) we only keep one with the less relatively path of reception signal Euclidean distance, and write down this path Euclidean distance and pairing path.
Figure 11 is a sketch map of this testing process, please notes in our example, at the five-stage that calculates, i.e. t=7 Δ T
sThe time, each surviving path (being relative optimal path) is respectively:
at this moment the initial part in best relatively path is-1 entirely;-1,1.Therefore we can enter a judgement:
Because the initial part of all relative optimal paths all is them, they are optimal path just naturally.
If each surviving path does not have common initial part, then calculate and to continue, till they have had common part.Therefore the judgement of MLSD algorithm is at random; It might not adjudicate output for a long time, and judgement output also is not necessarily by symbol, possibly once have only a judgement output; A plurality of judgement outputs also might once be arranged, but maximum judgement time-delay is length L+K-1 of Trellis figure.This is because for L symbol arranged, the system that an adjacent K symbol is overlapped is because its Trellis schemes the longest L+K-1 that has only; And its end-state is complete zero (0; 0 ...., 0); Final each paths must merge, so the time-delay of Maximum likelihood sequence detection MLSD algorithm is at most the L+K-1 step.
Because this characteristics of MLSD algorithm, we can produce following two queries very naturally:
First, because the MLSD algorithm is to have common initial part that judgement output is just arranged with each surviving path, promptly judgement will be through the time-delay of one section randomness.Is when L → ∞, the judgement time-delay much for the probability of ∞ so?
Second, MLSD algorithm require each state that two memories are arranged, an Euclidean distance (Euclidean Distance) that is used to store the relative optimal path of this state of arrival, a relative optimal path that is used to store this state of arrival.Should the capacity of this memory select much so?
For first problem, the inventor is verified for the system of L → ∞, and its judgement time-delay is zero for the probability of ∞, the document that sees reference (Li Daoben, the statistics detection and the estimation theory of signal, Science Press, 2004).
For in second problem first, i.e. path Euclidean distance memory, generally speaking because the existence of noise, no matter any paths, always it is increasing apart from receiving signal distance, from this point see this memory as if capacity should for ∞.But because of our the interested just relative distance between them, so we can make its maximum (or minimum) apart from being zero after each calculating, promptly all deduct this maximum (or minimum) distance to the distance of each surviving path.So, what we stored will be the relative value of distance, and its capacity is exactly limited naturally.As for second, i.e. surviving path memory is got 5K or 4K is just passable according to the general length of experimental result, can ignore basically because surviving path is longer than the probability of 5K.In case if these memories are filled with and adjudicate and can adjudicate by force when also not carrying out, promptly export the initial bit that minimum range is arranged as judgement this moment.Sometimes also available majority logic judgement is promptly exported the majority in each surviving path initial bit as judgement.The very simple but somewhat inferior properties of the equipment that adopts the latter is in first kind of mode, but because the probability of judgement is very little by force, the performance loss that is caused thus also is very little.
From the above mentioned, we find different with other any communication technologys, should handle in time-domain the input of overlapping time division multiplex system, and the most handy digital form are handled problems.This just requires the receiver of this system to disperse at first to received signal and digitized processing.What is interesting is that very the overlapping people of symbol generally believe that meeting produces serious phase mutual interference; The inventor finds that symbol is overlapping and not only can not produce interference, and it is a kind of utilizable coding bound relation on the contrary, and symbol is overlapping serious more; Coding bound is long more, and coding gain is high more on the contrary.Certain this coding is a kind of encoding relation that nature forms, and is not necessarily the forced coding restriction relation.The inventor believes firmly that the symbol overlapping multiplexing cooperates goes up the encoding relation that suitable coding makes it to form optimum, can further improve systematic function.
The inventor has proved theoretically and has passed through great amount of calculation machine simulating, verifying: in random time-varying channel, for mark-hold bandwidth width B
0, the mark-hold length T
s, we can be with reducing mark space Δ T
sIncrease overlapping tuple K, at this moment the inventor finds because system bandwidth is constant, the then spectrum efficiency of system proportional increase along with the increase of K, but the transmission reliability of system (being order of diversity) will remain unchanged basically; But if the total bandwidth B of simultaneously proportional increase system (for example spread spectrum and CDMA mode realize can be earlier to superpose filtering then or other by rectangle or broadband symbol); Then the spectrum efficiency of system will remain unchanged basically; And performance is really become better and better, and transmission reliability is increasingly high, and this moment is when K → ∞; Random time-varying channel will be transformed into the best perseverance of channel performance participation property white gaussian channel, i.e. awgn channel gradually.Therefore in random time-varying channel, as long as the system linearity degree is secure with transmitting power, the way that we just can the overlapping tuple K of the relieved increase of usefulness audaciously improves the spectrum efficiency or the transmission reliability of system, or the two is taken into account.Certainly the complexity of system handles also can increase thereupon.
Below, will with embodiment 1 and embodiment 2 overlapping time-division multiplexing method and overlapping time division multiplex system be described respectively respectively.
Embodiment 1
The performing step of time-division multiplexing method of the present invention is described through present embodiment below.
Step 1: according to given channel parameter, system parameters determines the most basic some design parameters:
1) channel parameter: mainly contain the maximum time diffusing capacity Δ (second) of channel or the coherence bandwidth of channel
(conspicuous); The peak frequency diffusing capacity of channel
; The coherence time of (conspicuous) or channel
; (second);
2) system parameters: mainly contain system bandwidth B (conspicuous); Requirement to spectrum efficiency; The linearity etc.;
3) design parameter: mainly contain:
A) basic modulation level is counted M=2
Q, wherein Q is the information bit of every modulation symbol institute load.
Because under same spectrum efficiency, system complexity and M have nothing to do, can suitably select as the case may be, because system complexity is decided by the stable state number, promptly 2
Q (K-1)=M
K-1Can know according to (1) formula and (2) formula, preset time bandwidth product B
0T
sThe time, the stable state base originally is decided by spectrum efficiency.
B) basic symbol length
(second) (symbol lengths
), basic spectrum modulation signal
Width B
0(conspicuous);
In order to reduce system complexity; So the status number of system will be constant basically in random time-varying channel, be convenient to Project Realization can to let
;
If only be concerned about the transmission reliability of system; And the variation or the complexity of system mode number are indifferent to, can let
to compare mutually even less than Δ with Δ;
B
0Select bigger, T
sSelect longlyer, in random time-varying channel, can automatically produce the implied diversity gain, improve systematic function.Wherein:
Latent frequency diversity tuple
Latent time diversity tuple
The total diversity tuple of system
is the two long-pending (the branch collection also should be multiplied by space diversity tuple
if having living space).
Wherein
expression minimum positive integer.
C) relative symbol shift amount (mark space) Δ T
sOr the overlapping tuple K of symbol:
Less Δ T
sThereby bigger K can improve the spectrum efficiency of system, but system complexity and allow all corresponding increase of level number will decide according to actual conditions and needs, and its fundamental relation is following:
(K-1)ΔT
s<T
s≤KΔT
s
Wherein: T
sIn except the basic symbol width
Maximum time propagation Δ (multipath broadening) the equal time expansion factor that also should comprise system outward.
At Δ T
sMuch smaller than the coherence time of channel
The time, the displacement tap coefficient in the overlapping time-division system channel model
To be punctured into some sample values (numerical value), on the contrary the displacement tap coefficient in the overlapping time-division system channel model
To be some time waveforms.
D) system symbol sum L (or frame length T):
Because the frame length T=T of system
s+ (L-1) Δ T
s, then
When concrete system design; Preferably let coherence time
that frame length is shorter than channel promptly
this be owing to so can think constant basically with the interior characteristic of channel at total frame length T, be convenient to system engineering and realize and arrange pilot signal.
When giving fixed system and channel parameter, design parameter B
0, Δ T
s, K, L, T
sAnd factor such as overall system bandwidth B is that mutual restriction is closely connected together, should select repeatedly and carries out optimal design according to actual conditions.
The spectrum efficiency η of system is:
Give regularly less Δ T at T and K
sThereby less T
s, will cause bigger L, higher η;
Too small B
0Naturally the latent frequency diversity tuple that can cause system
Descend.But as long as overall system bandwidth B enough wide (can realize the total bandwidth B than broad through modes such as spread spectrum, CDMA or multicarriers) just there is no need to take into account it when design.Because we still can be through the latent frequency diversity tuple that interweaves, technological means such as coding improves system; Because of system is sayed; Its implied diversity tuple be decided by
rather than
just the latter understand nature and form, and the former need pass through additional technological means and just can obtain.
Step 2: according to the given characteristic of channel, system parameters and design parameter design overlapping time-multiplexed emission system.
Because multi-transceiver technologies such as overlapping time-division multiplex technology and OFDM are the same also to be a parallel Synthronous data transmission system, different just its demodulation complete different with detection mode.That yes is essentially identical to its its transmitter architecture of each symbol and conventional digital communication transmitter.
Figure 12 is exactly the sketch map of an overlapping time division multiplexing (not containing straight expansion or CDMA) system transmitter.For spread spectrum or cdma system, can add the spread spectrum arithmetic section.Its l (l=0,1,2 ..., L-1) in the symbol time interval, the complex envelope of computing that transmitter is realized (not containing the spread spectrum computing) is:
t∈[0,T] (10)
In the formula: E
0Be every symbol emitted energy;
is normalization complex modulated signal envelope, and it satisfies:
is l (l=0; 1; ..., the complex data that L-1) individual symbol transmitted.
In the formula: this modulation signal of complex radical
Frequency spectrum do
Its bandwidth is B
0Conspicuous, promptly
Basic point among the figure is to form l=0 road complex modulated signal envelope waveform with digital form earlier
Homophase a
c(t) with quadrature a
s(t) waveform; They will obtain l=1 behind shift register; 2 ..., the homophase of other each road modulation signal of L-1 and quadrature envelope waveform; With the homophase envelope waveform of said each modulation signal and quadrature envelope waveform in-phase data signal and orthogonal data signal multiplication with each corresponding symbol, obtain each modulation signal through the filtered modulation signal waveform of data-modulated; With described each modulation signal waveform addition, formation transmits.This has just guaranteed the consistency of each road complex envelope.
Step 3: it is synchronous in receiver, to form symbol time; Under synchronous condition;
handles to the reception signal in each frame, and its basic step is following:
According to sampling theorem, select suitable sampling frequency, carry out digitized processing to received signal, form the time-domain Serial No. that receives signal.Digitized processing can also can be carried out in base band at intermediate frequency, is decided in its sole discretion by the designer fully.If the designer is ready to adopt non-numeric analog form to handle problems, can certainly removes digitized processing from and directly analog signal is handled.
Step 4: actual channel is measured, found out the valuation that receives complex envelope
in the distinct symbols time interval.
Valuation to
can be used any method; For example utilize special " pilot signal " to measure; Or utilize the information of having adjudicated through the mode of received signal computing is calculated its valuation; Perhaps get the combination of the two, even can find the solution its valuation with blind estimating method.
Step Five: Using Step four found
TDM systems overlap to form a valuation channel model tap coefficients
Step 6: the basic modulation level according to system adopted is counted M=2
Q, overlapping tuple K lists whole states of system; State comprises initial and end-state, preceding transition state, back transition state and stable state, totally five types.So-called state S is stored in the pairing Q dimension of modulating data
binary data in the time domain shift register channel model (+or-) or 0 data, wherein:
Initial and respectively one of end-state, they all are
(they are meant that all Q dimension data are 0 state);
Stable state has 2
Q (K-1)=M
K-1Individual, they are respectively:
(they are meant that all Q dimension data all are states of binary (+or-) data).
Preceding transition state all has M+M with the back transition state
2+ M
3+ ...+M
K-2Individual.
Transition state is meant that the individual Q dimension data in front some (but being less than K-2) is zero state before so-called.
So-called back transition state is meant that back some (but being less than K-2) individual Q dimension data is zero state.
Initial condition is only with to 2
QTransition state shifts before individual, if K=2, then directly to 2
QIndividual stable state shifts;
End-state can only be from the front 2
QIndividual back transition state shifts comes, if K=2, then directly from 2
QIndividual stable state shifts and comes;
The forward transition state can only be from the front state (initial condition or preceding transition state) shift and come, but can be rearwards 2
QIndividual state (preceding transition state or stable state) shifts; When the forward transition state only is present in node l among the Trellis figure<K-1.
The back can be from the front 2 to transition state
QIndividual state (preceding transition state or stable state) shifts and comes, but a state (back transition state or final transition state) shifts rearwards.When the back only is present in node l among the Trellis figure>L-1 to transition state.
Owing to always have a new Q dimension binary or 0 yuan of new data to get into channel model at every turn, 0 yuan of K-1 Q dimension in front or two senior statesman's data are left channel model simultaneously, and Q dimension binary data has 2
QKind the combination, and Q tie up 0 metadata have only one maybe, so the aforesaid state transfer relationship is just arranged.
Transition state is the peculiar state of overlapping time division multiplexing, and this is the difference with general convolution code or the pairing finite state machine of Trellis sign indicating number.
Step 7: according to the state transitions relation, the state diagram of the system of drawing, Trellis figure or tree graph.And by state transitions relation respectively by (4) to (6) formula calculate respectively shift branch road coding output
promptly:
Wherein: l ∈ (0; 1; 2 ..., L-K+1) l symbol of expression input; But behind l>L-1,
S representes to shift the state that branch road is arrived at the l node;
M representes to arrive the path of this state, for preceding transition state m=1; Other state m=2
Q=M;
Because Trellis figure just finishes behind the L-K+1 node, so l might be greater than L-1 in (4) formula, but when arriving the L-K+1 node, Trellis schemes must shrink in final all-zero state.Step 8: each stable state S is prepared two memories, and one of them storage arrives the surviving path U of this state S
S, l=[u
S, 0, u
S, 1..., u
S, l], l=0,1 ..., L+K-1, wherein u
S, lBe Q dimension binary data; Another then stores this surviving path U
S, lPairing coding is exported before the l node and is received burst
Between path Euclidean distance d
S, l, (l=0,1,2 ..., L-K+1).
Can temporarily use the memory of arbitrary stable state for transition state.Because stable state has 2
Q (K-1)=M
K-1Kind, therefore every kind of memory all needs M
K-1Individual, need 2M altogether
K-1Individual memory.Step 9: in Trellis figure, implement Maximum likelihood sequence detection MLSD, its substep is following:
1 ') to make the path Euclidean distance of start node state (l=0) be d
0,0=0;
2 ') to the l node (l=1 ..., all the state S in L-K+1) calculate its whole m (m=1 or 2
Q=M) bar from last state transitions so far state the branch road code signal with receive the burst
Between branch road Euclidean distance d
S, m(l, l+1).
3 ') will arrive the branch road Euclidean distance d of this state to every state S
S, m(l is l+1) with they set out separately path Euclidean distance d of state S '
S ', l-1Addition forms m new path Euclidean distance, and therefrom selects reckling, as the path Euclidean distance d of l node state S
S, l, upgrade the path Euclidean distance memory that deposits this state S in.
4 ') node l (l=1,2 ..., L-K+1), each state S is found out the pairing surviving path U of its path Euclidean distance
S, lRenewal deposits the surviving path memory of this state S in.
To l+1 node iteron step 2 '), 3 '), 4 '), until the l=L+K-2 node.This moment, surviving path was inevitable only remaining one, and then the pairing data sequence of this surviving path is exactly that our needed conclusive judgement is exported.
5 ') when L is big; In order to use short surviving path memory; Its length can be decided to be 4K~5K, and can check the surviving path memory of each state this moment at any time in substep 4, in case find in the path that they keep identical initial part is arranged; Then that this is identical initial part is vacateed corresponding memory space simultaneously as judgement output.
6 ') for the capacity of the path Euclidean distance memory that reduces each state S; Avoid overflowing; After each step is accomplished, can make minimum (greatly) person of path Euclidean distance is zero distance, the Euclidean distance memory of other each state only store with its difference (plus or minus) be relative Euclidean distance.
Step 10: when the overlapping tuple K of symbol is excessive; Though step 9 has optimum performance; Be that it can find out and receive the path that signal has real minimum Eustachian distance, but, utilize step 9 will cause sequential detector too complicated for excessive K; Other rapid serial decoding algorithm that can consider to use for reference this moment in the convolutional encoding reduces the complexity of sequential detector, can be adapted to overlapping time division multiplex system but must they be put into the melting pot.But the reduction of any Sequence Detection complexity will be a cost with sacrificial system thresholding signal to noise ratio all.
Embodiment 2
Through present embodiment time division multiplex system of the present invention is described below.
Time division multiplexing provided by the invention system is shown in figure 13, comprises transmitter and receiver.Wherein, transmitter also comprises overlapping time division multiplexing modulating unit and transmitter unit; Receiver comprises receiving element, pretreatment unit and Sequence Detection unit.
In transmitter, input data sequence forms overlapped the transmitting on time-domain of a plurality of symbols through overlapping time division multiplexing modulating unit, by transmitter unit said transmitting is transmitted into receiver again; The receiving element of receiver receives the signal of transmitter unit emission; Form the receiving digital signals that suitable Sequence Detection unit detects through pretreatment unit; The Sequence Detection unit carries out detecting by data sequence in the time-domain to received signal, thus the output judgement.
Shown in figure 14 is the block diagram of the overlapping time division multiplexing modulating unit of transmitter of the present invention, and this overlapping time division multiplexing modulating unit comprises digital waveform generator, shift register, serial-parallel converter, multiplier and adder.
At first form the phase wiggles and the orthogonal waveforms of first modulation signal envelope waveform with digital form by digital waveform generator; Be shifted by the phase wiggles and the orthogonal waveforms of shift register again, with homophase envelope waveform and the quadrature envelope waveform that produces other each modulation signal first modulation signal envelope waveform of digital waveform generator generation; Then, serial-parallel converter converts the data sequence of serial input into the parallel in-phase data signal and the orthogonal data signal of corresponding each modulation signal; The homophase envelope waveform and the quadrature envelope waveform of the modulation signal that multiplier is then corresponding with each with each in-phase data signal of said serial to parallel conversion unit output and orthogonal data signal multiply each other, with obtain each modulation signal through the filtered modulation signal waveform of data-modulated; At last by adder with each modulation signal of multiplier output through the filtered modulation waveform addition of data-modulated, transmit with formation.
Like Figure 15 another design frame chart for transmitter of the present invention, be different with the transmitter among Figure 13, this transmitter also comprises interleave unit, coding unit and spectrum-spreading unit except comprising overlapping time division multiplexing modulating unit.
Wherein spectrum-spreading unit is used to increase overall system bandwidth, thereby the same with the effect of interleave unit and coding unit, thereby improves the latent frequency diversity tuple or the latent time diversity tuple of system.
Shown in figure 16, be the block diagram of the pretreatment unit of receiver of the present invention.This pretreatment unit is used for the auxiliary synchronous receiving digital signals sequence that forms in each frame, comprises synchronizer, channel estimator and digitized processing device.It is synchronous that wherein synchronizer forms symbol time to received signal in receiver; Then channel estimating unit is estimated channel parameter; The digitized processing device carries out digitized processing to the reception signal in each frame, is fit to the receiving digital signals sequence that the Sequence Detection unit carries out Sequence Detection thereby form.
Shown in figure 17 is the block diagram of the Sequence Detection unit of receiver of the present invention, and this Sequence Detection unit comprises analytic unit memory, comparator and a plurality of surviving path memory and Euclidean distance memory or weighted euclidean distance memory (not shown).In testing process, the analytic unit memory is made the plural convolutional encoding model and the trellis structure of overlapping time division multiplex system, and lists whole states of overlapping time division multiplex system, and storage; And comparator searches out the path with receiving digital signals minimum Eustachian distance or weighting minimum Eustachian distance according to the trellis structure in the analytic unit memory; Surviving path memory and Euclidean distance memory or weighted euclidean distance memory then are respectively applied for surviving path and the Euclidean distance or the weighted euclidean distance of storage comparator output.Wherein, described surviving path memory and Euclidean distance memory or weighted euclidean distance memory need respectively be prepared one for each stable state.Described surviving path memory length can be preferably 4K~5K.Described Euclidean distance memory or weighted euclidean distance memory are preferably only stores relative distance.
Above embodiment only is used to explain the present invention, but not is used to limit the present invention.