Method and system for detecting spread spectrum signal

FIELD: radio engineering, communication.

SUBSTANCE: invention discloses a method (100) for detecting a spread spectrum signal (s(t)) with a direct code sequence, which is transmitted at a carrier frequency and is modulated by a code signal whose length is equal to Nc unit elements, for determining code delay of said spread spectrum signal and Doppler shift relative to said carrier frequency, wherein said determination is carried out in a discrete two-dimensional space M of possible code delays and F possible frequency shifts, wherein the method (100) comprises the following steps: receiving and sampling (102) said spread spectrum signal to obtain a sampled spread spectrum signal; compressing (103) said sampled spread spectrum signal with the signal of a local copy of said code signal by performing said compression for a plurality of possible code delays between said sampled signal and said copy signal; performing parallel frequency search (104, 105, 107, 108, 109) based on results of said compression step by carrying out a step (104) calculating Fourier transform of said results. The Fourier transform is fractional Fourier transform.

EFFECT: faster scanning of a search grid and reduced computational costs.

15 cl, 12 dwg

The present invention relates to communication systems, spread spectrum and, in particular, relates to a method for detection of a signal with spread spectrum, in accordance with the generic concept of the first claim. The present invention relates to the detection signal with spread spectrum.

Among the various so-called systems with spread spectrum currently the most popular system, extended range with application code direct sequence spread spectrum (DSSS).

For example, using the technique of extended spectrum using code direct sequence spread spectrum (DSSS provides a wide bandwidth channel that is shared by a method of multiple access code division (CDMA).

Access method CDMA allows you to recognize that multiple transmitters share the same channel and that each of them uses different pseudo-random code with zero cross-correlation of these different codes. Then the receiver should choose a suitable pseudo-random code and to carry out the method of detection, which allows the receiver to determine the frequency and the delay code of the received signal. In some applications, for example in applications of GNSS (global Navigation Satellite System), the receiver does not know in advance Billy signal transmitted from a particular transmitter, to apply the method detection using possible copies of different pseudo-random codes that are available for the receiver.

In particular, for the GNSS and the first task is to search for signals transmitted by the satellites. This process is called discovery. When the signal is detected, this signal is monitored to obtain the information required to determine the location of the receiver. If the signal is for some reason lost, the detection is carried out again to search for it on the same or on a different satellite. Method detection signal displays the search in a two-dimensional time-space to determine the phase or delay of the code) and in frequency (Doppler shift of the carrier, i.e. the frequency shift relative to the nominal carrier frequency due to the fact that the satellites are moving on orbital trajectories with velocities of several kilometers per second, creating a significant Doppler shift, which is also caused by the mismatch of the frequency signals of the transmitter and receiver). More specifically, the search in two-dimensional space is defined in the area of uncertainty of such a space. The search space is quantized in both dimensions, and time and frequency is divided into a finite number of units or "cells".

In the art known to walk the ways to detect GNSS signal or in a General sense, there are various ways to detect the signal with the spread spectrum to determine the receiver delay code and Doppler shift for the received signal.

Known, for example, detection methods, which are called "matched filtering" (or serial search). In methods of this type are scanned in sequence possible "cell" Doppler shift and delay code. Methods of this type are described, for example, in US 2003/0161543, have the disadvantage due to the need for a longer period of time to scan a two-dimensional search spaces.

Also known detection methods with so-called "parallel search", which contain parallel analysis of the size of a two-dimensional grid search. Since this two-dimensional grid, there are two ways a parallel search of various types:

- parallel code search (PCS)and all the cells of the delay code are tested at the same time;

parallel frequency search (PFS)and all the cells of the frequency shift are tested simultaneously.

How PCS and PFS guarantee good results regarding the timing of the scanning grid search. However, the ways PCS require a large memory capacity, whereas the methods PFS have large computational costs.

The present invention is to provide to the te method detection with operating parameters that represent the trade-off between scan time grid search, the computational cost and the requirements of the device memory.

These disadvantages are avoided in the method of detection in accordance with paragraph 1 of the attached claims.

Other embodiments of the invention are described in the following claims.

Another objective of the present invention is a detection system according to paragraph 13 of the formula.

Further characteristics and advantages of the present invention will become apparent from the following description of preferred and non-limiting examples of the present invention, and:

Fig. 1 depicts a functional block diagram of a detection signal with spread spectrum;

Fig. 2 is a diagram of one of the stages of the detection method of Fig. 1;

Fig. 3a-3c, 4a-4c and 5a-5c is a diagram of signals related to the detection method of Fig. 1; and

Fig. 6 is an illustrative functional block diagram of the detection signal with spread spectrum.

In the drawings, the same elements are the same numerical designation.

In Fig. 1 the reference position 100 indicates in General the detection signal s(t) spread spectrum using code direct sequence spread spectrum (DSSS) or method detection signal. N the example, the signal s(t) represents the signal provided by the satellite system, GNSS, and preferably represents a signal (CDMA multiple access code division). In accordance with an additional embodiment, the signal s(t) is a CDMA signal such as provided in the UMTS system.

The signal s(t) is the signal modulated pseudo-random code signal, or PRN code with length Nc, and which is transmitted at the carrier frequency. For example, the code signal consists of Nc=4092 micro frames that are periodically transmitted every Tc=4 MS (Tc displays the so-called multi-code), as occurs, for example, in the case of E1-B and E1-C signals provided by the GALILEO satellites.

The method 100 of detecting includes a step 101 of the reception signal s(t). Preferably, the step 101 of receiving includes a step appropriate processing of the signal s(t)to demodulate and convert the above-mentioned signal in the original signal (or almost in the original signal), the phase selection phase and quadrature component of the original signal and the phase sampling at a frequency f_sto convert the received signal into a stream of complex digital samples. The above stages are known to experts in the art and therefore not described in detail.

In accordance with the embodiment stage 101 of the reception of the holding stage thinning stream of complex digital samples, to reduce the frequency of this thread from the values of f_sto low values of f_kthat displays the operating frequency or clock rate of the method 100 of detection. For example, you can set the f_k=f_s/2. In this case, as will be clarified below, to compress the received signal using a local copy of the sample PRN code with a frequency f_kwhich consists of M=tc*fk samples. Preferably, at step 101 of the reception before decimation is filtered stream of digital samples through a low pass filter, for example by means of the FIR filter, in order to avoid or reduce the effect of aliasing that occurs in the thinned stream.

In an advantageous embodiment, the step 101 of receiving includes a step for preservation of digital samples of the received signal, which in this example is filtered and weeded on a delay line provided with N taps. Preferably, the number of taps N is a number contained an integer times the number of samples M a local copy of the PRN code, i.e. there exists an integer Mn, for which the following property holds : Mn=M/N.

It should be noted that the preservation of digital samples of a received signal in a delay line provided with N taps, provides, at any clock time, a number of N consecutive samples of the received signal, i.e. Genty with length N, having different clock times and the delay code, increasing between 0 and N-1 with respect to locally available reference code sequence.

As is known, the received signal s(t) shows a relatively coming from the satellite signal delay code, or phase delay, and Doppler shift relative to the transmitted carrier frequency. Therefore, the method 100 detection is to determine the delay code, or phase delay, received signal spread spectrum and Doppler shift relative to the carrier frequency of the transmitted signal (in this example - companion), and the said definition is easily performed in a two-dimensional discrete space M of possible code delays and F possible frequency shifts, and M is the length of the local copy of the PRN code of the sample with a frequency fk, and F is a positive integer.

In Fig. 1, the method 100 of detecting includes a step 102 of providing a signal to the local copy of the pseudo-random code signal, or PRN code. This signal displays reference code sequence.

In particularly preferred embodiments, the implementation of the M samples of the PRN code of the sample with the operating frequency f_kstored in the lookup table, that is, in the structure of the data matrix containing the vectors with length N, each provided for cochraneneonatal with length N PRN code, and with successive segments of the PRN code are sequentially stored in sequential vectors mentioned structure of the data matrix. Further assume, without any limitations, the data structure formed by a matrix with N rows and Mn columns and that the first N samples of the PRN code stored in the first column C1 of the structure of the data matrix, the next N samples PRN stored in the second column C2 structure of the data matrix, and so on. In accordance with this embodiment, the step 102 of providing a local copy of the PRN code includes a step of reading from a lookup table stored in her samples PRN to provide vectors of samples of length n

In addition, the detection method 100 includes a step 103 perform compression of the sample signal with spread spectrum, in this example, the operating frequency fk, via the signal from the local copy of the PRN code, when executing the above-mentioned compression for a variety of possible code delays between the discretized signal with a signal copy of the PRN code.

In accordance with the embodiment stage 103 compression includes the steps of compression by Mn=M/N consecutive segments of length N signal local copy, Mn=M/N segments of length N of the sample signal, by performing for each of the above-mentioned Mn segments of the local signal is Noah copies of compression for N possible sequential delays between the signal of the local copy and the above-mentioned discretized signal. The compression stage 103 to provide the output set of vectors, which will be referred to as "compressed vectors containing Mn compressed vectors for each delay code.

Preferably, the shrink operation is a by-bit multiplication of two vectors of length N. Then almost exclusively be considered stage 103 compression, which compresses by performing bit vector multiplication, without imposing any restrictions.

In Fig. 2 reference position 200 marked delay line that contains the at any time N samples of the received signal, and 201 specifies the lookup table contains sample the local copy of the PRN code. At the moment N stored samples in line 200 delays are multiplied by N stored samples code in column Cl table 201 compliance. In the next step, the content of line 200 delay is updated by shifting one position of the samples stored in the previous step, and the old sample 205 is withdrawn from line 200 delay, and a new sample of 206 is introduced into the delay line. Updated content line 200 delay is multiplied, using vector multiplication 203 N samples of column Cl table 201 compliance. This is equivalent to the compression of the segments with the length N of a received signal through a segment with length N PRN code with additional delay code, equal to the delay between the received signal and the PRN code, relative to the previous stage. This operation is performed for N consecutive moments, i.e. for N delay code that updates at each stage of the content of line 200 delay, but without changing the column Cl. You then multiply the contents of the line 200 delay on column C2 for additional N delay code that is updating N times the content line 200 delay, and so on, until reaching the multiplication on column CMn. As only N multiplication of the contents of the line 200 delay and the contents of the CMn also performed, i.e. after the complete scanning of column after column of table 201 compliance, for each delay code between 0 and N-1, Mn compressed vectors with length N become available.

In connection with Fig. 1, the result of step 103 is subjected to compression the next step 104 calculate the Fourier transform for parallel frequency search results provided at the output stage 103 compression. If compression 103 is performed using a bit vector multiplication, we can see that this multiplication, combined with Fourier transform shows the phase correlation, and additionally due to the multiplication by a complex exponent, is included in the Fourier transform, is performed parallel to delete multiple Doppler frequencies.

Preferably applied Fourier transform Ave is dstanley a fractional Fourier transform. This phase transformation allows one to compute the discrete Fourier transform (DFT), which applies only to the part of the spectrum with higher resolution, as opposed to its use for the range of all possible frequencies [-f_k/2, f_k/2]. Preferably, the fractional Fourier transform is calculated for F samples that are uniformly distributed in the interval [- ∆F_w/F + ∆ F_w/F], and more preferably, for F=N, where N is the dimensionality of the vectors, the output stage 103 compression. The interval [- ∆F_w/F + ∆ F_w/F] can therefore be selected as an interval suitable to cover the most common Doppler frequency, which signal can have.

Thus, the output of the transformation is a mapping of the input vector in the space of N frequency samples, which are separated by an interval ∆ F_d=Δf_w/F= ∆ F_w/N.

In accordance with the embodiment of the fractional Fourier transform is performed by running two fast Fourier transforms (FFT) on 2*N_optsamples, where N_optfor example, selected as the lowest degree twice greater than N. it Should be noted that the relatively traditional FFT conversion, saving computational cost is very high, so as to achieve the same resolution with the FFT algorithm may have many choice is OK, equal to f_s/Δf_d=(f_sN)/(∆F_w)>>N_opt.

In a particularly preferred embodiment, the calculation of the Fourier transform 104 contains the calculation of the Fourier transform of each of the compressed vectors, the output stage 103, to obtain a set of corresponding vectors, transformed into the discrete frequency domain.

In connection with Fig. 1, method 100 includes, in addition, the step 105, 106, accumulation coherent way compressed vectors, the output stage 103 and associated with the same delay code so that there were many total vectors, each of which is associated with a corresponding delay code.

In accordance with a particularly preferred embodiment, the step 105, 106 accumulation occurs after Fourier transformation 104 through coherent accumulation in N total compressed vectors vectors associated with the same delay between 0 and N-1 and transformed into the frequency domain stage 104 Fourier transform. It should be noted that in accordance with the embodiment, corresponding to the diagrams in Fig. 1 and 2, during step 105 accumulation of two compressed vector associated with the same delay, will be available for accumulation in the same aggregate vector relative to the mentioned delay every N input samples line 200 delay. In fact, after suggesting the N new samples in line 200 delays the received and sampled signal is shifted by N samples, but also a local copy of the PRN code is shifted by the same value because of the shift column in the table 201 compliance. It may be noted that in this example the frequency converted compressed vector displays the contribution of the correlation associated with the delay code that is added (i.e., accumulates) correlation to other deposits associated with the same delay.

In a particularly preferred embodiment, step 105 accumulation includes a step 106 of synthesizing signal phase correction, the correction by the above-mentioned signal phase correction of the accumulation 105 at the previous stage, associated with a given delay between 0 and N-1, and adding the accumulated and adjusted to the new compressed vector associated with the aforementioned delay between 0 and N-1. Practically, the phase correction is applied to the content of each aggregate vector before adding a new compressed vector or, more specifically, a Fourier transform to the total content of the above vector.

Preferably, the correction phase is performed by multiplying each element of v(i), where i is between 0 and N-1, the total vector v, the corresponding correction factor phase p_i=e^j2πf_D,iwith 0≤i≤N, where f_D,idiscrete Doppler frequency (resolution often is), associated with i-th element of the converted cumulative vector.

In the example shown, starting with N multiplications content line 200 delay on the contents of column C1 by scanning a matrix 201 PRN code, column after column, when N multiplications content line 200 delay and the contents of the completed CMn, for each delay between 0 and N-1, Mn compressed vectors with length N will be accumulated, which is converted into the spatial region of the discrete Doppler frequencies. In other words, at the end of the stage 105 savings will be obtained matrix of complex numbers or matrix search, which in this example is a matrix N×N, which identifies the search space N delays and N discrete frequencies.

In relation to Fig. 1 method detection 100 contains:

serial stage 107 calculation module, simple or quadratic elements of the above matrix search;

sequential step 108 search for the maximum, or peak, of the modules of the matrix elements of the search;

serial stage 109 comparison of the maximum threshold for determining whether there has been a capture signal.

As shown in Fig. 1, step 110 of choice, if the detection of the satellite, the method 100 also includes a step 111 selecting another lookup table containing sample PRN code relative to another satellite and a new iteration is the procedure 100 for detecting said satellite, and so on, for possible additional satellites grouping GNSS, which should be detected.

As indicated by step 110 of choice, if the comparison with the threshold has not provided a signal is detected, the above steps of the method 100 is repeated from the beginning:

- reset the cumulative vectors, recorded on stage 105; and

the step 103 compression, in this case, since the second column C2 of table 201 compliance.

If after this iteration, at step 110 of choice, the detection signal is not set, the steps of method 100 is repeated, starting with the third column of the lookup table, and so on until it reaches the last column CMn. As can be easily seen from the above description, therefore, the search of M possible delay code is performed by examining consecutive segments N delays.

In accordance with a modification, instead of having to search each time to a maximum, after scanning segment N possible code delays, it is possible to repeatedly perform the search, always by dividing into sub-searches of segments N potential delays, completely scanning the entire PRN code many times and accumulating incoherent way (that is, after execution of step 107 calculation module) the value of N_acc(an arbitrary integer) NxN-dimensional grid search, obtained by repeating scanned the I matrix, contains sample PRN code from the same column. If the peak is not found, the search is performed on serial N PRN delays code for the new PRN code. If after scanning for M possible delays peak is not found, the search for a new PRN code begins again (in this practical example, the search for a new companion). It should be noted that this detection strategy is that it increases the search time, although this increases the sensitivity of the detection method. So it is convenient to set N_acc=1 with relatively high ratios of carrier power to noise power (this is equivalent to the fact that is the normal way as described in connection with Fig. 1), and instead of installing N_acc>1 for relatively low relationship of carrier to noise power.

In accordance with an additional embodiment it is possible to predict the stage of assessment of real-time sound power, adaptive varying based on assessment of the threshold used for phase comparison, at step 109. In a particularly preferred embodiment, the estimation of noise power is performed based on the converted compressed vectors, the output stage 104 conversion. For example, we computed the quadratic module (or module) complex vectors, the output stage 104 conversion, and generates the average value of all converted is Avodah (the noise power is invariant with respect to frequency) and then filtered by time, preferably using a filter with an infinite impulse response (IIR) of the first order. This filtering helps to reduce the estimation noise.

In accordance with a particularly preferred embodiment the threshold is set as a real multiple of the number of estimated power of noise, and referred to a multiple of the number is determined based on the specific requirements of probability of false alarms and updated in real time or almost real time.

The aforementioned method can be implemented by software or a combination of hardware and software, such as FPGA.

In addition a modified embodiment, which is easy to carry when in the accepted secondary signal PRN code, or a sequence of characters, or information bits, is superimposed on the primary PRN code sequence characteristics data or the secondary code of the original code is well known to experts in the art; in other words, the basic premise of the secondary code of the same length as a primary code, and the symbol or data bits of the same length as an integer number of times the primary code), since this secondary code or sequence data can distort the results of the coherent accumulation, can what to anticipate, what stage 103 compression is performed, keeping the same initial matrix column 201 for two or more complete scans all columns C1-CMn mentioned matrix.

Below the above-mentioned problem and its solution are described more clearly.

In Fig. 3a, 3b and 3c shows the variation in time of the code phase of a received signal (i.e. the input signal of the delay line signal is a local copy of the PRN code (i.e. the one that is retrieved from a lookup table) and the content of the coherent memory (step 105 of the method of Fig. 1), respectively. For clarity of presentation in Fig. 3a and 3b, the sequence of the primary PRN code displayed increasing sawtooth signal, although in reality it is a rectangular sequence of positive and negative symbols; Fig. 3c shows the amplitude of the module content coherent drive. The diagrams in Fig. 3a, 3b, 3c were obtained under the condition that the secondary code is not present. The phase code is represented graphically conventionally, as an increasing integer from 0 to the maximum number of basic assumptions in the code minus 1. In this case the code length of 1023. As can be seen from Fig. 3a, 3b, 3c, signal (Fig. 3a) and the signal PRN code (Fig. 3b) perfectly aligned, i.e. they have the same phase code in each moment of time and the coherent accumulation (Fig. 3c) always increases, reaching max the maximum value, equal to the length of the code.

In Fig. 4a, 4b, 4c displays graphs that correspond shown in Fig. 4a, 4b, 4c, for the case when there is secondary PRN code, i.e. code, the basic premise is the same length as a primary PRN code (1023 in this example), and assuming a transition character between elementary parcel secondary code and the following elementary parcel. Graphically this inversion of the sign is represented as a change of sign in the code phase of the received signal (Fig. 4a). Obviously, this inversion can occur only when the phase code is 0, i.e. at the beginning of each primary code. As in the situation of Fig. 3a and 3b, the received signal and the code signal perfectly aligned, but coherent accumulation in Fig. 4c is unable to achieve the previous maximum value (Fig. 3c), because when the sign of elementary parcel secondary code it starts to fall and the final value is almost zero. The final value depends on the initial code phase of a received signal.

To resolve this problem, found the solution involves keeping the same initial column matrix that contains the code for two periods of the code instead of only one. Period code is a temporary code length. So it must correspond to increase the coherent integration time up to 2 periods of the code.

The piano is, 5a, 5b, 5c, Fig. 5a and 5b relate to the same case as Fig. 4a and 4b (as they are identical), whereas Fig. 5c shows the result of accumulation in the case of the above-mentioned decision. Almost coherent drive is restored at the beginning of a new period code, that is, when the phase signal of the PRN code starts again from 0, so again received the maximum (absolute) value of 1023 (the sign "-" in the diagram only occurs because of the transition of the sign).

However, the result of the above operations is to increase twice as full-time detection and, in General, more than one accumulation, this entails a loss of time only one period of the code.

In Fig. 6 very schematically for purposes of illustration, is shown a block diagram of a system 600 detection for performing the above method of detection. This system, for example, is a component of the GNSS receiver as satellite Navigator.

The system 600 includes a delay line with N taps 601 for storing complex samples of a received signal of the base band or almost the base band, which is optionally pre-filtered and weeded.

The system 600 further contains at least one memory device 602, such as a lookup table to save a local copy of the PRN code, in accordance with the already described procedure in the tie is with the method 100. Preferably, the system 600 includes a memory 602 to save more than one lookup table, each of which refers to the PRN code of the corresponding satellite.

The system 600 further contains a vector multiplier 603 to perform step 103 compression method.

The system 600 also contains a module 604 for performing Fourier transformation of the complex vectors produced and output by the multiplier 603, in accordance with the description of method 100. In addition, the system includes a coherent discharged by the drive 605 to output the aggregate vectors, each of which is associated with a corresponding delay code between kN and kN-1 (where k is an integer), the module 607 for calculation module or quadratic module, and which is preferably equipped with reset non-coherent memory 614 to integrate with N_acc>1 consecutive periods of the code, in accordance with the description of method 100. Zeroing the drive 614 is performed before beginning the search for a new code signal from the satellite.

In addition, the system 600 contains detection synthesizer 606 phase correction phase by means of the multiplier 612 total vectors to allow coherent accumulation of vectors, which are sequentially provided at the output of block 604 conversion. In addition, the system 600 contains detection line 613 delay the La delays the output signal of the multiplier 613 N clock cycles, moreover, the mentioned delay takes into account the fact that the update of each of the aggregate vectors of contribution output unit 602 occurs every N clock cycles, as already described for method 100 of detection.

The system 600 also comprises a control logic block 611 to correctly reference the columns in the lookup table stored in memory 602, and to send with the correct timing required by the aforementioned method 100, the reset signal for coherent memory 604.

In Fig. 6 for clarity not shown parts of the system 100 relating to the estimation of noise power to the installation and upgrade threshold, calculation of maxima and comparison of the threshold.

The system 100 can be implemented by means of a suitable combination of hardware and software, such as c FPGA or ASIC.

Experimental results, obtained by simulation using packages MATLAB™ or GRANADA™, confirmed that the method of detecting the above-mentioned type has the advantage that it is fast, accurate, adaptable, able to detect signals of more than one type and can be used for other signals with spread spectrum than CDMA signals in the GNSS.

As an example, was modeled detection signal E1c system GNSS GALILEO size code M=T_c×F_kequal to 19 000 (ie : the samples in the period code with T _c=4 MS and F_k=4,75 MHz with four samples for each elementary parcel), and the window of Doppler search was selected as 20 kHz with 120 possible look for the frequency offset (element resolution frequency). The following table compares the characteristics of the above-mentioned method ("ACSE", as indicated in the table, for the algorithm of the mechanism of acceleration detection) with methods known in the art. Listed in the table of computational costs represent the number of complex multiplications required for scanning the entire grid search.

As you can see from the provided in this data table, the method ACSE has a computational cost that is lower than for the other algorithms, except for the parallel search phase, which, on the other hand, requires a longer FFT, which is problematic when the method is implemented on FPGA.

It is therefore evident that the method of the aforementioned type has advantages in that it requires relatively low computational cost comparable with those for detection with parallel search phase, and much lower than those for methods of another type), a relatively short time of the search and has a relatively lower requirements to the volume and memory configuration.

Specialists in the art to meet specific needs can make some modifications and variations to the above method and detection system, which nevertheless are within the scope of protection as defined in the enclosed claims.

1. The method (100) for the detection of si is Nala (s(t)) extended range with direct code sequence, which is transmitted on the carrier frequency which is modulated by the signal code with length equal Nc elementary parcels to determine the delay code signal with the spread spectrum and the Doppler shift relative to that of the carrier frequency, and the above determination is performed in a discrete two-dimensional space M of possible code delays and F possible frequency shifts, and the method (100) includes the following stages:
receiving and sampling (102) mentioned signal spread spectrum to obtain the discretized signal with spread spectrum;
the operation (103) mentioned compression of the sample signal with the spread spectrum signal a local copy of the above code signal when executing a compression set of possible code delays between these discretized signal and said signal copies;
performing a frequency search(104, 105, 107, 108, 109) through the implementation stage (104) calculate the Fourier transform;
characterized in that
these frequency search(104, 105, 107, 108, 109) is a parallel frequency search performed by the aforementioned compression stage,
mentioned Fourier transform represents the fractional Fourier transform is performed on the above results.

2. The method (100) according to claim 1, in which et is p (103) compression includes the steps of compression by Mn=M/N consecutive segments of length N of the sample signal, performing, by each of the said segments of the Mn signal local copy, compression for N possible sequential delays between the signal of the local copy and the above-mentioned discretized signal.

3. The method (100) according to claim 2, in which step (103) of the compression is to provide at the output a lot of compressed vectors, the above-mentioned set contains Mn compressed vectors for each delay code.

4. The method (100) according to claim 3, whereby the above-mentioned step of calculating the Fourier transform includes the step of calculating the Fourier transform for each of the above-mentioned compressed vectors mentioned many, to obtain a set of corresponding transformed vectors in the discrete frequency domain.

5. The method (100) according to claim 2, and
step (103) compression ratio includes a step of multiplying, sampling, waveform segments to the appropriate segments copies, providing the appropriate product of vectors, and moreover, the above-mentioned step of calculating the Fourier transform is performed on each of the mentioned works of vectors.

6. The method (100) according to claim 2, additionally containing phase coherent accumulation, in the aggregate vectors, compressed vectors associated with the same delay code, providing total N vectors, each of which is associated with a corresponding delay.

7. The method (100) according to claim 6, in which step (105) estimates the program is executed after the above-mentioned step of calculating the Fourier transform (104) and thus referred to the transformed vectors coherently accumulate in the frequency domain.

8. The method (100) according to claim 7 in which the said step (105) accumulation includes a step (106) the cancellation of the phase shift of the partial result of these savings, before adding a new contribution in the above-mentioned partial result.

9. The method (100) according to any one of items 1 to 8, in which the said step of receiving and sampling (102) includes a step of preserving mentioned a discrete signal in the delay line (200).

10. The method (100) according to claim 9, in which the delay line (200) is a delay line with N taps.

11. The method (100) of claim 10 in which the said signal is a local copy is stored in a matrix structure (201) data containing Mn vectors with length N, each for storing a segment of N samples of the mentioned code, and with successive sample segments mentioned code sequentially stored in sequential vectors mentioned matrix, and the step (103) compression ratio includes a step of multiplying N times one vector of the above matrix (201) on the content of the above delay (200), and said delay line is updated samples mentioned received signal.

12. The method (100) according to claim 11, further containing the step of finding a maximum (108), after performing at least one scanning of the Finance, through the above-mentioned multiplication for all vectors mentioned matrix (201), and, in addition, contains the establishment phase of the detection by comparing (109) the above maximum threshold.

13. The method (100) according to item 12, further comprising stages:
evaluation, since the above-mentioned step of calculating the Fourier transform of the noise power;
adaptive variation of the above-mentioned threshold on the basis of these estimated power of the noise.

14. Detection system (600)containing a processing unit for implementing the method according to any of items 1 to 13.

15. The receiver of the GNSS signal containing a detection system for 14.