Disclosure of Invention
Aiming at the technical problems, the method for the differential recursion demodulation of the BPSK+DSSS starts from the signal expression of the BPSK+DSSS, derives the differential recursion demodulation method of the BPSK+DSSS, analyzes the problems existing in the differential recursion demodulation of the BPSK+DSSS, and provides a corresponding solution on the basis, namely the differential recursion demodulation method of the BPSK+DSSS based on multi-phase weighted signal synthesis.
In order to achieve the above object, the technical scheme of the present invention provides a bpsk+dsss differential recursive demodulation method based on multiphase weighted signal synthesis, which includes the following steps:
S1, preprocessing the radio frequency signal, including low-noise amplification, frequency mixing and filtering, to obtain a filtered intermediate frequency signal;
s2, sampling the filtered intermediate frequency signal obtained in the step S1 to obtain the following ADC sampling sequence:
sn＝Ancos(ωcnTs)+ξn (a)
Where an denotes the signal amplitude, ωc denotes the radio frequency or intermediate frequency carrier phase, Ts is the ADC sampling period,Fs is the ADC sampling frequency, ζn represents the noise component;
s3, removing the DC offset component in the sampling sequence Sn obtained in the step S2, and obtaining by using a formula (b)Sequence:
Wherein M is the sampling times corresponding to each baseband symbol before spreading, K is the baseband symbol spacing before spreading;
S4 based on step S3The sequence is calculated by using the formula (c) to carry out accumulation and smooth despreadingSequence:
S5, utilizing the formula (d) to pairThe sequence is subject to speed-down extraction to obtainSequence:
Wherein X is a speed-down extraction multiple,M is the length of the spread spectrum sequence, k is the sampling times of each symbol after spread spectrum;
s6 based on step S5Sequence, VK sequence was calculated using formula (e):
VK sequence characterizes the correspondingEquivalent signal amplitude of the sequence, wherein the greater VK, the correspondingThe larger the equivalent amplitude of the sequence signal, the higher the signal-to-noise ratio, where len represents the number of data samples used for signal amplitude evaluation;
S7, carrying out energy evaluation normalization, finding the maximum value VMAX＝VP in the VK sequence, and then carrying out the normalization according to the formula (f)Normalizing the sequence:
s8, for normalizedPerforming cross-correlation calculation by using a formula (g) according to the P value determined in the step S7;
Wherein, theRepresenting slaveThe extraction is carried out according to the mode of extracting 1 data every lambda data in the sequence, L is the length of an original frame synchronous sequence SYN0, SYNP is a new sequence obtained by multiplying every two elements with interval P in the original frame synchronous sequence SYN0, L-P is the length of a new sequence SYNP,
In the cross-correlation sequence Rn, find out the maximum value Rmax, and use the j point of the position where Rmax is located as the reference point, according to the formula (h)Extracting the sequence to obtain a sequence
S9, at the time of obtainingAfter the sequence, phase difference cosine values corresponding to different K values are calculated using formula (i):
Wherein Il corresponds to SYNl,Il+K in the original frame synchronization sequence SYN0 and SYNl+K,il in the original frame synchronization sequence SYN0 is obtained by formula (h);
S10, performing recursion judgment by using a formula (j):
Wherein In+1 is a symbol to be decided, In+1-K is a symbol decided before, L is the sequence length of the original frame synchronization sequence SYN0,Obtained by formula (h), a recursive decision is made starting from the last few bits of the frame header frame synchronization sequence (e.g., from the 6 th-to-last bit), thus obtaining the symbols of the entire transmission sequence.
Further, in step S1, a filtering process is performed using a ceramic filter or an LC filter.
Further, in step S2, the ADC sampling frequency fs is calculated based on the baseband rate after spreading and the number of samples corresponding to each symbol.
Further, in step S2, the ADC sampling frequency fs＝Mfb＝mkfb, where fb is the baseband symbol rate before spreading, M is the spreading sequence length, k is the number of samples of each symbol after spreading, and M is the number of samples corresponding to each baseband symbol before spreading.
Further, in step S5, lambda is set to a value in the range of lambda 5.
The technical scheme of the invention also provides a BPSK+DSSS differential recursion demodulation device based on multiphase weighted signal synthesis, which comprises the following modules:
The signal preprocessing module is used for preprocessing the radio frequency signals, including low-noise amplification, frequency mixing and filtering, so as to obtain filtered intermediate frequency signals;
the ADC sampling module is used for sampling the filtered intermediate frequency signals obtained by the signal preprocessing module to obtain the following ADC sampling sequences:
sn＝Ancos(ωcnTs)+ξn (a)
Where an denotes the signal amplitude, ωc denotes the radio frequency or intermediate frequency carrier phase, Ts is the ADC sampling period,Fs is the ADC sampling frequency, ζn represents the noise component;
a signal sequence calculation module for calculating by performing the following steps S11-S14, respectivelyA sequence of,A sequence of,Sequence and VK sequence:
S11, removing the DC offset component in the sampling sequence Sn obtained by the ADC sampling module, and obtaining by using a formula (b)Sequence:
Wherein M is the sampling times corresponding to each baseband symbol before spreading, K is the baseband symbol spacing before spreading;
s12 based on step S11The sequence is calculated by using the formula (c) to carry out accumulation and smooth despreadingSequence:
S13, utilizing the formula (d) pairSequence speed-down extraction to obtainSequence:
Wherein X is a speed-down extraction multiple,M is the length of the spread spectrum sequence, k is the sampling times of each symbol after spread spectrum;
s14 based on step S5Sequence, VK sequence was calculated using formula (e):
VK sequence characterizes the correspondingEquivalent signal amplitude of the sequence, wherein the greater VK, the correspondingThe larger the equivalent amplitude of the sequence signal, the higher the signal-to-noise ratio, where len represents the number of data samples used for signal amplitude evaluation;
An energy evaluation normalization module for finding the maximum value VMAX＝VP by in the VK sequence, and then for the pair according to formula (f)Normalizing the sequence:
A cross-correlation calculation module for normalizedA sequence, performing cross-correlation calculation by using a formula (g) according to the P value determined by the energy evaluation normalization module;
Wherein, theRepresenting slaveThe extraction is carried out according to the mode of extracting 1 data every lambda data in the sequence, L is the length of an original frame synchronous sequence SYN0, SYNP is a new sequence obtained by multiplying every two elements with interval P in the original frame synchronous sequence SYN0, L-P is the length of a new sequence SYNP,
In the cross-correlation sequence Rn, find out the maximum value Rmax, and use the j point of the position where Rmax is located as the reference point, according to the formula (h)Extracting the sequence to obtain a sequence
A phase difference cosine value calculation module for obtainingAfter the sequence, phase difference cosine values corresponding to different K values are calculated using formula (i):
Wherein Il corresponds to SYNl,Il+K in the original frame synchronization sequence SYN0 and SYNl+K,il in the original frame synchronization sequence SYN0 is obtained by formula (h);
The multi-channel signal synthesis judgment module is used for carrying out recursion judgment by using a formula (j):
Wherein In+1 is a symbol to be decided, In+1-K is a symbol decided before, L is the sequence length of the original frame synchronization sequence SYN0,Obtained by formula (h), a recursive decision is made starting from the last few bits of the frame header frame synchronization sequence (e.g., from the 6 th-to-last bit), thus obtaining the symbols of the entire transmission sequence.
Further, in the signal preprocessing module, a ceramic filter or an LC filter is adopted for filtering processing.
Further, in the ADC sampling module, the ADC sampling frequency fs is calculated based on the baseband rate after spreading and the number of samples corresponding to each symbol.
Further, in the ADC sampling module, the ADC sampling frequency fs＝Mfb＝mkfb, where fb is the baseband symbol rate before spreading, M is the spreading sequence length, k is the number of samples of each symbol after spreading, and M is the number of samples corresponding to each baseband symbol before spreading.
Further, when the signal sequence calculation module executes the step S13, the value range of lambda is lambda not less than 5.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
As shown in fig. 2, for burst-type DATA communication, the frame structure thereof is generally composed of two parts, namely, a SYN part and a DATA part. Wherein the SYN part is used for frame synchronization, identifying the start of a frame, and the DATA part is used for carrying user information, and in order to combat DATA errors, the DATA part typically also contains CRC check bytes for checking the integrity of the DATA and FEC coding information for error correction.
For the SYN field, there are two general ways, one is to use a pseudorandom sequence with sufficient length and good autocorrelation, for example, in DMR or PDT, the SYN of the voice frame is 24 symbols in length, and the corresponding sequence is:
+3-3-3-3+3-3-3+3+3+3-3+3+3-3+3+3+3-3-3+3-3-3-3+3
at the receiving end, the frame header is searched for using correlation calculations, defining the beginning of the frame.
Another SYN identifies the beginning of a frame by a series of alternating 01 sequences of fixed length + one or several bytes of signature, as in AIS (Automatic Identification System ), the SYN field is composed of alternating 01 sequences of 24 bits + one byte of signature (0 x 7E). Likewise, at the receiving end, the SYN is found by correlation calculation to achieve delimitation of the frame.
In modern communications, the former approach is more adopted, i.e. the SYN field is served by an auto-correlated pseudo-random sequence.
In either case, the SYN field is known in advance, and once a correlation peak or feature word is found, it can be considered to be the SYN field, which can be populated with the corresponding information without calculation.
For convenience we define SYN as follows:
SYN0＝[syn1,syn2,syn3,syn4,syn5,syn6,syn7,syn8......synL](1)
Wherein synn = ±1, the amplitude value thereof is normalized to 1, and the length of syn0 sequence is L.
From SYN0 we can calculate a new set of sequences as follows:
SYN1＝[syn1×syn2,syn2×syn3,syn3×syn4......,synL-1×synL](2)
SYN2＝[syn1×syn3,syn2×syn4,syn3×syn5......,synL-2×synL](3)
i.e. the new sequence is obtained by multiplying every two elements of the original sequence, which are spaced apart by K, i.e. the elements of the new sequence are defined as follows:
SYNK＝[syn1×syn1+K,syn2×syn2+K,syn3×syn3+K...,synL-K×synL] (4)
The length of the corresponding new sequence is L-K, where k=1, 2, 3, 4.
We define DSSS spreading sequences as:
c=[c0,c1,c2,......,cm-2,cm-1](5)
the spreading sequence has a length m.
The spreading is carried out k times, namely each spreading sequence element is simply repeated k-1 times, and an spread spreading sequence is obtained, namely:
C=[C0,C1,C2,......,CM-2,CM-1] (6)
Wherein, the
CJ＝cj＝cJ/k 0≤j<m 0≤J<M=k×m(7)
"/" Is the integer divide operator.
The length of the spread spectrum sequence after spreading is as follows:
M=k×m(8)
for a baseband signal after spreading, each symbol sample is sampled k times, which is equivalent to M times for a baseband signal before spreading, where k, M and M are positive integers.
For a bpsk+dsss modulation system, spreading processing in the baseband digital domain is required before BPSK modulation, that is,:
An＝ANM+J＝INCJ N=n/M,J=n%M(9)
"/" is the integer divide operator and "%" is the remainder operator. M represents the number of samples corresponding to a baseband symbol before spreading, i.e. the system sampling frequency is m=k×m times the baseband symbol rate. IN represents the nth baseband symbol value before spreading, either +1 or-1.
It is apparent that an, aNM+J, represents a J-th sample value corresponding to an nth symbol of the baseband before spreading.
For BPSK signals, it is generally expressed as follows:
s(t)=A(t)cos(ωct+φ0)+ξ(t) (10)
Where a (t) represents the signal amplitude (or +1 or-1) at time t, ωc represents the radio frequency or intermediate frequency carrier phase, Φ0 represents the carrier initial phase, which we will ignore in the following analysis process since the initial phase does not affect the subsequent analysis.
In the discrete domain, the above formula can be expressed as:
sn＝Ancos(ωcnTs)+ξn (11)
bringing formula (9) into:
sn＝sNM+J＝ANM+Jcos(ωcnTs)+ξn＝INCJcos(ωcnTs)+ξn(12)
Namely:
sn＝INCJcos(ωcnTs)+ξn n=NM+J(13)
Similarly, we obtain:
sn+KM＝IN+KCJcos(ωcnTs+ωcKMTs)+ξn+KM n＝NM+J(14)
we define:
According to formulae (13), (14) and (15), there is obtained:
The above formula was developed to obtain 3 parts, each as follows:
PART 3：INCJcos(ωcnTs)ξn+KM+IN+KCJcos(ωcnTs+ωcKMT)ξn(19)
The term CJCJ is negligible because of CJCJ ≡1.
The PARTs PART 2 and PART 3 are modulated by high-frequency signals, and can be filtered or weakened by a low-pass filtering or accumulating mode, so that the two latter two PARTs are not considered, and the following results are obtained:
It is clear that when NM≤n < (n+1) M, ININ+Kcos(ωcKMTs is a constant value, which does not vary with N, so the effective signal can be enhanced by accumulation, while ζnξn+KM is an uncorrelated or weakly correlated random variable, which is attenuated by accumulation to cancel each other out, thus we define:
it is clear that when n=nm, there will be an extremum (maximum or minimum) that occurs, as shown in the following formula:
I.e. by sliding window accumulation, the useful signal is enhanced M times, whereas the noise component is not enhanced by the accumulation operation due to uncorrelation or weak correlation.
According to the statistical theory, the noise component approaches 0 when M→infinity.
From the equation (22), in order to obtainThe value, M times accumulation is required. In order to obtain larger statistical gain, the M value is generally large, and the calculation is needed to be repeated for a plurality of times, so that the above formula is properly transformed to obtain
Namely:
by transformation we get aRecursive calculation method, using the recursive formula, calculating aThe values need only one addition and one subtraction, i.e. each value needs only two calculations, independent of the value of M. It is apparent that the use of recursive formulas can greatly reduce the amount of computation, which is very important for low-power systems based on MCUs.
As described above, when M is large,Negligible, we therefore get:
In order to avoid calculation in a high-speed domain, a timely deceleration process is needed, so that a downsampling operation is needed, the downsampling rate is generally defined to be 3-15 times of the baseband rate before despreading, the higher the theoretical downsampling rate is, the higher the time domain resolution is, the better the best sampling judgment is realized, and the higher downsampling rate is needed to be adopted as much as possible under the conditions of MCU calculation power and memory space permission.
For this we define:
i.e. every other X data, whereM is the degree of the spreading sequence, k is the sampling number of each symbol after spreading, and M corresponds to the sampling number corresponding to the symbol of the baseband before spreading.
In case the downsampling rate is high enough, i.e. the temporal resolution is large enough, i.e. the basic morphology and amplitude values of the waveform remain unchanged.
From (25), namelyIt is known that,The magnitude of (a) depends on two factors, namely, the M value, which is determined by the system parameters, and cos (ωcKMTs), which is a fixed value, and cos (ωcKMTs) which varies due to drift in carrier and local oscillator frequencies and variation in the K value.
For a certain value of K, cos (ωcKMTs) is a semi-static random variable, i.e. a constant is seen for a short time, but this constant is slowly varying over time, randomly varying between-1 and +1. It is apparent that when cos (ωcKMTs) =0,When only noise information is provided, no signal component information is provided, and when cos (omegacKMTs) = ±1, the information component is maximum, so that the subsequent signal processing is most facilitated. For subsequent signal processing, we first have to evaluate the signal amplitude, for which we define the signal equivalent amplitude value:
Instant takingIs a sufficiently long sequence, and its root mean square value is calculated, thereby estimating its equivalent signal amplitude.
Thereby obtaining a set of data, i.eComparing the sizes, confirming the largest group,
The method comprises the following steps:
Vmax＝max(V1 V2 V3 V4 V5 V6......)(28)
and for all at its maximum valueThe signal is normalized by dividing all data by the Vmax maximum:
Assuming that VP＝Vmax is maximum (P is a certain value, or is 1 or 2 or other value), this indicates the correspondingThe signal component is the largest and the noise component is the smallest in the sequenceIn order to obtain the optimal sampling decision point, SYNK (see formula (2)) is used for carrying out cross-correlation calculation on the information of the product of two adjacent baseband signals with the interval distance of K, and the position of a correlation peak is found. If the correlation peak is greater than a set threshold, the signal is considered valid and corresponds to the position of the correlation peak as the best sampling decision point of the sequence
The cross-correlation calculation is as follows:
Wherein L is the length of the frame original synchronization sequence SYN, and L-P is the length of the SYNP sequence.
Traversing Rn sequence, if the corresponding j-point cross correlation value is maximum, Rj is maximum and exceeds a certain threshold, considering that a data frame is received, taking the data frame as a reference point, and comparingAnd extracting at lambda-fold speed reduction to obtain a new sequence.
Where k=1, 2, 3, 4, 5, 6, 7, 8,j are the position offset values corresponding to the correlation peaks.
In the case where the sampling density is sufficient,
The method comprises the following steps:
in this new sequence, each sample value corresponds to the product of two adjacent baseband symbols of distance K and their phase difference cosine values.
And (3) transforming the formula (33) to obtain:
The method comprises the following steps:
for the header, i.e. SYN part, In、In+K is known, whereIs calculated, and for part In、In+K, there are:
In＝synn 0≤n<L(36)
In+K＝synn+K K≤n+K<L(37)
Therefore, the corresponding cos (ωcKMTs) can be calculated by using the formula (35), and in order to improve the accuracy of the cos (ωcKMTs) and reduce the influence of accidental errors, the processing can be performed by an accumulated smoothing method, namely:
Obtained by the formula (34):
The method comprises the following steps:
And then obtain:
using formula (41), it is easy to obtain:
By using (42) a recursive formula, provided that the In value is specified,The value and cos (omegacKMTs) value, we recursively make decisions to arrive at the whole baseband symbol sequence.
However, it should be noted that, with the above recursive formula, there is a serious problem that error diffusion, that is, when a decision is wrong, a subsequent decision is wrong, so this way of relying on a signal is only effective when the signal is good, and once a burst error occurs, error diffusion is necessarily caused, and finally, the whole data is not usable.
For this, we have to introduce multiple signals to perform in-phase superposition to prevent error diffusion, the principle is as follows:
From equation (41), we readily obtain:
For other K values, and so on, no further listing is made.
The above steps are accumulated to obtain:
Due to
Ignoring fixed coefficientsThe method comprises the following steps:
Unlike equation (42), equation (51) shows that the value of the next symbol is not only dependent on the value of the last symbol, but is also dependent on the first R values, so that the signal components are mutually enhanced and the noise components are mutually cancelled by introducing multiple paths of signals and by in-phase weighted superposition (phase adjustment is realized by multiplying cos (ωcKMTs)), thereby greatly improving the signal-to-noise ratio. Meanwhile, due to the fact that a plurality of components are introduced, error code diffusion is avoided, and even if a certain symbol is misjudged, continuous misjudgment, namely error code diffusion phenomenon, can be avoided as long as the weight of the symbol is not more than 50%.
Theoretically, the more signal branches R, the greater the statistical gain, the higher the signal-to-noise ratio,
The lower the bit error rate, the less the error spreading phenomenon. However, from the practical standpoint, an increase in R leads to an increase in the amount of computation and an increase in the memory space requirement, so from the engineering standpoint, R is generally taken to be 5 or 6. When r=6, it can be ensured that in any case, the specific gravity of any one component is much lower than 50%, i.e. for any value of θ:
Error diffusion can be effectively inhibited, so that the problems of cost effectiveness and performance are solved, and R=6 or R=5 is usually adopted.
From the above analysis, we get the demodulation process of the bpsk+dsss burst communication system as shown in fig. 3:
(1) Amplifying, mixing and filtering
To facilitate subsequent processing, the radio frequency signal needs to be preprocessed, including signal low noise amplification, mixing, and filtering. For example, through the above processing, the radio frequency signal can be shifted to a suitable intermediate frequency (e.g. 10.7 MHz), and then the unwanted out-of-band signal can be filtered by a filter (e.g. a ceramic filter or an LC filter);
(2) ADC sampling
According to the baseband rate after spreading and the sampling number corresponding to each symbol, the sampling frequency of the ADC is obtained by calculation, and the filtered intermediate frequency signal is sampled to obtain an ADC sampling sequence:
sn＝Ancos(ωcnTs)+ξn (11)
Where Ts is the ADC sampling period, which is the inverse of the ADC sampling frequency fs, i.e
And fs＝Mfb＝mkfb, where fb is the symbol rate of the baseband before spreading, M is the length of the spreading sequence, k is the number of samples corresponding to each symbol period after spreading, and m=mk is the number of samples corresponding to each baseband symbol before spreading.
(3) Calculation ofSequence(s)
For the ADC sample data, an ADC sample original sequence sn containing no dc component is obtained after removing the dc offset component, using formula (15), namely:
Where M is the number of samples corresponding to one symbol (before spreading) period, and K is the baseband symbol (before spreading) spacing, e.g., calculated for the cases of k=1, 2,3,4, 5, 6, respectively, to obtain a new sequence.
Based on equation (14) above,
sn+KM＝An+KMcos(ωcnTs+ωcKMTs)+ξn+KM
For the cases k=1, 2, 3, 4, 5, 6, respectively, 6 sets of new sequences were calculated, i.e
(4) Accumulated smooth despreading, calculationSequence(s)
For the above 6 sets of sequencesUsing formula (24), i.e
Respectively calculating 6 groups of new sequences, namely
(5) Extracting at a reduced speed to obtainSequence(s)
For the above 6 sets of sequencesUsing equation (26), i.e
Wherein X is a reduced speed extraction multiple, λ is the number of samples corresponding to each baseband symbol after reduced speed extraction, and m=λx.
On the pair ofThe sequence is subjected to a deceleration extraction to obtain 6 sets of signal sequences, i.e
The extraction principle is that λ should be as large as possible, with the larger λ corresponding to a higher temporal resolution, which is more advantageous for the subsequent decision, if the computational effort allows. In general, λ is greater than or equal to 5, i.e., the number of samples corresponding to each baseband symbol period is guaranteed to be an integer and greater than or equal to 5.
(6) Calculation of the VK sequence
According to formula (27), i.e
Respectively for the 6 groupsThe sequence, V1、V2、V3、V4、V5、V6, is a new set of sequences, which characterize the corresponding sequenceThe greater the equivalent signal amplitude of the sequence of VK, the correspondingThe larger the equivalent amplitude of the sequence signal, the higher the signal-to-noise ratio, where len represents the number of samples of data used for signal amplitude evaluation, and in theory, the larger len, the more accurate the evaluation, typically len will be greater than 2M/X, i.e., corresponding to more than two symbols of sample data.
(7) Normalization
In the VK sequence, the maximum VMAX＝VP is found and then according to formula (29), namely:
for 6 groupsThe sequences were normalized.
(8) Cross-correlation calculation to find out optimal sampling decision point
For normalizedPerforming cross-correlation calculation by using a formula (30) according to the P value determined in the step (7);
Wherein, theRepresenting slaveThe sequence is extracted according to the mode of extracting 1 data every lambda data, L is the length of an original frame synchronizing sequence SYN0, SYNP is a new sequence obtained by multiplying every two elements with the interval P in the original frame synchronizing sequence SYN0, and L-P is the length of a synchronizing sequence SYNP.
In the cross-correlation sequence, a maximum value Rmax, is found, and the j point of the position where Rmax is located is taken as a reference point, according to the formula (31), namely:
For a pair ofThe sequences are subjected to downsampling to obtain 6 groups of new sequences, namely
(9) Calculating the cosine value of the phase difference
In the process of obtainingAfter the sequence, equation (38) is used, namely:
Phase difference cosine values corresponding to different K values are calculated.
Its Il corresponds to element SYNl,Il、Il+K in the frame synchronization original sequence SYN0, and to SYNl、synl+K in the original frame synchronization sequence SYN0, i.e. I0＝syn0,I1＝syn1, and so on. Wherein il is based on the result of step (8)Sequence.
(10) Recursive decision
On the basis of the above work, on the basis of obtaining the relevant information, the formula (51) is used, that is:
Where In+1 is the symbol to be decided, In+1-K is the symbol that has been decided before, and starting from the last few bits (e.g., from the 6 th bit) of the frame header frame synchronization sequence, a recursive decision can be made, so as to obtain the symbol of the entire transmission sequence. Wherein, theBased on step (8)Sequence.
The invention avoids phase discrimination operation and despreading operation in a high-speed digital domain through algorithm optimization design, utilizes product information among adjacent symbols through careful algorithm design, then performs accumulation, thereby completing DSSS despreading operation requiring great computation force through phase transformation, then rapidly reduces data from a high-speed domain to a low-speed domain (from MSps to dozens kSps) through a downsampling technology, then performs complex signal processing and logic judgment, comprises complex operations such as signal amplitude evaluation, comparison, correlation calculation, phase difference cosine value calculation, multipath signal synthesis and the like, and obtains corresponding processing gain through the complexity of the algorithm, thereby improving the system receiving sensitivity.
The algorithm of the invention has the following advantages:
(1) The direct despreading algorithm is avoided, so that the computational power requirement is reduced, the algorithm can be used in MCU with medium performance or without using a special ASIC chip or FPGA chip or a high-performance DSP chip, the advantages of low cost, low power consumption, small volume and the like are achieved, and meanwhile, the development difficulty is reduced and the development period is shortened;
(2) The algorithm has robustness, is insensitive to frequency offset, can be used for a large frequency offset system, and reduces the requirement on an RF system;
(3) The algorithm improves the receiving sensitivity and the system stability by synthesizing multiple paths of in-phase weighted signals, and under the condition that the synthesizing path is more than 5, the receiving sensitivity is superior to that of a coherent receiving system, and simulation calculation shows that the method has about 3dB statistical gain;
(4) The algorithm ensures that the high signal-to-noise ratio branch signals occupy larger components in signal synthesis and the signal-to-noise ratio branch signals occupy smaller components through a multipath in-phase weighted signal synthesis technology, and avoids signal-to-noise ratio deterioration caused by equal weight synthesis;
(5) By multi-branch signal synthesis, error code diffusion caused by overlarge weight of a certain branch signal is avoided, and block or grid error code phenomenon caused by wrong judgment of a certain symbol is avoided, so that FEC failure is caused;
(6) The algorithm does not need carrier frequency recovery, is suitable for burst communication, and reduces the complexity of a hardware circuit.
The foregoing is only a preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art, who is within the scope of the present invention, should make equivalent substitutions or modifications according to the technical scheme of the present invention and the inventive concept thereof, and should be covered by the scope of the present invention.