A signal processing method and system for eliminating the influence of cumulative phase deviation

Abstract

The invention discloses a signal processing method and system for eliminating the influence of cumulative phase deviation, comprising the steps of convolving a sampling sequence of a chirp signal on an I-path and a sampling sequence of a chirp signal on a Q-path with a local chirp sequence to obtain an I-path convolution sequence and a Q-path convolution sequence; performing square sum addition and square root operation on the I-path convolution sequence and the Q-path convolution sequence to obtain an absolute value sequence of a correlation peak and a first timestamp corresponding to a maximum peak; finding a carrier phase deviation and a phase deviation rotation factor corresponding to the first timestamp; multiplying the I-path convolution sequence and the Q-path convolution sequence with the phase deviation rotation factor to assemble into a correlation peak complex sequence; selecting the real part of the correlation peak complex sequence to obtain a first correlation peak real sequence; grouping and accumulating the sampling points in the first correlation peak real sequence to obtain a second correlation peak real sequence, finding the maximum peak in the second correlation peak real sequence to perform multiple interpolation to obtain a high-precision timestamp.

Description

Technical Field

[0001] The present invention relates to the technical field of communication signal processing, and in particular to a signal processing method and system for eliminating the influence of cumulative phase deviation. Background Art

[0002] In modern radar and wireless communication systems, sampling frequency offset mainly comes from the deviation between the actual operating frequency of hardware components such as analog-to-digital converters (ADCs) and their nominal value. This deviation can be caused by a variety of factors such as temperature changes, power supply fluctuations, or device aging.

[0003] When there is a sampling frequency offset, the initial sampling time of each chirp signal will gradually increase at the beginning of each new chirp. Specifically, assuming there is no sampling frequency offset, the starting point of each chirp should be consistent, but due to the presence of sampling frequency offset, each time a new chirp starts, its actual starting sampling point will be slightly offset relative to the ideal position. Over time, this offset will gradually accumulate, causing the initial sampling time of each chirp to be continuously delayed. This cumulative effect directly affects the ability to extract high-resolution time from the current chirp. Summary of the invention

[0004] In order to solve the problems existing in the prior art, the present invention provides a signal processing method and system for eliminating the influence of cumulative phase deviation.

[0005] The technical solution adopted by the present invention is:

[0006] A first aspect of the present application provides a signal processing method for eliminating the influence of cumulative phase deviation, which is applied to a signal processing system for eliminating the influence of cumulative phase deviation, and includes the following contents:

[0007] The wireless signal receiving end samples the I-channel baseband signal and the Q-channel baseband signal to obtain a sampling sequence of the chirp signal on the I-channel and a sampling sequence of the chirp signal on the Q-channel;

[0008] Convolving the sampling sequence of the chirp signal on the I path and the sampling sequence of the chirp signal on the Q path with the local chirp sequence respectively to obtain an I-path convolution sequence and a Q-path convolution sequence;

[0009] Perform square root operation on the I-way convolution sequence and the Q-way convolution sequence, add the squares and then take the square root to obtain the absolute value sequence of the correlation peaks;

[0010] Traversing the absolute value sequence, and finding a first timestamp corresponding to a maximum peak value in the absolute value sequence;

[0011] Calculating an I-way convolution sequence value and a Q-way convolution sequence value corresponding to the first timestamp;

[0012] Based on the I-path convolution sequence value and the Q-path convolution sequence value, a carrier phase deviation and a phase deviation rotation factor are calculated;

[0013] Multiplying the I-way convolution sequence and the Q-way convolution sequence by the phase rotation factor respectively to assemble into a correlation peak complex sequence;

[0014] Select the real part of the correlation peak complex sequence to obtain the first correlation peak real sequence;

[0015] The sampling points in the first correlation peak real number sequence are grouped and accumulated to obtain a second correlation peak real number sequence;

[0016] The maximum peak value in the second correlation peak real number sequence is found, and multiple interpolation is performed on both sides of the maximum peak value in the second correlation peak real number sequence to obtain a high-precision timestamp.

[0017] Preferably, the expression of the I-channel baseband signal is:

[0018]

[0019] The expression of the Q-path baseband signal is:

[0020]

[0021] in, is the carrier signal of the transmitter, is the baseband chirp signal transmitted by the transmitter, is the carrier frequency of the receiver, is the initial phase of the receiver, is the carrier frequency of the transmitter, is the initial phase of the transmitter, is the carrier frequency deviation, is the carrier phase deviation, is the phase angle that varies with time.

[0022] Preferably, the sampling sequence of the chirp signal on the I path is:

[0023]

[0024] The sampling sequence of the chirp signal on the Q path is:

[0025]

[0026] in, is the actual sampling interval of the receiver, Indicates that the chirp received by the receiver is not equal to the chirp sent Integer multiples of the delay time, Indicates the sequence number of the chirp signal. is a random phase.

[0027] Preferably, the baseband chirp signal transmitted by the transmitter is:

[0028]

[0029]

[0030] Where A represents the signal amplitude, k is the slope of the angular frequency changing with time, is the phase angle that varies with time.

[0031] Preferably, the local chirp sequence is:

[0032]

[0033] in, is the sampling interval of the transmitter, Indicates the sequence number of the chirp signal.

[0034] A second aspect of the present application provides a signal processing system for eliminating the influence of cumulative phase deviation, including a receiver, and applying the above-mentioned signal processing method for eliminating the influence of cumulative phase deviation.

[0035] The beneficial effects of the present invention are at least one of the following:

[0036] By calculating and applying the phase deviation rotation factor to correct the initial phase deviation of each chirp, the accumulated phase deviation caused by the sampling frequency deviation is effectively eliminated. This ensures that no additional errors are introduced due to phase inconsistency during the accumulation of multiple chirps.

[0037] By grouping and accumulating multiple chirps in the same frame, the strength of the useful signal is enhanced, while the noise and other interference components are averaged out, significantly improving the signal-to-noise ratio.

[0038] The maximum peak value in the second correlation peak real number sequence is found, and multiple interpolation is performed on both sides of the maximum peak value in the second correlation peak real number sequence, thereby further improving the accuracy of the timestamp. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 This is a flow chart of the method of embodiment 1 of this application DETAILED DESCRIPTION

[0040] The embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0041] Embodiment 1 provides a signal processing method for eliminating the influence of cumulative phase deviation. This method can be applied to extract high-precision timestamps in a ranging and positioning system.

[0042] In this embodiment, the wireless signal receiving end may be a receiver having an ad9361 sampling chip therein, and the wireless signal transmitting end may be a transmitter.

[0043] Step 1: The wireless signal receiving end samples the I-channel baseband signal and the Q-channel baseband signal to obtain a sampling sequence of the chirp signal on the I-channel and a sampling sequence of the chirp signal on the Q-channel.

[0044] In the specific implementation process, before the receiver samples the I-channel baseband signal and the Q-channel baseband signal, it needs to first perform correlation demodulation on the RF signal transmitted by the transmitter, so the transmitted signal is modeled first. The expression of the baseband chirp signal transmitted by the transmitter is:

[0045] (1)

[0046] in:

[0047] (2)

[0048] represents the phase angle change of the chirp, and k is the slope of the angular frequency changing with time.

[0049] Since only the Q part is not 0 and the I part is 0 during transmission, when the transmitting antenna is used When (1) is transmitted with a carrier frequency, the signal expression on the antenna is:

[0050] (3)

[0051] The carrier frequency used by the receiver is , may not be equal to , the initial phase of the receiver is , or it may not be equal to When the receiver is used Multiply , and then low-pass filtering is performed, the signal obtained from the receiver part I is:

[0052] (4)

[0053] Similarly, the receiver uses Multiply , and then low-pass filtering is performed, the signal of the receiver Q part is:

[0054] (5)

[0055] in:

[0056] (6)

[0057] in is the carrier frequency deviation.

[0058] (7)

[0059] represents the carrier phase deviation. From equations (3) and (4), it can be seen that when = 0, it becomes the modeling equation when there is only carrier phase offset; if there is no phase deviation at the same time, that is =0, then 0, , so that the I part completely obtains the same signal as the Q part of the transmitter, but usually , but a random phase uniformly distributed between 0 and 360 degrees. It can be extracted from the received IQ signal.

[0060] Then the sampling frequency deviation is modeled. First, the expression between the sent single chirp and the received single chirp is modeled.

[0061] When t is substituted into equations (1) and (2), The local chirp sequence is:

[0062] (8)

[0063] However, the actual sampling interval of the receiver and the sampling interval of the transmitter Different, will Substituting into equation (1) and equation (2), we get the expression of the sampling sequence of the chirp signal on the I path of the receiver:

[0064] (9)

[0065] The expression of the sampling sequence of the chirp signal on the Q path of the receiver is:

[0066] (10)

[0067] In the above formula Indicates that the chirp received by the receiver is not equal to the chirp sent Integer multiples of the delay time, that is, the high-precision delay time that needs to be extracted; is a random phase.

[0068] In the specific implementation process, if complex numbers are used to express equations (9) and (10), they can be expressed as:

[0069] (11)

[0070] If there is only carrier frequency deviation and phase deviation but no sampling frequency deviation, we only need to replace Replace That is:

[0071] (12)

[0072] The subscript d in the formula means "expected".

[0073] First, let's estimate the size of the various deviations in the above formula. Assuming that the total length of the frame is 51.2μs, the carrier frequency deviation between the transmitter and the receiver and the deviation in the sampling formula are Assuming the carrier frequency of the transmitter is 2.4 GHz, multiply by We get 2400Hz. That is:

[0074] (13)

[0075] The sampling frequency of the receiver is 100MHz, multiplied by We get 100Hz, which means we assume the sampling frequency error is 100Hz.

[0076] Assumptions:

[0077] (14)

[0078] (15)

[0079] That is, assuming the actual receiver sampling frequency Than the sampling frequency of the transmitter Small , in the above assumptions Then we can use Taylor series to calculate and Indicated ,

[0080] (16)

[0081] That is to say Only equivalent to is larger by a small amount, and that amount is When it is very small, it changes linearly with the sampling frequency deviation.

[0082] Substituting formula (16) into formula (11), we obtain:

[0083] (17)

[0084] Comparing equation (17) with equation (12), it can be seen that after considering the sampling frequency deviation, the chirp signal of the first cycle is different from the chirp signal of the first cycle without considering the sampling frequency deviation.

[0085] For example, in formula (17) Compared with the formula (12) Both represent the baseband chirp signal part of the original transmission; when there is a sampling frequency deviation, Relative to There is distortion; when there is no sampling frequency deviation, but there is carrier frequency deviation and carrier phase deviation, it is equivalent to =0, at this time the baseband signal is not distorted.

[0086] For example, in formula (17) Represents a frequency and phase modulation of the baseband signal, in which both the carrier frequency deviation and the sampling frequency deviation can affect the modulation frequency. =0, indicating that there is only carrier frequency deviation, and the modulation frequency is ,when When the modulation frequency is , the phase of modulation is Looking at the initial phase of modulation in equation (12), we find that the magnitude is similar to Exactly the same, that is, only the carrier frequency deviation and the transmitted signal as well as It will affect the initial phase offset of the first chirp, but has nothing to do with the sampling frequency deviation. However, this does not mean that the initial phase of the second, third, and nth chirp received will not change relative to the initial phase of the first chirp. In fact, in the case of multiple subsequent chirps, the initial phase of the chirp will accumulate.

[0087] For another example, from equation (17), it can be seen that the amplitude of each sampling code of the baseband signal caused by the carrier frequency deviation is multiplied by the rotation factor The phase angle is It changes linearly with the number of sampling points in the chirp, that is:

[0088] (18)

[0089] This linear change shows that when multiple chirps are sent continuously, the phenomenon that the phase increases with the increase of n will make the initial phase of the subsequent chirps larger. This increase in initial phase can be estimated. For example, the carrier frequency deviation is 10 -6 , and assuming the carrier is 2.4 GHz, , Hz, so the average of each sampling point will cause the phase to shift by about 10% relative to the previous sampling point. Radians. If you multiply this number by 80, which is the number of sampling points in each chirp, you get 0.012 radians. Converted to degrees, it is 0.6912 degrees. There are 16*3=48 chirps between the first chirp sent and the last chirp, so the difference between the initial phase of the last chirp and the initial phase of the first chirp is 0.6912*48=33.18 degrees. This degree is still relatively large.

[0090] It should be noted that in formula (17) The part represents the effect of sampling frequency deviation on the received signal.

[0091] The sampling time in formula (17) is determined by the normal become:

[0092] (19)

[0093] From this phenomenon, we can see that not only for a single chirp code, as n increases, the position of each sampling moment inside will deviate; but also the baseband initial phase of the second code will shift relative to the baseband initial phase of the first chirp. For example, according to formula (17), the initial sampling moment of the baseband signal of the 0th chirp code is (obtained from n=0). Similarly, according to the conclusion of equation (17), the initial sampling time of the first chirp code will be equal to . And so on, The initial sampling time of the chirp is:

[0094] (20)

[0095] It can be seen from formula (19) that when there is a sampling frequency deviation, the time point of each sampling is longer than the originally expected time point. , then each chirp will go through 80 sampling points, so the initial time of the baseband of the next chirp will be delayed compared to the initial time of the previous chirp Similarly, the initial time of chirp number n is delayed relative to the initial time of chirp number 0. , so the expression of the initial moment of chirp number n is as shown in formula (19).

[0096] This conclusion shows that when there is a sampling frequency deviation, the initial sampling time of each chirp is cumulatively increasing, and this accumulated initial sampling time is exactly what needs to be extracted from the current chirp to represent the high-resolution time content, but the sampling frequency deviation causes this information to change.

[0097] When n=1, Substituting into formula (20), we get: , that is, the precise time extracted from chirp No. 1 (assuming it can be infinitely refined) is relative to the real precise time Too much to say , according to a Here, the electromagnetic wave transmission is 3 meters (the sampling frequency is 100MHz), and the distance deviation is:

[0098] (twenty one)

[0099] Visible when Very small, each chirp relative to its previous chirp high-precision time will have a small deviation, but when When it is larger, such as , the distance deviation caused by the high-precision time difference between adjacent chirps will be equal to 2.4cm.

[0100] Let n in equation (20) be equal to 47 to obtain, , according to a In the calculation, the electromagnetic wave transmission is 3 meters, and the distance deviation will be equal to This deviation turns out to be very small.

[0101] However, due to various factors such as temperature changes, power supply fluctuations or device aging, if , the deviation will be equal to 112.8cm. This deviation is very large.

[0102] Therefore, in order to solve the technical problem that when there is a sampling frequency deviation, the initial sampling moment of each chirp is accumulated and increased, resulting in a change in the accumulated initial sampling moment, after sampling the I-channel baseband signal and the Q-channel baseband signal to obtain the sampling sequence of the chirp signal on the I channel and the sampling sequence of the chirp signal on the Q channel, steps 2-6 need to be performed in sequence.

[0103] Step 2: Convolve the sampling sequence of the chirp signal on the I path and the sampling sequence of the chirp signal on the Q path with the local chirp sequence respectively to obtain an I-path convolution sequence and a Q-path convolution sequence.

[0104] Among them, the I-way convolution sequence is the convolution of formula (9) and formula (8), and its expression is:

[0105] (twenty two)

[0106] The Q-way convolution sequence is the convolution of formula (10) and formula (8), and its expression is:

[0107] (twenty three)

[0108] In a specific implementation process, the convolution function can be directly called to perform a convolution operation.

[0109] Step 3, perform square root operation on the I-way convolution sequence and the Q-way convolution sequence, add the squares and then take the square root operation to obtain the absolute value sequence of the correlation peaks.

[0110] In the specific implementation process, the sum(A^2) function can be called to square and sum the I-way convolution sequence and the Q-way convolution sequence, and then the sqrt function can be called to perform a square root operation.

[0111] As mentioned above, there are 16*3=48 chirps between the first chirp and the last chirp sent, so the initial phase difference between the last chirp and the first chirp is equal to 0.6912*48=33.18 degrees. This degree is still relatively large. However, after performing square root operations on the I-way convolution sequence and the Q-way convolution sequence, this degree difference can be immuned. Immunity means that the position of the correlation peak can be found smoothly without being affected by the carrier phase deviation. And the influence of the initial phase of each chirp.

[0112] Step 4: traverse the absolute value sequence to find the first timestamp corresponding to the maximum peak value in the absolute value sequence.

[0113] Step 5: Calculate an I-way convolution sequence value and a Q-way convolution sequence value corresponding to the first timestamp.

[0114] Step 6: Calculate the carrier phase offset and the phase offset rotation factor based on the I-path convolution sequence value and the Q-path convolution sequence value.

[0115] In the specific implementation process, the function of calculating the angle can be called to calculate the carrier phase deviation and the phase deviation rotation factor. The impact is obtained when capturing the first timestamp corresponding to the first maximum peak and the phase rotation factor j .

[0116] Step 7: Multiply the I-way convolution sequence and the Q-way convolution sequence by the phase rotation factor respectively to assemble them into a correlation peak complex sequence.

[0117] It should be noted that this is equivalent to correcting the phase deviation of the original convolution sequence containing random carrier phase deviation, so that the real part concentrates most of the energy of the convolution sequence. Multiply the convolution complex sequence of each chirp after the first timestamp by the phase deviation rotation factor j , that is, their phase deviations are corrected to near the 0 phase, and the initial phase of the nth chirp is , the phase that should be corrected is , but only corrected , there are still No, because the IQ part of the sampling point of the signal of this period is correlated with the local chirp, so the correlation is equivalent to multiplying the I part and the Q part respectively. and , correct After that, the IQ sequence of the nth chirp is equivalent to being multiplied by and .

[0118] Step 8: Select the real part of the correlation peak complex sequence to obtain the first correlation peak real sequence.

[0119] Step 9: Group and accumulate the sampling points in the first correlation peak real number sequence to obtain a second correlation peak real number sequence.

[0120] Step 10, finding the maximum peak value in the second correlation peak real number sequence, performing multiple interpolation on both sides of the maximum peak value in the second correlation peak real number sequence, and obtaining a high-precision timestamp.

[0121] Embodiment 2 provides a signal processing system for eliminating the influence of cumulative phase deviation, including a receiver, and applies the above-mentioned signal processing method for eliminating the influence of cumulative phase deviation.

[0122] The above-mentioned embodiments only express the specific implementation of the present invention, and the description thereof is relatively specific and detailed, but it cannot be understood as limiting the scope of the present invention. It should be pointed out that for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention.

Claims

1. A signal processing method for eliminating the influence of cumulative phase deviation, applied to a signal processing system for eliminating the influence of cumulative phase deviation, characterized in that: Includes the following: The wireless signal receiving end samples the I-channel baseband signal and the Q-channel baseband signal to obtain a sampling sequence of the chirp signal on the I-channel and a sampling sequence of the chirp signal on the Q-channel; Convolving the sampling sequence of the chirp signal on the I path and the sampling sequence of the chirp signal on the Q path with the local chirp sequence respectively to obtain an I-path convolution sequence and a Q-path convolution sequence; Perform square root operation on the I-way convolution sequence and the Q-way convolution sequence, add the squares and then take the square root to obtain the absolute value sequence of the correlation peaks; Traversing the absolute value sequence, and finding a first timestamp corresponding to a maximum peak value in the absolute value sequence; Calculating an I-way convolution sequence value and a Q-way convolution sequence value corresponding to the first timestamp; Based on the I-path convolution sequence value and the Q-path convolution sequence value, a carrier phase deviation and a phase deviation rotation factor are calculated; Multiplying the I-way convolution sequence and the Q-way convolution sequence by the phase rotation factor respectively to assemble into a correlation peak complex sequence; Select the real part of the correlation peak complex sequence to obtain the first correlation peak real sequence; The sampling points in the first correlation peak real number sequence are grouped and accumulated to obtain a second correlation peak real number sequence; The maximum peak value in the second correlation peak real number sequence is found, and multiple interpolation is performed on both sides of the maximum peak value in the second correlation peak real number sequence to obtain a high-precision timestamp.

2. A signal processing method for eliminating the influence of cumulative phase deviation according to claim 1, characterized in that: The expression of the I-channel baseband signal is: The expression of the Q-path baseband signal is: in, is the carrier signal of the transmitter, is the baseband chirp signal transmitted by the transmitter, is the carrier frequency of the receiver, is the initial phase of the receiver, is the carrier frequency of the transmitter, is the initial phase of the transmitter, is the carrier frequency deviation, is the carrier phase deviation, is the phase angle that varies with time.

3. A signal processing method for eliminating the influence of cumulative phase deviation according to claim 2, characterized in that: The sampling sequence of the chirp signal on the I path is: The sampling sequence of the chirp signal on the Q path is: in, is the actual sampling interval of the receiver, ) indicates that the chirp received by the receiver is not equal to the chirp sent. Integer multiples of the delay time, Indicates the sequence number of the chirp signal. is a random phase.

4. A signal processing method for eliminating the influence of cumulative phase deviation according to claim 2, characterized in that: The baseband chirp signal transmitted by the transmitter is: Where A represents the signal amplitude, k is the slope of the angular frequency changing with time, is the phase angle that varies with time.

5. A signal processing method for eliminating the influence of cumulative phase deviation according to claim 4, characterized in that: The local chirp sequence is: in, is the sampling interval of the transmitter, Indicates the sequence number of the chirp signal.

6. A signal processing system for eliminating the influence of cumulative phase deviation, comprising a receiver, characterized in that: A signal processing method for eliminating the influence of cumulative phase deviation as described in any one of claims 1 to 5 is applied.

Images