CN104935407B - A kind of Turbo DFH coding and decoding methods of nonopiate frequency spectrum - Google Patents
A kind of Turbo DFH coding and decoding methods of nonopiate frequency spectrum Download PDFInfo
- Publication number
- CN104935407B CN104935407B CN201510292488.4A CN201510292488A CN104935407B CN 104935407 B CN104935407 B CN 104935407B CN 201510292488 A CN201510292488 A CN 201510292488A CN 104935407 B CN104935407 B CN 104935407B
- Authority
- CN
- China
- Prior art keywords
- frequency
- signal
- frame
- turbo
- nonopiate
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L1/00—Arrangements for detecting or preventing errors in the information received
- H04L1/004—Arrangements for detecting or preventing errors in the information received by using forward error control
- H04L1/0056—Systems characterized by the type of code used
- H04L1/0059—Convolutional codes
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04B—TRANSMISSION
- H04B1/00—Details of transmission systems, not covered by a single one of groups H04B3/00 - H04B13/00; Details of transmission systems not characterised by the medium used for transmission
- H04B1/69—Spread spectrum techniques
- H04B1/713—Spread spectrum techniques using frequency hopping
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L1/00—Arrangements for detecting or preventing errors in the information received
- H04L1/004—Arrangements for detecting or preventing errors in the information received by using forward error control
- H04L1/0045—Arrangements at the receiver end
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L1/00—Arrangements for detecting or preventing errors in the information received
- H04L1/004—Arrangements for detecting or preventing errors in the information received by using forward error control
- H04L1/0045—Arrangements at the receiver end
- H04L1/0054—Maximum-likelihood or sequential decoding, e.g. Viterbi, Fano, ZJ algorithms
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L1/00—Arrangements for detecting or preventing errors in the information received
- H04L1/004—Arrangements for detecting or preventing errors in the information received by using forward error control
- H04L1/0056—Systems characterized by the type of code used
- H04L1/0064—Concatenated codes
- H04L1/0066—Parallel concatenated codes
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Artificial Intelligence (AREA)
- Error Detection And Correction (AREA)
Abstract
A kind of Turbo DFH coding and decoding methods of nonopiate frequency spectrum, belong to wireless communication technology field.Including signal framing, Turbo codings, modulated signal generation, signal demodulation, Turbo decoding procedures, increase n register in the signal framing step, n is positive integer, i.e., RSC registers number is N+n;Adjacent frequency is set in the modulated signal generation step at intervals of 1/ (2n‑1*Tb) (Hz), i.e. the frequency points of output signal are changed into original 2nTimes;(2 are added in the signal demodulation step after each frequency sampling signal receivedn1) * M zero, pass through | FFT |2Its preceding M value is taken, obtains the random vector of observation space, finally decoding output.This method improves the frequency points in same band by nonopiate spectrum technology in the frequency bandwidth of restriction, using its coding gain more than the influence that nonopiate spectral leakage is brought, the performance of improvement Turbo DFH systems.
Description
Technical field
The invention belongs to wireless communication technology field, and in particular to a kind of nonopiate frequency under Differential Frequency Hopping Systems
The Turbo differential jumping frequency coding and decoding methods of spectrum.
Background technology
Turbo code is that Berrou etc. is proposed in the meetings of ICC ' 93.Coding side it by by sub-encoders interweave
Device parallel cascade passes through iterative decoding structure reality to realize the coding thinking of the randomization long code in Shannon's theorems at decoding end
The decoding thought of long code is now randomized, has reached the performance close to shannon limit.
Typical Turbo encoder by two recursive systematic convolutional codes (Recursive System Convolution,
RSC) sub-encoders are formed by a random interleaver cascade, and corresponding decoder architecture is translated by two soft-output coding
Code device is formed by interleaver and deinterleaver serially concatenated, and wherein interleaver is identical with the interleaver used in encoder.
《Turbo-DFH coded modulations and iterative decoding》(《Beijing Institute of Technology's journal》O. 11th of volume 25 in 2005:981-984, make
Person:Pei little Dong, He Zunwen, Kuang Jingming) a kind of Turbo-DFH code modulating methods are proposed for the first time in a text, by Turbo code
It is combined with differential jumping frequency (Differential Frequency Hoppong, DFH) technology;As a result show, as a result of
Random coded and soft output iterative decoding, the performance of BER of Turbo-DFH systems are corrected compared with conventional error correction coding and false tripping
The DFH systems of algorithm have clear improvement.
Generally in the performance evaluation to Turbo code, first component code zero of Turbo code is typically assumed that, is passed through
Tail bit corresponding with RSC1 code registers states is added in every frame can be allowed to be zeroed.But due to the work of interleaver
With every postamble bit after intertexture will not typically have corresponding relation with code registers state, therefore can not make second
Individual component coder (RSC2) zero.However, using third generation partner program (3rd Generation Partnership
Project, 3GPP) in secondary arrangement multinomial (Quadratic Polynomial Permutation, QPP) interleaver
(《3GPP TS 36.212v10.6.0》In June, 2012:13-14) the interleaver as Turbo-DFH systems, after tested, in spy
Under fixed feedback factor, can realize makes RSC2 also be zeroed while RSC1 is zeroed, i.e., double return zero techniques.Double return zero techniques can be with
The decoding backward recursion factor and the forward recursion factor is had reliable initial value, improve the free distance of code word, improve
The bit error rate performance of Turbo-DFH systems.
Pei little Dong is in thesis for the doctorate《Capabilities of HF differential frequency-hopping key technology research》(Pei little Dong, Beijing Institute of Technology, 2005)
In propose Turbo-DFH System build code processes, comprise the following steps that:
Transmitting terminal A
Step A-1. signal framings:The information source transmission cycle is TbInformation bit sequence, form per frame length be K bit frame
Signal uk, wherein finally N number of bit is added according to corresponding RSC1 buffer status per frame, in order that RSC1 is zeroed,
Wherein N is the number of RSC registers;
Step A-2.Turbo is encoded:The frame signal u that step A-1 is generatedkRSC1 is sent into, is mapped according to buffer status
Frequency label sequence a is obtained to frequency state functionk, while the frame signal u that step A-1 is generatedkThe QPP being sent into K position 3GPP
Interleaver, the frame information after being interweaved are sent into RSC2, and being then mapped to frequency state function according to buffer status obtains frequency
Piont mark sequence bk;By frequency sequence akWith frequency sequence bkIt is sent into binary system on off keying (On-Off Keying, OOK) simultaneously
It is multiplexed, delta frame a length of 2K, cycle Tb/ 2 frequency label sequence akbk;
Step A-3. modulated signals generate:The frequency label sequence a that step A-2 is generatedkbkDirect Digital frequency is sent into close
Grow up to be a useful person (Direct Digital Synthesizer, DDS), and DDS is according to frequency label according to adjacent frequency at intervals of 2/Tb
(Hz) modulation frequency signal is generated.
Receiving terminal B
Step B-1. signals demodulate:By the every frame time-ofday signals received are sampled obtain M sampled value after, by quick
Fourier transformation (Fast Fourier Transformation, FFT) and modulus square (use | FFT |2Represent) obtain observation sky
Between random vector A1 B1 A2 B2 … AK BK, wherein Ak=(Ak,1,Ak,2,…,Ak,M)T, Bk=(Bk,1,Bk,2,…,Bk,M
)T, K is frame length, and k is the frame moment, k=1,2 ..., K, and T is transposition symbol, Ak,jAnd Bk,jRepresent in a frame in kth time-ofday signals
The energy value of j-th of frequency;M is the number of frequency set intermediate-frequeney point, and M=2N, N>0, N is the number of register in RSC, 1≤j
≤M;
Step B-2.Turbo is decoded:The signal A that step B-1 is obtained1 B1 A2 B2 … AK BKDemultiplexed by OOK
Into A1 A2 … AKAnd B1 B2 … BKAnd two sub-decoders are respectively fed to, and the two sub-decoders pass through K positions respectively
QPP interleavers and corresponding deinterleaver serially concatenated in 3GPP form;The signal vector for being sent into sub-decoder will be demultiplexed
It is expressed as Y={ Yi,j:1≤i≤K, 1≤j≤M }, wherein Yi,j>=0 is that the i-th jump signal passes through | FFT |2What is obtained corresponds to
The energy value of j-th of frequency in frequency set;Two sub-decoders are respectively received K and jump vector signal, are just iterated decoding,
Wherein sub-decoder related algorithm is as follows:
(1) forward recursion
Initialization:If be 0 per the dwell vessel original state in RSC sub-encoders under frame, then forward recursion it is initial
It is worth and is
Alpk(s) the forward path metric that buffer status is s under k-hop is represented.
If the dwell vessel original state in RSC sub-encoders is not 0, there is Alp0(s)=0.
For kth=1,2 ..., K, formula is defined
Wherein s is the state variable when front jumping, and s ' is the state variable of previous dive,Represent that k-1 is jumped off by uk
Branched measurement value in caused s ' → s forward recursion calculating;A is the range value constant of modulated signal, N0It is Gauss white noise
The one-sided power spectrum density of sound, and I0() is first kind zeroth order modified Bessel function, its available following formula approximate calculation:
Wherein ln (P (uk)) it is the priori log-likelihood ratio information that previous sub-decoder provides in iterative decoding process;
ukThe value of information of information source output when kth is jumped is represented, because information source output information is binary bits signal, so ukCan only be 0 or
1;To the ln (P (u of sub-decoder 1k)) can be approximated to be
Wherein L_e21 (uk) represent that sub-decoder 2 is output to the external information of sub-decoder 1, L_e12 (uk) represent son decoding
Device 1 is output to the external information of sub-decoder 2;It similarly can obtain the priori log-likelihood ratio information of sub-decoder 2;Wherein
Then to kth=1,2 ..., K, to iteration before can calculating according to formula above:
Finally to Alpk(s) it is normalized, prevents internal memory from overflowing, obtains Alp 'k(s):
(2) backward recursion
Initialization:If the buffer status zero under per frame in RSC sub-encoders, the initial value of backward recursion are
Wherein, Betk(s) the backward path metric value that buffer status is s under k-hop is represented.
If the register in RSC sub-encoders does not carry out return-to-zero, there is BetK(s)=0.
For kth=1,2 ..., K, formula is defined
Wherein,Represent under k-hop by ukBranched measurement value in caused s ' → s backward recursion calculating;Then
To k=1,2 ..., K, backward iteration can be calculated according to formula above:
Finally to Betk(s) it is normalized, prevents internal memory from overflowing, obtains Bet 'k(s):
(3) posteriority log-likelihood ratio
First to k=1,2 ..., K, formula is defined
Wherein Γk(s ', s) is represented under k-hop by ukBranch metric in caused s ' → s posteriority log-likelihood calculations
Value;Then obtained transmitting symbol u by following formulakPosteriority log-likelihood ratio:
Wherein L1(uk) and L2(uk) be respectively sub-decoder 1 and sub-decoder 2 posteriority LLR ratio, finally by
The estimate that following rule exports according to the posteriority log-likelihood ratio of sub-decoder 2 after deinterleaver to information sourceCarry out
Judgement
Wherein L '2(uk) it is L2(uk) obtained by deinterleaver;Current decoder is supplied to the elder generation of next decoder
Test information L_e21 (uk)、L_e12(uk) can be expressed as
Wherein L_e12 ' (uk) L_e12 (u are obtained by interleaverk), and L_e21 ' (uk) L_ obtained by deinterleaver
e21(uk)。
Computing is iterated according to above-mentioned algorithm, is exported after the completion of iteration according to obtained posteriority log-likelihood ratio decoding.
However, it can be taken very wide due to having used DFH technologies during Turbo-DFH System build codes in the application
Frequency spectrum, thus the band resource bottleneck that its further develops as restriction.
The content of the invention
A kind of the defects of present invention exists for background technology, it is proposed that the Turbo differential jumping frequencies compiling of nonopiate frequency spectrum
Code method.In the frequency bandwidth of restriction, the frequency in same band is mainly improved by nonopiate spectrum technology for this method
Point number, using its coding gain more than the influence that nonopiate spectral leakage is brought, the performance of improvement Turbo-DFH systems.
Technical scheme is as follows:
A kind of Turbo-DFH coding and decoding methods of nonopiate frequency spectrum, including signal framing, Turbo coding, modulated signal life
Into the step of the demodulation of, signal, Turbo decodings, it is characterised in that increase n register in the signal framing step, n is just
Integer, the i.e. number of RSC registers are N+n;Adjacent frequency is set in the modulated signal generation step at intervals of 1/ (2n-1*
Tb) (Hz), i.e. the frequency points of output signal are changed into original 2nTimes;In each frequency received in the signal demodulation step
(2 are added after M sampled point signal of rate samplingn- 1) * M zero, then pass through | FFT |2Taking its preceding M value, (wherein M is frequency set
The number of intermediate-frequeney point), the random vector of observation space is obtained, finally enters row decoding output.
A kind of Turbo-DFH coding and decoding methods of nonopiate frequency spectrum provided by the invention, specifically include following steps:
Transmitting terminal A
Step A-1 signal framings:The information source transmission cycle is TbInformation bit sequence, form per frame length be K bit frame letter
Number uk, wherein last N+n bit is added according to corresponding RSC1 buffer status in per frame, in order that RSC1 returns
Zero;Wherein N is the RSC register numbers under quadrature condition, and n is from being just sent to nonopiate increased RSC deposits under same band
Device number (while also imply that nonorthogonal factor Rp=1/2n);
Step A-2 Turbo are encoded:The frame signal u that step A-1 is generatedkIt is sent into RSC1 and carries out 1/2nNonopiate processing, i.e.,
RSC1 buffer status numbers are changed into original 2nTimes, RSC1 exports frame length according to buffer status to frequency state mapping function
For K frequency label sequence ak, while the frame signal u that step A-1 is generatedkThe QPP interleavers being sent into K position 3GPP, are handed over
Frame information after knitting is sent into RSC2 and carries out 1/2nNonopiate processing, i.e. RSC2 buffer status number are changed into original 2nTimes, then
Frequency label sequence b is obtained according to buffer status to frequency state mapping functionk;By frequency label sequence akWith frequency label
Sequence bkIt is sent into binary system on off keying simultaneously to be multiplexed, a length of 2K of delta frame, cycle Tb/ 2 frequency label sequence
akbk;
Step A-3 modulated signals generate:The frequency label sequence a that step A-2 is generatedkbkDDS is sent into, DDS is according to frequency
Label is according to adjacent nonopiate frequency at intervals of 1/ (2n-1*Tb) (Hz) generation modulation frequency signal;Due to Turbo-DFH systems
The frequency bandwidth of occupancy is constant, therefore the frequency points of output signal are changed into original 2nTimes;
Receiving terminal B
Step B-1 signals demodulate:By the every frame time-ofday signals received are sampled obtain M sampled value after, in each frequency
(2 are added after rate sampled signaln- 1) * M zero, then pass through | FFT |2Its preceding M value is taken, obtains the random vector of observation space
A1 B1 A2 B2 … AK BK, wherein Ak=(Ak,1,Ak,2,…,Ak,M)T, Bk=(Bk,1,Bk,2,…,Bk,M)T, K is frame length, and k is
Frame moment, k=1,2 ..., K, T are transposition symbols, Ak,jAnd Bk,jRepresent the energy of j-th of frequency in kth time-ofday signals in a frame
Value;M is the number of frequency set intermediate-frequeney point, and M=2N+n, N+n>1, N+n is the number of register in RSC, 1≤j≤M;
Step B-2 Turbo are decoded:The signal A that step B-1 is obtained1 B1 A2 B2 … AK BKDemultiplexed by OOK
Into A1 A2 … AKAnd B1 B2 … BKTwo sub-decoders are respectively fed to, and the two sub-decoders pass through K positions respectively
QPP interleavers and corresponding deinterleaver serially concatenated in 3GPP form;The signal vector for being sent into sub-decoder will be demultiplexed
It is expressed as Y={ Yi,j:1≤i≤K, 1≤j≤M }, wherein Yi,j>=0 is that the i-th jump signal passes through | FFT |2What is obtained corresponds to
The energy value of j-th of frequency in frequency set, two sub-decoders are respectively received K and jump vector signal, are iterated decoding, its
Middle sub-decoder related algorithm is as follows:
(1) forward recursion
Initialization:If be 0 per the dwell vessel original state in RSC sub-encoders under frame, then forward recursion it is initial
It is worth and is
Wherein, Alpk(s) the forward path metric that buffer status is s under k-hop is represented.
If the dwell vessel original state in RSC sub-encoders is not 0, there is Alp0(s)=0.
For kth=1,2 ..., K, formula is defined:
Wherein s is the state variable when front jumping, and s ' is the state variable of previous dive,Represent that k-1 is jumped off by uk
Branched measurement value in caused s ' → s forward recursion calculating;A is the range value constant of modulated signal, N0It is Gauss white noise
The one-sided power spectrum density of sound, and I0() is first kind zeroth order modified Bessel function, its available following formula approximate calculation:
Wherein ln (P (uk)) it is the priori log-likelihood ratio information that previous sub-decoder provides in iterative decoding process;
ukThe value of information of information source output when kth is jumped is represented, because information source output information is binary bits signal, so ukCan only be 0 or
1;To the ln (P (u of sub-decoder 1k)) can be approximated to be
Wherein L_e21 (uk) represent that sub-decoder 2 is output to the external information of sub-decoder 1, L_e12 (uk) represent son decoding
Device 1 is output to the external information of sub-decoder 2;It similarly can obtain the priori log-likelihood ratio information of sub-decoder 2;Wherein
Then to kth=1,2 ..., K, to iteration before can calculating according to formula above:
Finally to Alpk(s) it is normalized, prevents internal memory from overflowing, obtains Alp 'k(s):
(2) backward recursion
Initialization:If the buffer status zero under per frame in RSC sub-encoders, the initial value of backward recursion are
Wherein, Betk(s) the backward path metric value that buffer status is s under k-hop is represented.
If the register in RSC sub-encoders does not carry out return-to-zero, there is BetK(s)=0
For kth=1,2 ..., K, formula is defined
WhereinRepresent under k-hop by ukBranched measurement value in caused s ' → s backward recursion calculating;Then
To k=1,2 ..., K, backward iteration can be calculated according to formula above:
Finally to Betk(s) it is normalized, prevents internal memory from overflowing, obtains Bet 'k(s):
(3) posteriority log-likelihood ratio
First to k=1,2 ..., K, formula is defined
WhereinRepresent under k-hop by ukBranch metric in caused s ' → s posteriority log-likelihood calculations
Value;Then obtained transmitting symbol u by following formulakPosteriority log-likelihood ratio:
Wherein L1(uk) and L2(uk) be respectively sub-decoder 1 and sub-decoder 2 posteriority LLR ratio, finally by
The estimate that following rule exports according to the posteriority log-likelihood ratio of sub-decoder 2 after deinterleaver to information sourceCarry out
Judgement
Wherein L '2(uk) it is L2(uk) obtained by deinterleaver;Current decoder is supplied to the elder generation of next decoder
Test information L_e21 (uk) or L_e12 (uk) can be expressed as
Wherein L_e12 ' (uk) L_e12 (u are obtained by interleaverk), and L_e21 ' (uk) L_ obtained by deinterleaver
e21(uk);
Computing is iterated according to above-mentioned algorithm, is exported after the completion of iteration according to obtained posteriority log-likelihood ratio decoding.
Beneficial effects of the present invention are:
1st, the present invention adds nonopiate spectrum technology in Turbo-DFH systems, correspondingly have modified frequency interval, compiles
The some algorithm of code device structure, the mapping function of buffer status to frequency state and receiving end signal demodulation;In the frequency of restriction
In bands of a spectrum are wide, communicated using nonopiate Hopping frequencies collection, increased using the coding of nonopiate frequency spectrum Turbo-DFH encoders
Benefit is more than the influence that nonopiate spectral leakage is brought, and improves the performance of Turbo-DFH systems;Frequency of the inventive method in restriction
In bands of a spectrum are wide, the frequency points of communication are added, improve the utilization rate of frequency spectrum.
2nd, non-orthogonal manner is revised as in traditional Turbo-DFH communication system frequencies interval by the present invention from orthogonal manner.
Specifically, the definition at traditional Turbo-DFH orthogonal frequencies interval is:If the frequency label sequence of system transmission is in the time of time domain
At intervals of Tb/ 2, in order to ensure the orthogonality of system adjacent frequency, then need to meet condition Δ in frequency domain adjacent frequency interval delta f
F=2/Tb(Hz);And the definition at Turbo-DFH non-orthogonal frequencies interval proposed by the present invention:If Rp is nonorthogonal factor, limiting
Determine under bandwidth, if nonorthogonal factor Rp=1/2n, then Turbo-DFH non-orthogonal frequencies are at intervals of Δ f*Rp=1/ (2n-1*Tb)
(Hz), wherein n=1,2,3 ..., i.e. frequency points under same band are changed into original 2nTimes.
3rd, the modification of frequency interval improves frequency resolution in the present invention, adds frequency points, it is therefore desirable in original
Increase register number and feedback tap number on the basis of rsc encoder;Because nonopiate frequency spectrum changes of register
The number of number and feedback tap, therefore corresponding buffer status needs to change to frequency state mapping function.
Brief description of the drawings
Fig. 1 is the frame diagram of Turbo-DFH systems;(background technology is identical with the system framework figure of the present invention)
Fig. 2 is the performance comparison figure of background technology and the Turbo-DFH systems of the present invention.
Note:In Fig. 2 background technology with the present invention Turbo-DFH systems used by simulation parameter be:
Embodiment
Embodiment
Input the intelligence sample rate of nonopiate rsc encoder:fs=5KHz (i.e. be multiplexed frequency label sequence sample rate be
10KHz);
The register number of rsc encoder:3+1 (wherein, N=3, n=1);
QPP interleaver sizes:40, information source frame signal frame length:40;
Frequency set frequency points:16, modulation system:Hexadecimal frequency shift keying (16FSK);
Nonorthogonal factor:(adjacent frequency frequency interval is 5KHz to Rp=1/2,16) each frequency sampling points is;
Wireless channel environment:Rayleigh (Rayleigh) channel;
Channel SNRs are set:20dB;
Demodulation mode:Hexadecimal frequency shift keying it is soft demodulation (will each frequency sampling 16 points add 16 0, one
Rise and be sent into | FFT |2, then take each frequency energy value of preceding 16 values as soft demodulation output);
The Turbo-DFH decoder iteration numbers of nonopiate frequency spectrum:1;
Here we have found a specific RSC feedback factor 17, while RSC1 is zeroed, be interweaved by QPP
Can also RSC2 be set to be zeroed after device, i.e. dual homed zero.
Its buffer status is to frequency state mapped function relation such as following table
So frequency state transition rule such as following table
Transmitting terminal A
Step A-1. signal framings.The information source transmission cycle is Tb=200 μ s information bit sequence, it is 40 to form per frame length
The frame signal u of bitk:[1 01110000100011 ... 10000000 1], wherein frame
Last 4 bits [0 00 1] are obtained according to corresponding RSC1 buffer status, to cause RSC1 to be zeroed;
Step A-2.Turbo is encoded.This frame signal u that step A-1 is generatedkIt is sent into RSC1 and carries out 1/2 nonopiate place
Reason, RSC1 is according to the frequency label sequence a of buffer status to frequency state mapping function output frame a length of 40k:[12 14 4
9 6 10 12 14 2 7 13 4 1 1 1 …… 2 6 10 12 14 2 6 10 8];
Then by uk40 QPP interleavers are sent into, the sequencing table of wherein interleaver is:[13 6 19 12 25 18 31
24 37 30 3 36 92 15 ... 16 29 22 35 28 1 34 7 0], the frame information u ' after then being interweavedk:
[1 0 1 0 0 0 1 1 0 0 1 0 1 1 0 …… 0 1 1 0 1 0 0 0 1]
It is re-fed into RSC2 and carries out 1/2 nonopiate processing, output frequency label sequence bk:[12 14 4 1 9 7 5 15 5
3 10 12 0 3 11 …… 12 0 3 11 14 2 6 10 8]
By frequency sequence akWith frequency sequence bkBinary system on off keying (On-Off Keying, OOK) progress is sent into simultaneously
It is multiplexed delta frame a length of 80, frequency label sequence a of the frequency label at intervals of 100 μ skbk:[12 12 14 14 4 4 9 1 6
9 10 7 12 5 14 15 2 5 7 3 13 10 4 12 …… 10 3 12 11 14 14 2 2 6 6 10 10 8
8]。
Step A-3. modulated signals generate.The frequency label sequence that step A-2 is generated is sent into DDS, sets adjacent frequency
At intervals of 5KHz, because the frequency bandwidth that Turbo-DFH systems take is constant, therefore DDS exports the frequency of Turbo-DFH signals
Number is changed into original 2 times, i.e. frequency points are 16.For frequency signal corresponding to each frequency, sampling number 16, phase
The DDS output signals answered:
Receiving terminal B
Step B-1. signals demodulate.The signal received from Rayleigh channels is as follows:
The each frequency sampling signal received is added into 16 zero, such as sampling letter corresponding to first frequency label 8
Number processing is:[0.2885+0.6302j -0.9972+0.3911j -0.5616-0.8364j -0.1120+0.3936j -
0.8558-0.2177j 0.0447-0.9994j 0.5175+0.7857j-1.8441-0.1014j 0.4133+0.2623j
0.0626+0.9637j-0.3223-0.9623j 0.6817+0.4711j-0.8466+0.7316j-0.2359-0.0938j
0.0219+1.0805j-0.0729-0.1035j 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0]
Then pass through | FFT |2Its preceding 16 value is taken again, obtains the random vector A of observation space1 B1 A2 B2 … A15
B15, wherein Ak=(Ak,1,Ak,2,…,Ak,16)T, Bk=(Bk,1,Bk,2,…,Bk,16)T, k is the frame moment, and T is transposition symbol, Ak,j
And Bk,jRepresent the energy value of j-th of frequency in kth time-ofday signals in a frame;
Step B-2.Turbo is decoded:The signal A that step B-1 is obtained1 B1 A2 B2 … A40 B40Demultiplexed by OOK
Into A1 A2 … A40And B1 B2 … B40Two sub-decoders are respectively fed to, and the two sub-decoders pass through 40 respectively
QQP interleavers and corresponding deinterleaver serially concatenated form, wherein 40 QQP interleavers here due to RSC select it is specific
Feedback factor, so being the interleaver of dual homed zero;All represented for convenience of the signal vector for representing for demultiplexing to be sent into sub-decoder
For Y={ Yi,j:1≤i≤15,1≤j≤16 }, wherein Yi,j>=0 is that the i-th jump signal is done | FFT |2What is obtained corresponds to frequency set
In j-th of frequency energy value;That is A1A2…A40The signal input of corresponding sub-decoder 1, B1 B2 … B40Corresponding son decoding
The signal input of device 2.
First, the energy signal inputted for sub-decoder 1 is expressed as Y, as follows:
(1) forward recursion Alp is obtained according to algorithmic formula abovek(s) value (the wherein ∞ of initial value definition=
1010):
It can see that, as frame moment k=1, haveL_e21(uk)=0, so havingWhereinProbable value has
Wherein sj′≠0
ThenIt is most likely to be
It is otherSo have
For moment k=2,3 ..., 14 and calculate and obtain by above algorithmic formula.
(2) backward recursion Bet is obtained according to algorithmic formula abovek(s) value:
(3) due to being to enter row decoding for the first time, so initial L_e21 (uk)=0, then L_e12 ' (uk)=L1(uk), so
Prior information L_e12 ' (u from decoder 1 to decoder 2k) be:[29802.688104 -29802.688104
29802.688104 29802.688104 29802.688104 -31932.556711 -47775.049110 -
42090.940130 -42090.940130 36934.679076 -36934.679076 -36934.679076 ...... -
32526.953964 -32526.953964 32526.953964]
Then L_e12 (u are obtained by QQP interleaversk) be:[36934.679076 -47775.049110
17214.518101 -36934.679076 -17605.298582 -17214.518101 17542.675868
17605.298582 -32526.953964 -17542.675868 29802.688104 -28401.413627
36934.679076……-28401.413627 -42090.940130 29802.688104]
2nd, the energy signal inputted for sub-decoder 2 is expressed as Y:
(1) forward recursion Alp is obtained according to algorithmic formula abovek(s) value:
(2) backward recursion Bet is obtained according to algorithmic formula abovek(s) value:
(3) the posteriority log-likelihood ratio L of decoder 2 is obtained according to algorithmic formula above2(uk) value is:
[92113.515642 -129305.310403 106291.921874 -157496.405348 -104358.782002 -
92113.515642 106102.532821 101243.996895 -87543.875384 -102949.696818 …… -
109068.072643 110282.671021]
Then L ' is obtained by deinterleaving2(uk) value is:[110282.671021 -86855.410780
87543.875384 108418.572685 63263.867304 -86291.906076 -129305.310403 -
109068.072643 -107508.815385 101243.996895 -98996.243310 -71954.312070 ……-
87543.875384-70472.365909 88361.055666], hard decision, L ' are finally carried out2(uk) >=0 item decodes output
1, L '2(uk)<0 decoding output is 0, then last decoded frame output is:[1 0 1 1 1 0 0 0 0 1 0 0 …… 0 0
1], as a result with the output frame of information source being, explanation is correct decoding.
If be iterated twice and the above decoding, need the posteriority log-likelihood ratio L of decoder 2 above2
(uk) subtract the prior information L_e12 (u that decoder 1 arrives decoder 2k) obtain L_e21 ' (uk), then L_ is obtained by deinterleaver
e21(uk), sub-decoder 1 is then fed into, then repeat step B-2, until iterations is completed and decodes output.
As for the signal coding and decoding to next frame, then repeat step A-1 to step B-2;It should be noted that previous frame is most
Decoding of the signal at moment to the next frame moment afterwards is helpful, i.e., is related between front and rear frame, such as in embodiment
In enter row decoding to next frame, then in step B-2 first time iteration decoder 1 calculating forward recursion Alpk(s) during value,
As moment k=1, haveValue to be worth corresponding to the previous frame last moment.
Using Matlab to the Turbo-DFH proposed in the present invention and the small eastern thesis for the doctorate of background technology Pei in Rayleigh
Simulation comparison once is carried out with four iteration bit error rate performances in channel, its simulation result is as shown in Figure 2.Can from accompanying drawing 2
To find out, in Rayleigh channels, BER=1 × 10-4When, background technology orthogonal iteration 1 time, iteration are respectively required for offer 4 times
12dB and 7.2dB, iteration 1 time that nonorthogonal factor of the present invention is 1/4, iteration fall below about 9.7dB and 6.1dB 4 times, believed respectively
Make an uproar and more corresponding than gain improve about 2.3dB and 1.1dB.
Claims (1)
1. a kind of Turbo-DFH coding and decoding methods of nonopiate frequency spectrum, specifically include following steps:
Transmitting terminal A
Step A-1 signal framings:The information source transmission cycle is TbInformation bit sequence, form per frame length be K bit frame signal uk,
Last N+n bit is added according to corresponding RSC1 buffer status in wherein per frame, in order that RSC1 is zeroed;Wherein
N is the RSC register numbers under quadrature condition, and n is from being just sent to nonopiate increased RSC registers number under same band;
Step A-2Turbo is encoded:The frame signal u that step A-1 is generatedkIt is sent into RSC1 and carries out 1/2nNonopiate processing, RSC1 roots
According to the frequency label sequence a of buffer status to a length of K of frequency state mapping function output framek, while step A-1 is generated
Frame signal ukThe QPP interleavers being sent into K position 3GPP, the frame information after being interweaved are sent into RSC2 and carry out 1/2nNonopiate processing,
Then frequency label sequence b is obtained according to buffer status to frequency state mapping functionk;By frequency label sequence akAnd frequency
Label sequence bkIt is sent into binary system on off keying simultaneously to be multiplexed, a length of 2K of delta frame, cycle Tb/ 2 frequency label sequence
akbk;
Step A-3 modulated signals generate:The frequency label sequence a that step A-2 is generatedkbkDDS is sent into, DDS is according to frequency label
According to adjacent nonopiate frequency at intervals of 1/ (2n-1*Tb) generation modulation frequency signal;
Receiving terminal B
Step B-1 signals demodulate:By the every frame time-ofday signals received are sampled obtain M sampled value after, adopted in each frequency
(2 are added after sample signaln- 1) * M zero, then pass through | FFT |2Its preceding M value is taken, obtains the random vector of observation space
A1B1A2B2…AK BK, wherein Ak=(Ak,1,Ak,2,…,Ak,M)T, Bk=(Bk,1,Bk,2,…,Bk,M)T, K is frame length, when k is frame
Carve, k=1,2 ..., K, T is transposition symbol, Ak,jAnd Bk,jRepresent the energy value of j-th of frequency in kth time-ofday signals in a frame;M
It is the number of frequency set intermediate-frequeney point, and M=2N+n, N+n>1, N+n is the number of register in RSC;
Step B-2Turbo is decoded:The signal A that step B-1 is obtained1B1A2B2…AK BKA is demultiplexed into by OOK1A2…AKWith
B1B2…BKBe respectively fed to two sub-decoders, and the two sub-decoders respectively by the QPP interleavers in the 3GPP of K positions and
Corresponding deinterleaver serially concatenated forms;The signal vector for demultiplexing feeding sub-decoder is expressed as Y={ Yi,j:1≤i
≤ K, 1≤j≤M }, wherein Yi,j>=0 is that the i-th jump signal passes through | FFT |2Obtained j-th of the frequency corresponded in frequency set
Energy value, two sub-decoders are respectively received K and jump vector signal, are iterated decoding output.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510292488.4A CN104935407B (en) | 2015-06-01 | 2015-06-01 | A kind of Turbo DFH coding and decoding methods of nonopiate frequency spectrum |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510292488.4A CN104935407B (en) | 2015-06-01 | 2015-06-01 | A kind of Turbo DFH coding and decoding methods of nonopiate frequency spectrum |
Publications (2)
Publication Number | Publication Date |
---|---|
CN104935407A CN104935407A (en) | 2015-09-23 |
CN104935407B true CN104935407B (en) | 2017-12-15 |
Family
ID=54122391
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201510292488.4A Expired - Fee Related CN104935407B (en) | 2015-06-01 | 2015-06-01 | A kind of Turbo DFH coding and decoding methods of nonopiate frequency spectrum |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN104935407B (en) |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107544943B (en) * | 2017-08-28 | 2021-02-19 | 湖南时变通讯科技有限公司 | Partial frequency point calculation method and device based on FFT library function |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20100205510A1 (en) * | 2001-01-09 | 2010-08-12 | Von Der Embse Urbain Alfred | QLM decoding |
CN102025451A (en) * | 2010-12-08 | 2011-04-20 | 北京理工大学 | Turbo code differential frequency hopping method for cooperative communication |
-
2015
- 2015-06-01 CN CN201510292488.4A patent/CN104935407B/en not_active Expired - Fee Related
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20100205510A1 (en) * | 2001-01-09 | 2010-08-12 | Von Der Embse Urbain Alfred | QLM decoding |
CN102025451A (en) * | 2010-12-08 | 2011-04-20 | 北京理工大学 | Turbo code differential frequency hopping method for cooperative communication |
Non-Patent Citations (1)
Title |
---|
Turbo-DFH编码调制与迭代译码;裴小东等;《北京理工大学学报》;20051130;第25卷(第11期);全文 * |
Also Published As
Publication number | Publication date |
---|---|
CN104935407A (en) | 2015-09-23 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US7856591B2 (en) | Decoding method for tail-biting convolutional codes using a search depth viterbi algorithm | |
US6606724B1 (en) | Method and apparatus for decoding of a serially concatenated block and convolutional code | |
Berkmann | On turbo decoding of nonbinary codes | |
MXPA01010239A (en) | A system and method employing a modular decoder for decoding turbo and turbo-like codes in a communications network. | |
CN105634508A (en) | Realization method of low complexity performance limit approximate Turbo decoder | |
Zhou et al. | Turbo trellis-coded differential chaotic modulation | |
Yu et al. | Hybrid Parity-Check and CRC Aided SCL decoding for polar codes | |
US6795507B1 (en) | Method and apparatus for turbo decoding of trellis coded modulated signal transmissions | |
CN114826284A (en) | Iterative decoding method based on extended Turbo code and continuous phase modulation | |
KR20030036660A (en) | Method and apparatus for a complementary encoder/decoder | |
Maunder et al. | Evolutionary algorithm aided interleaver design for serially concatenated codes | |
CN104935407B (en) | A kind of Turbo DFH coding and decoding methods of nonopiate frequency spectrum | |
JP6427461B2 (en) | Receiving device, wireless communication system, and wireless communication method | |
CN115225202B (en) | Cascade decoding method | |
Ngo et al. | Extrinsic information transfer charts for characterizing the iterative decoding convergence of fully parallel turbo decoders | |
CN112398580B (en) | Modulation method and device | |
CN108400788A (en) | The hardware implementation method of Turbo decodings | |
Huahua et al. | Analysis of turbo decoding algorithm in lte system | |
Wang et al. | Low-complexity decoding architecture for rate-compatible puncturing polar codes | |
CN111371528A (en) | Interleaving mapping method and de-interleaving de-mapping method of LDPC code words | |
Bahirgonde et al. | BER analysis of turbo decoding algorithms | |
Naranjo et al. | CCSDS 131.2-B-1 Serial Concatenated Convolutional Turbo Decoder Architecture for Efficient FPGA Implementation | |
SINDHU | VLSI Implementation of Turbo codes for LTE Systems | |
JP2012182559A (en) | Decoder | |
KR100895670B1 (en) | Turbo decoding method using a sliding window scheme, turbo decoder performing the same, and rf receiver |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20171215 Termination date: 20200601 |
|
CF01 | Termination of patent right due to non-payment of annual fee |