Multipath interference elimination method and device, estimation method and device, and receiver

Abstract

The present disclosure relates to a multipath interference elimination method, a channel impulse response spectrum estimation method, a multipath interference elimination device, a channel impulse response spectrum estimation device and a receiver. The multipath interference elimination method includes obtaining a cross-correlation signal spectrum between a GNSS input signal and a receiver local signal; weighting the cross-correlation signal spectrum to obtain a weighted spectrum; performing sliding window summation on the weighted spectrum to obtain a sliding window summation spectrum; normalizing the sliding window summation spectrum to obtain a channel impulse response spectrum; transforming the channel impulse response spectrum to the time domain to obtain a time domain waveform of the channel impulse response, and obtaining a code phase deviation of a direct path by identifying the time domain waveform through a code phase deviation, so as to eliminate multipath interference. The present disclosure can accurately estimate the code phase deviation after the multipath signal overlaps with the direct path.

Description

Technical Field

[0001] The present disclosure relates to the field of communication technology, and in particular to a multipath interference elimination method, a channel impulse response spectrum estimation method, a multipath interference elimination device, a channel impulse response spectrum estimation device and a receiver. Background Art

[0002] The Global Navigation Satellite System (GNSS) can achieve all-weather, global and high-precision continuous navigation and positioning, and has developed rapidly in recent years. When receiving the signal sent by the GNSS, the ground receiver needs to calculate the deviation of the current code phase relative to the received GNSS code phase.

[0003] In the prior art, the shape information of the autocorrelation function of a PRN code (Pseudo Random Noise Code) is usually used to estimate the code phase deviation.

[0004] However, when GNSS signals are transmitted to the ground, they will be reflected multiple times, thus forming multipath signals that interfere with each other. After the multipath signals overlap with the direct path, the above-mentioned PRN code autocorrelation function method cannot accurately estimate the code phase deviation.

[0005] It should be noted that the information disclosed in the above background technology section is only used to enhance the understanding of the background of the present disclosure, and therefore may include information that does not constitute the prior art known to ordinary technicians in the field. Summary of the invention

[0006] The purpose of the present disclosure is to provide a multipath interference elimination method, a channel impulse response spectrum estimation method, a multipath interference elimination device, a channel impulse response spectrum estimation device and a receiver, thereby at least to a certain extent overcoming the existing problem that the code phase deviation cannot be accurately estimated after the multipath signal overlaps with the direct path.

[0007] Other features and advantages of the present disclosure will become apparent from the following detailed description, or may be learned in part by the practice of the present invention.

[0008] According to a first aspect of the present disclosure, a multipath interference elimination method is provided, the method comprising:

[0009] Obtain the cross-correlation signal spectrum of the GNSS input signal and the receiver local signal;

[0010] Weighting the cross-correlation signal spectrum to obtain a weighted spectrum;

[0011] Performing sliding window summation on the weighted spectrum to obtain a sliding window summed spectrum;

[0012] Normalizing the sliding window sum spectrum to obtain a channel impulse response spectrum;

[0013] The channel impulse response spectrum is transformed into the time domain to obtain the time domain waveform of the channel impulse response, and the time domain waveform is identified by code phase deviation to obtain the code phase deviation of the direct path, so as to eliminate multipath interference.

[0014] Optionally, weighting the cross-correlation signal spectrum to obtain a weighted spectrum includes:

[0015] The cross-correlation signal spectrum is multiplied by a weighting coefficient of each frequency point to obtain the weighted spectrum, wherein the weighting coefficient of each frequency point is a quotient of a spectrum of a corresponding frequency point in a non-multipath situation and a noise variance of the corresponding frequency point.

[0016] Optionally, performing sliding window summation on the weighted spectrum to obtain a sliding window summation spectrum includes:

[0017] A spectrum segment is selected for each frequency point of the weighted spectrum through a window function, and the spectrum segments of each frequency point are summed to obtain a sliding window summed spectrum of the corresponding frequency point.

[0018] Optionally, the window function is a rectangular window function.

[0019] Optionally, the window function widths corresponding to different frequency points are different.

[0020] Optionally, normalizing the sliding window sum spectrum to obtain a channel impulse response spectrum includes:

[0021] Sum the normalized parameters preset within the window function range of each frequency point to obtain the normalized coefficient of the corresponding frequency point;

[0022] The sliding window sum spectrum of each frequency point is divided by the corresponding normalization coefficient to obtain the channel impulse response spectrum of the corresponding frequency point.

[0023] According to a second aspect of the present disclosure, a method for estimating a channel impulse response spectrum is provided, comprising:

[0024] Obtaining a cross-correlation signal spectrum between an input signal and a local signal, or obtaining a cross-correlation signal spectrum between an input signal and a pilot signal;

[0025] Weighting the cross-correlation signal spectrum to obtain a weighted spectrum;

[0026] Performing sliding window summation on the weighted spectrum to obtain a sliding window summed spectrum;

[0027] The sliding window sum spectrum is normalized to obtain a channel impulse response spectrum.

[0028] According to a third aspect of the present disclosure, a multipath interference elimination device is provided, the device comprising:

[0029] A first spectrum acquisition module, used to acquire a cross-correlation signal spectrum between a GNSS input signal and a receiver local signal;

[0030] A second spectrum acquisition module, used for weighting the cross-correlation signal spectrum to obtain a weighted spectrum;

[0031] A third spectrum acquisition module, configured to perform sliding window summation on the weighted spectrum to obtain a sliding window summed spectrum;

[0032] A fourth spectrum acquisition module, used for normalizing the sliding window sum spectrum to obtain a channel impulse response spectrum;

[0033] The code phase deviation identification module is used to transform the channel impulse response spectrum into the time domain to obtain the time domain waveform of the channel impulse response, and obtain the code phase deviation of the direct path through the code phase deviation identification of the time domain waveform to eliminate multipath interference.

[0034] According to a fourth aspect of the present disclosure, there is provided a device for estimating a channel impulse response spectrum, comprising:

[0035] A first spectrum acquisition module, used to acquire a cross-correlation signal spectrum between an input signal and a local signal, or to acquire a cross-correlation signal spectrum between an input signal and a pilot signal;

[0036] A second spectrum acquisition module, used for weighting the cross-correlation signal spectrum to obtain a weighted spectrum;

[0037] A third spectrum acquisition module, configured to perform sliding window summation on the weighted spectrum to obtain a sliding window summed spectrum;

[0038] The fourth spectrum acquisition module is used to perform normalization processing on the sliding window sum spectrum to obtain a channel impulse response spectrum.

[0039] According to a fifth aspect of the present disclosure, a receiver is provided, the receiver comprising: the multipath interference elimination device as described above.

[0040] The technical solution provided by the present disclosure may have the following beneficial effects:

[0041] In the exemplary embodiments of the present disclosure, the multipath interference elimination method, the channel impulse response spectrum estimation method, the multipath interference elimination device, the channel impulse response spectrum estimation device and the receiver, on the one hand, weight the cross-correlation signal spectrum of the obtained GNSS input signal and the receiver local signal to obtain a weighted spectrum, which can increase the weight of the spectrum with a high signal-to-noise ratio in the channel impulse response spectrum estimation, achieve unbiased estimation, and minimize the impact of noise. On the other hand, by performing sliding window summation on the weighted spectrum, the direct path spectrum can be enhanced, the noise and multipath spectrum can be attenuated by utilizing the properties that the direct path spectrum changes slowly, the multipath spectrum changes faster, and the noise correlation of each frequency point is weak, so as to achieve the purpose of eliminating noise and multipath spectrum interference. On the other hand, by normalizing the sliding window sum spectrum, the problem of amplitude distortion caused by inconsistent spectrum gains of each frequency point during the weighting and sliding window summation process can be eliminated, so as to achieve the purpose of recovering the signal. On the other hand, after eliminating the noise and multipath spectrum interference, the code phase deviation of the direct path is obtained by code phase deviation identification, and only a small number of correlators are needed to obtain a more accurate linear output of the code phase deviation. In addition, the multipath channel impulse response spectrum obtained after the weighting, sliding window summation, and normalization processing is transformed into the time domain as a linear superposition of delayed impulse functions, which has better noise reduction and multipath influence elimination performance, and greatly improves the anti-multipath performance.

[0042] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] The drawings herein are incorporated into the specification and constitute a part of the specification, showing embodiments consistent with the present disclosure, and together with the specification, are used to explain the principles of the present disclosure. Obviously, the drawings described below are only some embodiments of the present disclosure, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. In the drawings:

[0044] Figure 1 A direct path and multipath superposition waveform diagram according to an exemplary embodiment of the present disclosure is schematically shown;

[0045] Figure 2 Another direct path and multipath superposition waveform diagram according to an exemplary embodiment of the present disclosure is schematically shown;

[0046] Figure 3 A flowchart of a multipath interference elimination method according to an exemplary embodiment of the present disclosure is schematically shown;

[0047] Figure 4A schematic diagram of an autocorrelation waveform of a satellite PRN code according to an exemplary embodiment of the present disclosure is schematically shown;

[0048] Figure 5 A schematic diagram of an impulse function waveform according to an exemplary embodiment of the present disclosure is schematically shown;

[0049] Figure 6 Schematically shows an autocorrelation function spectrum diagram according to an exemplary embodiment of the present disclosure;

[0050] Figure 7 Schematically shows a noise spectrum diagram of a GPS L1C / A signal after being correlated with a local PRN according to an exemplary embodiment of the present disclosure;

[0051] Figure 8 A noise spectrum diagram after dividing a correlation signal spectrum by a reference correlation function spectrum according to an exemplary embodiment of the present disclosure is schematically shown;

[0052] Fig. 9 A multipath channel impulse response waveform diagram according to an exemplary embodiment of the present disclosure is schematically shown;

[0053] Fig.10 A flowchart of a GNSS receiver from antenna reception of satellite signals to baseband tracking according to an exemplary embodiment of the present disclosure is schematically shown;

[0054] Fig.11 A processing flow chart of a single tracking channel according to an exemplary embodiment of the present disclosure is schematically shown;

[0055] Fig.12 The flowchart of frequency domain code phase error estimation involved in an exemplary embodiment of the present disclosure is schematically shown;

[0056] Fig.13 The flowchart of the frequency domain anti-multipath algorithm involved in the exemplary embodiment of the present disclosure is schematically shown;

[0057] Fig.14 A flow chart schematically illustrates a method for estimating a channel impulse response spectrum according to an exemplary embodiment of the present disclosure;

[0058] Fig.15 A block diagram of a multipath interference elimination device according to an exemplary embodiment of the present disclosure is schematically shown;

[0059] Fig.16 A block diagram of a channel impulse response spectrum estimation device according to an exemplary embodiment of the present disclosure is schematically shown. DETAILED DESCRIPTION

[0060] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in a variety of forms and should not be construed as limited to the embodiments set forth herein; rather, these embodiments are provided so that this disclosure will be comprehensive and complete and will fully convey the concepts of the example embodiments to those skilled in the art. The same reference numerals in the figures represent the same or similar parts, and thus their repeated description will be omitted.

[0061] In addition, the described features, structures or characteristics may be combined in one or more embodiments in any suitable manner. In the following description, many specific details are provided to provide a full understanding of the embodiments of the present disclosure. However, those skilled in the art will appreciate that the technical solutions of the present disclosure may be practiced without one or more of the specific details, or other methods, components, devices, steps, etc. may be adopted. In other cases, known structures, methods, devices, implementations, materials or operations are not shown or described in detail to avoid blurring various aspects of the present disclosure.

[0062] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities may be implemented in software form, or these functional entities or parts of functional entities may be implemented in one or more software hardened modules, or these functional entities may be implemented in different networks and / or processor devices and / or microcontroller devices.

[0063] When receiving signals from GNSS, the ground-based GNSS receiver needs to align the receiver's PRN code with the satellite's PRN code to calculate the pseudorange between the receiver antenna and the satellite.

[0064] Since the relative position between the receiver and the satellite is constantly changing, the receiver's PRN code needs to constantly adjust the code phase to keep it aligned with the received satellite's PRN code, or in other words, the receiver needs to track the satellite's PRN code. During the tracking process, the receiver needs to calculate the deviation of the current code phase relative to the received satellite's PRN code phase, and then adjust the DLL (Delay-locked Loop) according to the code phase deviation to align the receiver's PRN code with the received satellite's PRN code.

[0065] The commonly used method at present is to use the shape information of the PRN code autocorrelation function to estimate the code phase deviation, and then adjust the DLL to align the local code with the received satellite PRN. For example, in GPS L1C / A (Global Positioning System, where the L1 band is the only band with two different GPS civilian signals), the satellite PRN code autocorrelation function is a symmetrical triangle with a width of two chips. If the local instant code is aligned with the satellite transmitted PRN code, then its absolute value |P| related to the transmitted PRN code is aligned with the vertex of the autocorrelation function. The advance code E and the lag code L that are equidistant from the instant code, then the absolute values ​​after correlation with the transmitted PRN code are symmetrically distributed on both sides of the vertex, such as Figure 1 (a) and Figure 2 (a). If the transmitted PRN code is earlier than the local instant code, then Figure 1 (c) shows that |E| is greater than |L|; conversely, Figure 2 (c) shows that |L| is greater than |E|. Using this relationship, we can get two commonly used EML (Early-Minus-Late) code phase discriminators (Code Discriminator):

[0066]

[0067]

[0068] The essence of the above two code phase discriminators is to use the shape information of the known PRN code autocorrelation function to estimate the code phase deviation. In addition, there are other discriminators with the same basic principles as the above discriminators but different forms, which will not be repeated here.

[0069] The above two code phase discriminators are methods of estimating code phase deviation using the relative size of E, P and L. This method can obtain a linear output of code phase deviation in the absence of multipath signals and only requires a small number of correlators, which has been widely used.

[0070] However, in the presence of multipath signals, the PRN code autocorrelation function is a triangle with a duration of two chips. After the multipath signals within plus or minus one chip overlap with the direct path, the output waveform after the received signal is correlated with the local code is no longer symmetrical (i.e. Figure 1 (c) and Figure 2 E and L in (c) are no longer symmetric), which causes an error in the code phase discriminator output; and the greater the power ratio of the multipath to the direct path and the greater the multipath delay, the greater the discriminator output error; in addition, the multipath error of this identification method based on the correlation peak shape will not decrease as the signal-to-noise ratio increases, that is, a stronger satellite signal will not be beneficial in improving the multipath error.

[0071] It should be noted that Figure 1 middle, Figure 1 (a) shows the waveform of the direct path. Figure 1 (b) shows a waveform of a multipath, where the amplitude of the multipath is 1 / 2 of the direct path, the delay is 0.1 code chip, and the multipath phase is in phase with the direct path. Figure 1 (c) Yes Figure 1 The direct path in (a) and Figure 1 (b) is the waveform diagram after multipath superposition.

[0072] Figure 2 middle, Figure 2 (a) shows the waveform of the direct path. Figure 2 (b) shows a waveform of a multipath, where the amplitude of the multipath is 1 / 2 of the direct path, the delay is 0.1 code chip, and the phase of the multipath is opposite to that of the direct path. Figure 2 (c) Yes Figure 2 The direct path in (a) and Figure 2 (b) is the waveform diagram after multipath superposition.

[0073] Large multipath errors require the receiver to take more time to reach convergence and deteriorate positioning accuracy and stability, which is a recognized problem in today's GNSS navigation and positioning. Although there are some relatively effective anti-multipath algorithms, they cannot effectively eliminate the impact of multipath environment on pseudorange calculation. I will not go into details here.

[0074] The exemplary embodiment of the present disclosure provides a multipath interference elimination method based on the above-mentioned PRN code autocorrelation function. While retaining the advantage of high accuracy of estimating code phase deviation through shape information of the PRN code autocorrelation function, this method can be applied to direct path signals interfered by multipath signals, and can effectively eliminate the influence of multipath signals on pseudorange calculation.

[0075] Implementation Method 1

[0076] Reference Figure 3 , shows a flow chart of a multipath interference elimination method according to an exemplary embodiment of the present disclosure. Figure 3 As shown, the multipath interference elimination method may include the following steps:

[0077] Step S310, obtaining a cross-correlation signal spectrum between a GNSS input signal and a receiver local signal;

[0078] Step S320, weighting the cross-correlation signal spectrum to obtain a weighted spectrum;

[0079] Step S330, performing sliding window summation on the weighted spectrum to obtain a sliding window summed spectrum;

[0080] Step S340, normalizing the sliding window sum spectrum to obtain a channel impulse response spectrum;

[0081] Step S350: transform the channel impulse response spectrum into the time domain to obtain the time domain waveform of the channel impulse response, and identify the time domain waveform through the code phase deviation to obtain the code phase deviation of the direct path to eliminate multipath interference.

[0082] According to the multipath interference elimination method in the exemplary embodiment of the present disclosure, on the one hand, the cross-correlation signal spectrum of the obtained GNSS input signal and the receiver local signal is weighted to obtain a weighted spectrum, which can increase the weight of the spectrum with a high signal-to-noise ratio in the channel impulse response spectrum estimation, achieve unbiased estimation, and minimize the impact of noise. On the other hand, by performing sliding window summation on the weighted spectrum, the direct path spectrum can be enhanced, the noise and multipath spectrum can be attenuated by utilizing the properties that the direct path spectrum changes slowly, the multipath spectrum changes faster, and the noise correlation of each frequency point is weak, so as to achieve the purpose of eliminating noise and multipath spectrum interference. On the other hand, by normalizing the sliding window summation spectrum, the problem of amplitude distortion caused by inconsistent spectrum gains at each frequency point during the weighted and sliding window summation process can be eliminated, so as to achieve the purpose of recovering the signal. On the other hand, after eliminating the noise and multipath spectrum interference as mentioned above, the code phase deviation of the direct path is obtained by code phase deviation identification, and only a small number of correlators are needed to obtain a more accurate linear output of the code phase deviation. In addition, the multipath channel impulse response spectrum obtained after the above weighting, sliding window summation and normalization processing is transformed into the time domain as a linear superposition of delayed impulse functions, which has good noise reduction and multipath influence elimination performance, greatly improving the anti-multipath performance.

[0083] The following will take the reception of GPS L1C / A as an example to describe in detail the multipath interference elimination method provided by the exemplary embodiment of the present disclosure:

[0084] Among them, the waveform function required for the multipath interference elimination method provided in this exemplary embodiment includes: Figure 4 The autocorrelation waveform of the satellite PRN code shown, the function of the autocorrelation waveform in this exemplary embodiment is called the reference correlation function r0(n); Figure 5 The impulse function waveform shown, its corresponding function is recorded as δ(n).

[0085] In step S310 , a cross-correlation signal spectrum between a GNSS input signal and a receiver local signal is obtained.

[0086] For ease of representation, the amplitude after direct path correlation is normalized to 1. When multipath exists, the waveform function r(n) of the input signal after cross-correlation with the local signal can be expressed as:

[0087]

[0088] Among them, the complex number c i Represents the amplitude and phase of each multipath, usually |c i |<1; l is the number of multipaths; τ0 represents the tracking deviation of the current direct path; τ i It represents the tracking deviation of the direct path after the delay time i, and is usually greater than τ0; ζ(n) is Gaussian noise with variance ρ.

[0089] It should be noted that the autocorrelation function and spectrum of the noise in the GNSS input signal after being correlated with the local PRN code will change and is not a flat spectrum. Since the direct path amplitude has been normalized to 1, we have:

[0090]

[0091] The signal-to-noise ratio (SNR) of the correlation output is cor =SNR cor =CN0·L coh .

[0092] Among them, CN0 represents the ratio of the satellite signal power received by the antenna to the thermal noise spectral density N0, L coh is the coherent accumulation time length of the correlation operation.

[0093] The multipath signal can be expressed in the form of convolution as:

[0094]

[0095] Represents the convolution operation. Let:

[0096]

[0097] Where δ(n) is the discrete impulse function and h(n) is the channel impulse response. The spectra of r0(n), r(n), h(n), and ζ(n) can be expressed as:

[0098]

[0099] Where DTFT (Discrete-time Fourier Transform) represents discrete Fourier transform. jω )=DTFT(r(n)) is the cross-correlation signal spectrum. If the received signal is a GNSS signal, it is the cross-correlation signal spectrum between the GNSS input signal and the receiver local signal.

[0100] The relationship between the cross-correlation signal spectrum and the channel impulse response spectrum can be obtained through the convolution theorem:

[0101] R(e jω )=R0(e jω )H(e jω )+ψ(e jω ) (8)

[0102] The power spectrum of r0(n) |R0(e jω )| 2 like Figure 6 As shown. With the spectrum E(|ψ(e jω )| 2 ) will vary with SNR cor Inversely proportional to GPS, 24MHz bilateral bandwidth, 30MHz sampling rate, SNR cor = 1, E(|ψ(e jω )| 2 )like Figure 7 As shown in Figure 1, where E() represents the operation of taking the mathematical expectation of the random variable. In the case of stable tracking, SNR cor can be reliably estimated, and it can be considered that E(|ψ(e jω )| 2 ) and |R0(e jω )| 2 is a known quantity.

[0103] like Figure 4 As shown in , the spectrum of the reference correlation function shows a downward trend as the frequency increases, and there are multiple zero points, resulting in the corresponding frequency domain estimation of the multipath channel after division. The higher the frequency, the greater the noise. In addition, at the zero point of the spectrum of the reference received signal or the reference correlation function, the information of the frequency domain multipath channel impulse response is completely lost, and the noise is significantly amplified, as shown in Figure 5 shown.

[0104] The impulse response of a multipath channel can be represented by a linear combination of impulses with different delays, as shown in equation (7). Corresponding to the frequency domain, it is a linear combination of complex exponential functions of different frequencies, and the frequency is proportional to the delay time of the corresponding time domain impulse pulse. In addition, the random noise spectrum is distributed in the entire frequency domain. In the process of GNSS signal tracking, the tracking deviation can usually be limited to a lower range, so that the impulse corresponding to the direct path in the channel impulse response belongs to a slowly changing low-frequency component in the frequency domain. Specifically, if the local code is completely aligned with the direct path signal, it is a direct current; the impulse corresponding to the multipath with a larger delay in the channel impulse response belongs to a rapidly changing high-frequency component in the frequency domain.

[0105] There are two points to note about the correlated signal in the frequency domain: 1. The spectrum of the noise contained in the correlated signal is no longer flat, and the noise intensity at different frequencies varies greatly, resulting in large variations in the noise variance at different frequencies. 2. The spectrum of the correlated signal is the product of the channel spectrum and the known reference correlation function spectrum, with noise superimposed on it; by dividing the spectrum of the correlated signal by the spectrum of the reference correlation function, an estimate of the signal spectrum can be obtained, but the noise deteriorates significantly near integer multiples of the code frequency.

[0106] As mentioned above, for the spectrum of the received GNSS signal, the low-frequency part of the channel impulse response that changes slowly in the frequency domain and approaches DC contains the direct path signal that is expected to be processed, while the high-frequency component that changes rapidly in the frequency domain contains the multipath components and noise that are expected to be eliminated. Then, through sliding window filtering, the direct path signal can be retained while attenuating the multipath signal and reducing the noise.

[0107] Since the random spectrum noise distributions at each frequency point vary greatly, the multipath interference elimination method proposed in the exemplary embodiment of the present disclosure includes:

[0108] Through steps S320 to S340, the cross-correlation signal spectrum is weighted to obtain a weighted spectrum; the weighted spectrum is summed by sliding window to obtain a sliding window sum spectrum; the sliding window sum spectrum is normalized to obtain a channel impulse response spectrum.

[0109] Among them, weighting the cross-correlation signal spectrum to obtain a weighted spectrum can specifically include: obtaining a weighting coefficient corresponding to the minimum signal-to-noise variance in the channel impulse response spectrum; initializing the weighting coefficient to obtain an initialized weighting coefficient and a normalization coefficient; and weighting the input signal by the initialized weighting coefficient to obtain a weighted spectrum.

[0110] In the exemplary embodiment of the present disclosure, since R0(e jω ), R(e jω ), H(e jω ),ψ(e jω ) is a continuous function of frequency. In actual processing, discrete sampling values ​​are used, and the calculation can be accelerated by fast Fourier transform (FFT).

[0111] Assume that the length of the correlator is N, the length of the FFT is M, and M is usually greater than N and is an integer power of 2. The commonly used normalized frequency is expressed as follows:

[0112]

[0113]

[0114]

[0115]

[0116] The spectrum of other signals can be expressed in a similar way. Then we have:

[0117] R(k)=R0(k)H(k)+ψ(k)

[0118] R0(k) is known and can be pre-calculated and stored; ψ(k) is the random noise spectrum. In the case of continuous tracking, due to SNR cor can be reliably estimated, E(|ψ(k)| 2 ) can also be calculated in real time.

[0119] At this time, H(k) can be expressed as:

[0120]

[0121] The final estimate is τ0, and during stable tracking, τ0 is usually small, indicating that the channel impulse response corresponds to the direct path component. belongs to the low-frequency component of H(k), while |τ i |>|τ0|(i>0), and the rest are higher frequency components of H(k). Sliding window filtering can attenuate more high frequency components of multipath, while retaining low frequency components of direct path. What is actually processed is R(k) containing noise spectrum ψ(k). Low pass filtering of R(k) can also attenuate high frequency noise, thus achieving noise reduction effect.

[0122] Let the estimated value

[0123]

[0124] Since the spectrum of the channel impulse response corresponding to the direct path is a low-pass component, as long as Right now It can be considered that H(k) does not change in the range of kW / 2≤k≤k+W / 2.

[0125] Taking M = 256 as an example, assuming that the loop converges after traction and the tracking deviation is less than 40 meters, that is, |τ0|<4, then W can be 16, that is, a low-pass filter with a coefficient length of 16. If the loop converges further, the filter length can continue to increase. Another basis for selecting W is SNR cor , when SNR cor When W is high, a smaller W can be selected to obtain better multipath resolution; otherwise, a larger W can be selected to obtain better noise performance. In practical applications, W can be dynamically adjusted according to the loop convergence degree and satellite signal strength.

[0126] As mentioned above, H(i) is considered constant within the considered window, so the estimation of H(i) requires:

[0127]

[0128] In addition, in order to minimize the estimated signal-to-noise variance, we need to:

[0129]

[0130] Considering that ψ(k) is a random noise spectrum, we take its mathematical expectation. Let:

[0131]

[0132] Defining SNR cor =1 hour can be obtained in advance, then:

[0133]

[0134] The optimization problem (10) is simplified to:

[0135]

[0136] The Lagrange extreme value method is used to determine the weight coefficient:

[0137]

[0138] Obtain the weighting coefficient c corresponding to the minimum noise variance of the channel impulse response spectrum estimation i :

[0139]

[0140] R0(k) is known, CN0 and SNR in the case of continuous tracking cor can be reliably estimated, so can also be reliably estimated. To complete this algorithm, first determine R0(k), Initialize weighting coefficients and normalization coefficients:

[0141]

[0142]

[0143] Then the above algorithm based on noise variance weighted filtering is divided into the following process: multiply the cross-correlation signal spectrum by the weighting coefficient of each frequency point to obtain a weighted spectrum, wherein the weighting coefficient of each frequency point is the quotient of the spectrum of the corresponding frequency point in the absence of multipath and the noise variance of the corresponding frequency point, that is:

[0144]

[0145] Next, a spectrum segment is selected for each frequency point of the weighted spectrum through a window function, and the spectrum segment of each frequency point is summed to obtain a sliding window summed spectrum of the corresponding frequency point.

[0146] Among them, the rectangular window function f(i) can be used for sliding window summation:

[0147]

[0148] The window function widths corresponding to different frequency points may be different, so that an appropriate window function width may be selected according to the actual frequency point conditions.

[0149] make Right now:

[0150]

[0151] The normalization coefficient of the corresponding frequency point is obtained by summing the normalization parameters preset within the window function range of each frequency point; the channel impulse response spectrum of the corresponding frequency point is obtained by dividing the sliding window sum spectrum of each frequency point by the corresponding normalization coefficient.

[0152] After the spectrum is weighted and windowed, the signal-to-noise ratio is improved, but the absolute value of the spectrum will be magnified or reduced. Figure 8 As shown. According to formulas (10) and (11), to fully implement the algorithm, the following normalization calculations are required:

[0153]

[0154] The problem of amplitude distortion caused by inconsistent spectrum gain at each frequency point in the weighted and sliding window summation process can be eliminated to achieve the purpose of signal recovery. Fig. 9 The waveform shown.

[0155] Fig. 9 In the waveform shown, the peak in the channel impulse response that is significantly larger than the noise floor represents the impulse response of the main path, where the earliest peak corresponds to the impulse response of the direct path, and the other peaks correspond to the impulse response of the multipath. As shown in the figure, P1 is the direct path component of the channel impulse response, and P2 is the multipath component.

[0156] Since the processed signal is a discrete signal, the peak of the impulse response function is usually not at the sampling point. It is necessary to adopt step S350 to transform the channel impulse response spectrum into the time domain to obtain the time domain waveform of the channel impulse response, and to identify the time domain waveform through the code phase deviation to obtain the code phase deviation of the direct path to eliminate multipath interference.

[0157] That is, zi Perform Fourier transform to get the time domain waveform u i , and input to Code Discriminator to obtain code phase deviation estimation. The Fourier transform may be IDFT (Inverse Discrete Fourier Transform) or IFFT (Invert Fast Fourier Transformation).

[0158] Through the multipath interference elimination method provided by the exemplary embodiment of the present disclosure, a more accurate estimate of the peak value of the earliest impulse response signal can be obtained, and the code phase estimate of the satellite's PRN code can be estimated through the peak value, which can eliminate multipath interference and improve the accuracy of code phase estimation. In addition, the frequency domain signal can use a high-efficiency FFT algorithm; the spectrum of the reference received signal or the reference correlation function can be obtained in advance and stored for repeated use without repeated calculation; the convolution in the time domain corresponds to the multiplication in the frequency domain (the deconvolution in the time domain corresponds to the division in the frequency domain), and the operation in the frequency domain is simpler. Thus, the amount of calculation is simplified.

[0159] Although the exemplary embodiment of the present disclosure takes GPS L1C / A as an example, since the estimated channel impulse pulse response is independent of the received signal, the proposed multipath interference elimination method is not affected by the specific satellite signal autocorrelation waveform in terms of working principle and is applicable to all current GNSS signals or other signals.

[0160] Based on the description of the multipath interference elimination method described above, this exemplary embodiment describes in detail the entire process steps of the GNSS receiver from receiving satellite signals to finally obtaining the code phase deviation of the direct path:

[0161] refer to Fig.10 , provides a flow chart of a GNSS receiver from receiving a satellite signal from an antenna to baseband tracking. To illustrate the position of the exemplary embodiment of the present disclosure in the GNSS signal tracking process, Fig.11 A processing flow chart of a single tracking channel is provided, describing the generation process of the cross-correlation signal, which is output to the post-processing module for estimating the code phase, carrier frequency and carrier phase deviation, wherein the multipath interference elimination algorithm is completed in the code phase deviation estimation. Fig.12 A flow chart of frequency domain code phase error estimation according to an exemplary embodiment of the present disclosure is provided. Fig.13 A flow chart of a frequency domain anti-multipath algorithm involved in an exemplary embodiment of the present disclosure is provided.

[0162] The specific steps include:

[0163] Step 1: See Fig.10, the GNSS carrier signal is first received by the antenna of the receiver 1001 .

[0164] Step 2: See Fig.10 The carrier signal received by the antenna is processed by the receiver 1001. The receiver 1001 usually includes modules such as low noise amplifier, mixer, filter, automatic gain control, etc., and finally outputs an intermediate frequency analog signal.

[0165] Step 3: See Fig.10 The intermediate frequency analog signal is converted into a digital signal by the analog-to-digital converter 1002, and subsequent processing will be performed in the digital domain.

[0166] Step 4: See Fig.10 ,To facilitate the subsequent tracking channel processing, the digital signal output by the analog-to-digital converter 1002 needs to be pre-processed by the pre-processor 1003, which includes digital mixing, anti-interference, downsampling, digital filtering, and weight quantization to a lower bit width, and finally output to the parallel tracking channel 1004 to track multiple satellites simultaneously.

[0167] Step 5: See Fig.11 In the tracking channel, the signal preprocessed by the preprocessor 1003 first passes through the digital mixer 1101 to compensate for the local oscillator deviation and the Doppler frequency deviation and phase of the tracked satellite. The local oscillator deviation is the same for the same type of satellites, while the Doppler deviation and phase of each satellite are different, which is related to the relative speed and distance between the satellite and the receiver. Another input of the mixer 1101 is a digital carrier signal from the carrier digital oscillator module 1102.

[0168] Step 6: See Fig.11 The carrier digital oscillator module 1102 is controlled by the frequency deviation and phase deviation signals output by the phase-locked loop / frequency-locked loop 1103, and outputs a complex exponential signal with the same frequency and phase as the received satellite signal to the mixer 1101 in step 5 above.

[0169] Step 7: See Fig.11 The code digital oscillator module 1104 is controlled by the code phase deviation signal output by the delay phase locked loop 1105 to generate the expected code rate signal and code phase signal that are the same as the received tracking satellite signal.

[0170] Step 8: See Fig.11 According to the code rate signal and code phase signal output by the code digital oscillator module 1104, the code generator module 1106 generates a specific pseudo-random code signal for the tracked satellite, that is, a local pseudo-random code discrete signal, which usually takes ±1.

[0171] Step 11: See Fig.11The local pseudo-random code signal generated by the code generator module 1106 is output to the code memory 1107. The length of the code memory 1107 is the same as the number of parallel correlators 1108 and corresponds one to one.

[0172] Step 12: See Fig.11 Each correlator 1108 first multiplies the pseudo-random code signal stored in a unit in the corresponding code memory 1107 by the signal output from the aforementioned step 1 with the frequency offset and phase offset removed, so as to remove the pseudo-random code modulation of the received signal.

[0173] Step 11: See Fig.11 , the signal outputted from step 12 without frequency offset, carrier phase deviation, and pseudo-random code modulation is accumulated, and the accumulated result is outputted to the post-processing module 1109 at a preset interval, and cleared to start the next accumulation. Multiple correlators 1108 are processed in parallel, and accumulation, output, and clearing are performed at the same time, and the time interval between adjacent correlators 1108 is the same as the sampling interval of the local pseudo-random code signal.

[0174] Step 12: See Fig.12 For the post-processing of frequency domain code phase deviation estimation, the first step is to perform a fast Fourier transform 1201 to transform the parallel correlator output results in the time domain into the frequency domain.

[0175] Step 13: See Fig.12 The spectrum of the correlation signal output by the correlator obtained in step 12 needs to be subjected to frequency domain noise reduction estimation 1202 to obtain a relatively reliable channel impulse response spectrum.

[0176] Step 14: See Fig.13 The frequency domain noise reduction estimation 1202 in the exemplary embodiment of the present disclosure first performs weighting 1301 based on the noise variance of each frequency point. Let R0(i) be the spectrum of the reference correlation function received at frequency point i in the absence of multipath, is the noise variance of the frequency point calculated based on channel 0. In order to reduce noise, the weight coefficient is adjusted along with the noise variance. As the reference spectrum |R0(i)| increases, the noise weight coefficient decreases. For example, the noise weight coefficient can be proportional to the reference spectrum amplitude |R0(i)| and proportional to the spectrum noise variance. Inversely proportional relationship, that is, the value of the weight coefficient of frequency point i Multiply the spectrum R(i) of frequency point i by the corresponding weight coefficient to obtain the weighted spectrum x i .

[0177] Step 15: See Fig.13 , the weighted results are summed 1302 with sliding windows to reduce noise and attenuate multipath, that is, for frequency point i, the weighted x iThe noise reduction output y is obtained by summing the nearby spectra i After weighting in step 14, the window function adopts a rectangular window function f(i), and the coefficient within the window function coverage range is 1, or is defined by formula (15). The width of the window function at different frequency points i is different. In this implementation, the window function width is fixed to W+1 for example, and W is an even number greater than zero. For the frequency close to the minimum frequency point i, specific The window function covers the range For the maximum frequency i, specific The window function covers the range For other frequency points i, the window function coverage range is The weighted spectrum output in step 14 is summed within the window function range of frequency point i as defined above to obtain the sliding window summation result y of frequency point i i .

[0178] Step 16: See Fig.13 , the result y after sliding window summation i Normalization 1303 is performed to obtain the channel impulse response spectrum estimation. After the processing of steps 14 and 15, the spectrum of each frequency point will be enlarged and reduced, and normalization is required to match the actual channel impulse response. First, a set of normalization parameters is initialized. For frequency point i, let R0(i) be the spectrum of the reference correlation function received at frequency point i in the absence of multipath, is the noise variance of the frequency point calculated based on channel 0. The normalized parameter value of frequency point i varies with |R0(i)| 2 Increase and increase, with Specifically, if the weight coefficients and window functions of steps 14 and 15 are used, the normalized parameters can be compared with the reference correlation function spectrum energy |R0(i)| 2 Proportional to the noise variance Inversely proportional, that is The normalization operation is as follows: sum the normalization parameters within the coverage range of the window function used in step 15 to obtain the normalization coefficient of frequency point i, and divide the spectrum after the sliding window summation obtained in step 15 by the normalization coefficient to obtain the CIR spectrum estimate z i .

[0179] Step 17: See Fig.12 , perform IFFT or IDFT1203 on the channel impulse response spectrum estimate obtained in step 13 to obtain a channel impulse response estimate in the time domain.

[0180] Step 18: See Fig.12A code phase deviation identifier 1204 is used to estimate the code phase deviation and output it to a delay-locked loop (DLL).

[0181] On the basis of the above-mentioned embodiments, the exemplary embodiments of the present disclosure provide a method for estimating a channel impulse response spectrum for processing various signals.

[0182] Fig.14 A flow chart of a method for estimating a channel impulse response spectrum according to an exemplary embodiment of the present disclosure is provided. Fig.14 As shown, the method for estimating the channel impulse response spectrum may include the following steps:

[0183] Step S1410: obtaining a cross-correlation signal spectrum between an input signal and a local signal, or obtaining a cross-correlation signal spectrum between an input signal and a pilot signal;

[0184] Step S1420, weighting the cross-correlation signal spectrum to obtain a weighted spectrum;

[0185] Step S1430, performing sliding window summation on the weighted spectrum to obtain a sliding window summed spectrum;

[0186] Step S1440: normalize the sliding window sum spectrum to obtain a channel impulse response spectrum.

[0187] The specific operation process and principle of the above steps S1410 to S1440 have been described in detail in the above embodiments, and will not be repeated here. The difference is that the input signal in step S1410 can be a GNSS input signal or other signals except the GNSS input signal; the pilot signal is a signal sent in the telecommunications network for the purpose of measurement or monitoring, and this signal is usually a single frequency.

[0188] According to the channel impulse response spectrum estimation method in the exemplary embodiment of the present disclosure, on the one hand, the cross-correlation signal spectrum of the obtained input signal and the local signal of the receiver is weighted to obtain a weighted spectrum, which can increase the weight of the spectrum with a high signal-to-noise ratio in the channel impulse response spectrum estimation, achieve unbiased estimation, and minimize the influence of noise. On the other hand, by sliding window summing the weighted spectrum, the direct path spectrum can be enhanced, the noise and multipath spectrum can be attenuated by utilizing the properties that the direct path spectrum changes slowly, the multipath spectrum changes faster, and the noise correlation of each frequency point is weak, so as to achieve the purpose of eliminating noise and multipath spectrum interference. On the other hand, by normalizing the sliding window summed spectrum, the problem of amplitude distortion caused by inconsistent spectrum gains of each frequency point during the weighted and sliding window summing process can be eliminated, so as to achieve the purpose of recovering the signal. On the other hand, the channel impulse response spectrum obtained after the above-mentioned weighting, sliding window summing, and normalization processing is a linear superposition of delayed impulse functions when transformed to the time domain, and has good noise reduction and multipath influence elimination performance.

[0189] It should be noted that although the steps of the method in the present disclosure are described in a specific order in the drawings, this does not require or imply that the steps must be performed in this specific order, or that all the steps shown must be performed to achieve the desired results. Additionally or alternatively, some steps may be omitted, multiple steps may be combined into one step, and / or one step may be decomposed into multiple steps, etc.

[0190] Furthermore, this example implementation also provides a multipath interference elimination device.

[0191] Fig.15 FIG. 1 schematically shows a block diagram of a multipath interference elimination device according to an exemplary embodiment of the present disclosure. Fig.15 According to an exemplary embodiment of the present disclosure, the multipath interference elimination device 1500 may include a first spectrum acquisition module 1510, a second spectrum acquisition module 1520, a third spectrum acquisition module 1530, a fourth spectrum acquisition module 1540 and a code phase deviation identification module 1550.

[0192] Specifically, the first spectrum acquisition module 1510 can be used to obtain the cross-correlation signal spectrum between the GNSS input signal and the receiver local signal; the second spectrum acquisition module 1520 can be used to weight the cross-correlation signal spectrum to obtain a weighted spectrum; the third spectrum acquisition module 1530 can be used to perform sliding window summation on the weighted spectrum to obtain a sliding window summation spectrum; the fourth spectrum acquisition module 1540 can be used to normalize the sliding window summation spectrum to obtain a channel impulse response spectrum; the code phase deviation identification module 1550 can be used to transform the channel impulse response spectrum into the time domain to obtain the time domain waveform of the channel impulse response, and obtain the code phase deviation of the direct path through the code phase deviation identification of the time domain waveform to eliminate multipath interference.

[0193] Since the functional modules of the multipath interference elimination device in the embodiment of the present disclosure are the same as those in the above method embodiment, they are not described in detail here.

[0194] Furthermore, this exemplary embodiment also provides a device for estimating a channel impulse response spectrum.

[0195] Fig.16 FIG. 1 schematically shows a block diagram of a device for estimating a channel impulse response spectrum according to an exemplary embodiment of the present disclosure. Fig.16 According to an exemplary embodiment of the present disclosure, the channel impulse response spectrum estimation device 1600 may include a first spectrum acquisition module 1610 , a second spectrum acquisition module 1620 , a third spectrum acquisition module 1630 , and a fourth spectrum acquisition module 1640 .

[0196] Specifically, the first spectrum acquisition module 1610 can be used to obtain the cross-correlation signal spectrum between the input signal and the local signal, or to obtain the cross-correlation signal spectrum between the input signal and the pilot signal; the second spectrum acquisition module 1620 can be used to weight the cross-correlation signal spectrum to obtain a weighted spectrum; the third spectrum acquisition module 1630 can be used to perform sliding window summation on the weighted spectrum to obtain a sliding window summed spectrum; the fourth spectrum acquisition module 1640 can be used to normalize the sliding window summed spectrum to obtain a channel impulse response spectrum.

[0197] Furthermore, this exemplary embodiment also provides a receiver. The receiver in this exemplary embodiment includes the above-mentioned multipath interference elimination device 1500. The specific details of the multipath interference elimination device 1500 have been described in detail in the aforementioned embodiment part, so they will not be repeated here.

[0198] Those skilled in the art will readily appreciate other embodiments of the present disclosure after considering the specification and practicing the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of the present disclosure that follow the general principles of the present disclosure and include common knowledge or customary technical means in the art that are not disclosed in the present disclosure. The specification and embodiments are to be considered exemplary only, and the true scope and spirit of the present disclosure are indicated by the claims.

[0199] It should be understood that the present disclosure is not limited to the exact structures that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present disclosure is limited only by the appended claims.

Claims

1. A multipath interference elimination method, characterized in that: The method comprises: Obtain the cross-correlation signal spectrum of the GNSS input signal and the receiver local signal; Weighting the cross-correlation signal spectrum to obtain a weighted spectrum; Performing sliding window summation on the weighted spectrum to obtain a sliding window summed spectrum; Normalizing the sliding window sum spectrum to obtain a channel impulse response spectrum; The channel impulse response spectrum is transformed into the time domain to obtain a time domain waveform of the channel impulse response, and the time domain waveform is subjected to code phase deviation identification to obtain a code phase deviation of a direct path, so as to eliminate multipath interference; The step of weighting the cross-correlation signal spectrum to obtain a weighted spectrum includes: The cross-correlation signal spectrum is multiplied by a weighting coefficient of each frequency point to obtain the weighted spectrum, wherein the weighting coefficient of each frequency point is a quotient of a spectrum of a corresponding frequency point in a non-multipath situation and a noise variance of the corresponding frequency point.

2. The multipath interference elimination method according to claim 1, characterized in that: Performing sliding window summation on the weighted spectrum to obtain a sliding window summation spectrum includes: A spectrum segment is selected for each frequency point of the weighted spectrum through a window function, and the spectrum segments of each frequency point are summed to obtain a sliding window summed spectrum of the corresponding frequency point.

3. The multipath interference elimination method according to claim 2, characterized in that: The window function is a rectangular window function.

4. The multipath interference elimination method according to claim 2, characterized in that: The window function widths corresponding to different frequency points are different.

5. The multipath interference elimination method according to claim 2, characterized in that: Normalizing the sliding window sum spectrum to obtain a channel impulse response spectrum includes: Sum the normalized parameters preset within the window function range of each frequency point to obtain the normalized coefficient of the corresponding frequency point; The sliding window sum spectrum of each frequency point is divided by the corresponding normalization coefficient to obtain the channel impulse response spectrum of the corresponding frequency point.

6. A method for estimating a channel impulse response spectrum, characterized in that: include: Obtaining a cross-correlation signal spectrum between an input signal and a local signal, or obtaining a cross-correlation signal spectrum between an input signal and a pilot signal; Weighting the cross-correlation signal spectrum to obtain a weighted spectrum; Performing sliding window summation on the weighted spectrum to obtain a sliding window summed spectrum; Normalizing the sliding window sum spectrum to obtain a channel impulse response spectrum; The step of weighting the cross-correlation signal spectrum to obtain a weighted spectrum includes: The cross-correlation signal spectrum is multiplied by a weighting coefficient of each frequency point to obtain the weighted spectrum, wherein the weighting coefficient of each frequency point is a quotient of a spectrum of a corresponding frequency point in a non-multipath situation and a noise variance of the corresponding frequency point.

7. A multipath interference elimination device, characterized in that: The device comprises: A first spectrum acquisition module, used to acquire a cross-correlation signal spectrum between a GNSS input signal and a receiver local signal; A second spectrum acquisition module, used for weighting the cross-correlation signal spectrum to obtain a weighted spectrum; A third spectrum acquisition module, configured to perform sliding window summation on the weighted spectrum to obtain a sliding window summed spectrum; A fourth spectrum acquisition module, used for normalizing the sliding window sum spectrum to obtain a channel impulse response spectrum; A code phase deviation identification module is used to transform the channel impulse response spectrum into the time domain to obtain a time domain waveform of the channel impulse response, and to identify the time domain waveform through a code phase deviation to obtain a code phase deviation of a direct path, so as to eliminate multipath interference; The step of weighting the cross-correlation signal spectrum to obtain a weighted spectrum includes: The cross-correlation signal spectrum is multiplied by a weighting coefficient of each frequency point to obtain the weighted spectrum, wherein the weighting coefficient of each frequency point is a quotient of a spectrum of a corresponding frequency point in a non-multipath situation and a noise variance of the corresponding frequency point.

8. A device for estimating a channel impulse response spectrum, characterized in that: include: A first spectrum acquisition module, used to acquire a cross-correlation signal spectrum between an input signal and a local signal, or to acquire a cross-correlation signal spectrum between an input signal and a pilot signal; A second spectrum acquisition module, used for weighting the cross-correlation signal spectrum to obtain a weighted spectrum; A third spectrum acquisition module, configured to perform sliding window summation on the weighted spectrum to obtain a sliding window summed spectrum; A fourth spectrum acquisition module, used for normalizing the sliding window sum spectrum to obtain a channel impulse response spectrum; The step of weighting the cross-correlation signal spectrum to obtain a weighted spectrum includes: The cross-correlation signal spectrum is multiplied by a weighting coefficient of each frequency point to obtain the weighted spectrum, wherein the weighting coefficient of each frequency point is a quotient of a spectrum of a corresponding frequency point in a non-multipath situation and a noise variance of the corresponding frequency point.

9. A receiver, characterized in that: The receiver comprises: the multipath interference elimination device as claimed in claim 7.

Images