A method and device for tracking carrier frequency offset

Abstract

The present application discloses a carrier frequency offset tracking method and device, which receives multiple Wi-Fi frames containing custom flags, obtains a phase sequence and a receiving time sequence, determines a training segment and a tracking segment according to the custom flag, then calculates a phase difference sequence of adjacent frames of the training segment according to the phase sequence, calculates a differential time series of the training segment according to the time series, performs first-order linear fitting to obtain a CFO estimate of the training segment, and finally initializes an unscented Kalman filter according to the CFO estimate of the training segment, predicts and corrects the CFO estimate of the tracking segment through the initialized unscented Kalman filter and the CFO estimate of the training segment, outputs a CFO long-term measurement value sequence, and performs long-term tracking of the carrier frequency offset of channel state information. The unscented Kalman filter can also achieve the effects of high accuracy, simple implementation, strong robustness, and applicability to a wide range of application scenarios.

Description

Technical Field

[0001] The present application relates to the field of wireless communication technology, and in particular to a method and device for tracking carrier frequency offset. Background Art

[0002] Wi-Fi sensing has the advantages of indoor ubiquity, privacy protection, fine-grained perception, and environmental robustness. High-precision channel state information (CSI) measurement is critical to Wi-Fi's perception capabilities, but the carrier frequency offset (CFO) caused by the frequency difference between the Tx and Rx local oscillators (LO) will accumulate phase errors and affect the accuracy of CSI measurements. In addition, LO is also very sensitive to temperature changes. Long-term operation of the device will cause the LO temperature to rise, resulting in CFO drift, further increasing the phase error of CSI measurement. Therefore, CFO has a serious impact on the development of Wi-Fi sensing technology based on CSI phase, hindering the development of new generation Wi-Fi technologies such as wireless phase synchronization, Wi-Fi phased array, and Wi-Fi imaging.

[0003] Although there are some studies on solving the CFO estimation problem in the prior art, there are often many problems. In a high-speed mobile communication system, if the pilot transmission period is too long, the cumulative error of the frequency offset may be large; if the pilot transmission period is too short, the length of the pilot sequence may not be enough to support accurate frequency offset estimation. These defects make it difficult for the prior art to be applied to a wide range of Wi-Fi sensing practices. More importantly, Wi-Fi sensing requires long-term tracking and estimation of CFO to calibrate the error, and short-term CFO estimation cannot meet the needs of Wi-Fi sensing.

[0004] Therefore, how to conduct research on the long-term tracking of CFO has become one of the many problems that technicians in this field need to solve. Summary of the invention

[0005] The purpose of the present application is to provide a method and device for tracking a carrier frequency offset in order to overcome the existing technical defects, so as to perform long-term tracking of the carrier frequency offset of channel state information.

[0006] The purpose of this application is achieved through the following technical solutions:

[0007] In a first aspect, the present application proposes a method for tracking a carrier frequency offset, which is applied to an RX receiving end, and the method includes:

[0008] Receive multiple Wi-Fi frames containing custom flags;

[0009] Obtaining a phase sequence and a receiving time sequence according to the Wi-Fi frame;

[0010] Determine the training segment and the tracking segment in the Wi-Fi frame according to the custom flag bit;

[0011] The phase difference sequence of adjacent frames in the training segment is calculated according to the phase sequence, the differential time series of the training segment is calculated according to the time series, and the phase difference sequence and the differential time series are fitted by a first-order linear fitting to obtain the CFO estimation value of the training segment;

[0012] The unscented Kalman filter is initialized according to the CFO estimate of the training segment, and the CFO estimate of the tracking segment is predicted and corrected by the initialized unscented Kalman filter and the CFO estimate of the training segment, and a CFO long-term measurement value sequence is output.

[0013] In a possible implementation, the step of predicting and correcting the CFO estimate of the tracking segment by using the initialized unscented Kalman filter and the CFO estimate of the training segment to output a CFO long-term measurement value sequence includes:

[0014] The state transfer function is constructed based on the CFO estimation value and phase difference sequence of the training segment and the differential time series of the tracking segment;

[0015] Generate Sigma points based on the initial values ​​of the covariance matrix and the initial values ​​of the state vector;

[0016] Using the state transfer function to predict the state of the Sigma point, a k+1-step priori state vector is obtained;

[0017] Calculate the posterior state and the covariance of the posterior state using the k+1 step prior state vector;

[0018] The CFO estimate of the posterior state is used as the CFO estimate of the next cycle to predict and correct the CFO estimate of the tracking segment until all tracking segments are executed and the CFO long-term measurement value sequence is output.

[0019] In a possible implementation, the step of calculating the a posteriori state and the covariance of the a posteriori state using the k+1-step priori state vector includes:

[0020] The predicted state of the Sigma point is calculated using the k+1-step prior state vector and the observation function, and the mean, variance and cross-correlation equation of the predicted state are calculated;

[0021] Calculating the Kalman gain according to the cross-correlation equation and the covariance matrix, and correcting the initial value of the covariance matrix and the time series of the training segment by the Kalman gain to obtain the posterior state and the covariance of the posterior state;

[0022] In a possible implementation, the steps of calculating a phase difference sequence of adjacent frames of a training segment according to a phase sequence, calculating a differential time series of the training segment according to a time series, and performing a first-order linear fitting on the phase difference sequence and the differential time series to obtain a CFO estimation value of the training segment include:

[0023] The phase difference sequence of adjacent frames of the training segment is calculated according to the phase sequence, and the differential time series of the training segment is calculated according to the time series;

[0024] Unwrapping the phase difference sequence to obtain an unwrapped phase difference sequence;

[0025] Perform a first-order linear fit on the differential time series of the training segment and the phase difference series after unwrapping to obtain the slope value of the fitting function;

[0026] A CFO estimate for the training segment is calculated based on the slope value.

[0027] In one possible implementation, the CFO estimate The calculation formula is: k is the slope value of the fitting function.

[0028] In a possible implementation manner, the state transfer function includes a CFO change rate, a CFO current value, an unwrapped phase accumulation amount, and an unwrapped phase accumulation amount.

[0029] In a possible implementation, the step of obtaining a phase sequence and a receiving time sequence according to the Wi-Fi frame includes:

[0030] measuring channel state information of the Wi-Fi frame,

[0031] Perform DC subcarrier interpolation on the channel state information to obtain a phase sequence;

[0032] A reception time series is constructed based on the reception time of the Wi-Fi frames.

[0033] In a second aspect, the present application proposes a carrier frequency offset tracking device, which is applied to an RX receiving end and includes:

[0034] A receiving module, used for receiving multiple Wi-Fi frames containing custom flags;

[0035] A data processing module, used to obtain a phase sequence and a receiving time sequence according to the Wi-Fi frame;

[0036] A training-tracking segment determination module, used to determine a training segment and a tracking segment in a Wi-Fi frame according to the custom flag bit;

[0037] An estimation value calculation module is used to calculate the phase difference sequence of adjacent frames of the training segment according to the phase sequence, calculate the differential time series of the training segment according to the time series, and perform first-order linear fitting on the phase difference sequence and the differential time series to obtain the CFO estimation value of the training segment;

[0038] The output module is used to initialize the unscented Kalman filter according to the CFO estimation value of the training segment, predict and correct the CFO estimation value of the tracking segment through the initialized unscented Kalman filter and the CFO estimation value of the training segment, and output the CFO long-term measurement value sequence.

[0039] In a third aspect, the present application further proposes a computer device, comprising a processor and a memory, wherein the memory stores a computer program, and the computer program is loaded and executed by the processor to implement the carrier frequency offset tracking method as described in any one of the first aspects.

[0040] In a fourth aspect, the present application further proposes a computer-readable storage medium, in which a computer program is stored. The computer program is loaded and executed by a processor to implement the carrier frequency offset tracking method as described in any one of the first aspects.

[0041] The above-mentioned main scheme of the present application and its further options can be freely combined to form multiple schemes, all of which are schemes that can be adopted and claimed for protection in the present application; and in the present application, (non-conflicting options) options and other options can also be freely combined. After understanding the scheme of the present application, those skilled in the art can understand that there are multiple combinations based on the prior art and common knowledge, all of which are technical schemes to be protected by the present application, and they are not exhaustively listed here.

[0042] The present application discloses a carrier frequency offset tracking method and device, which first receives multiple Wi-Fi frames containing custom flags, then obtains a phase sequence and a receiving time sequence according to the Wi-Fi frames, determines a training segment and a tracking segment in the Wi-Fi frame according to the custom flag, then calculates a phase difference sequence of adjacent frames of the training segment according to the phase sequence, calculates a differential time series of the training segment according to the time series, performs first-order linear fitting to obtain a CFO estimate of the training segment, and finally initializes an unscented Kalman filter according to the CFO estimate of the training segment, predicts and corrects the CFO estimate of the tracking segment through the initialized unscented Kalman filter and the CFO estimate of the training segment, outputs a CFO long-term measurement value sequence, and performs long-term tracking of the carrier frequency offset of channel state information. The unscented Kalman filter can also achieve higher accuracy, simple implementation, good robustness, and wide applicability to practical application scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1A flow chart of a carrier frequency offset tracking method proposed in an embodiment of the present application is shown.

[0044] Figure 2 Another flowchart of the carrier frequency offset tracking method proposed in an embodiment of the present application is shown.

[0045] Figure 3 The residual diagram between the predicted CFO and the actual measurement for different tracking methods proposed in the embodiments of the present application is shown. DETAILED DESCRIPTION

[0046] The following describes the embodiments of the present application through specific examples, and those skilled in the art can easily understand other advantages and effects of the present application from the contents disclosed in this specification. The present application can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present application. It should be noted that the following embodiments and features in the embodiments can be combined with each other without conflict.

[0047] Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making any creative work shall fall within the scope of protection of this application.

[0048] In the existing technology, although CFO can be effectively estimated, there are some problems, such as sensitivity to noise and large amount of calculation of autocorrelation function, strict requirements on the selection and transmission of pilots, and in high-speed mobile communication systems, if the pilot transmission period is too long, the cumulative error of frequency offset may be large; if the pilot transmission period is too short, the length of the pilot sequence may not be enough to support accurate frequency offset estimation. Wi-Fi sensing requires long-term tracking and estimation of CFO to calibrate the error, and short-term CFO estimation cannot meet the needs of Wi-Fi sensing.

[0049] Unscented Transform (UT) uses a fixed number of parameters to approximate a Gaussian distribution. Its implementation principle is: select some points in the original distribution according to a certain rule so that the mean and covariance of these points are equal to the mean and covariance of the original state distribution; substitute these points into the nonlinear function, and obtain the corresponding nonlinear function value point set, through which the mean and covariance of the transformation can be obtained. For any nonlinear system, when the Gaussian state trace is transmitted through the nonlinear system, the posterior mean and covariance accurate to the third order moment can be obtained by using this set of sampling points. The characteristic of unscented transformation is to approximate the probability density distribution of nonlinear functions instead of nonlinear functions. Even if the system model is complex, it does not increase the difficulty of algorithm implementation; and the accuracy of the statistics of the obtained nonlinear function can reach the third order; in addition, it does not need to calculate the Jacobian matrix and can handle non-differentiable nonlinear functions. Unscented Kalman Filter (UKF) is a combination of Unscented Transform (UT) and standard Kalman filter system.

[0050] Therefore, in order to solve the above problems, an embodiment of the present application proposes a method and device for tracking carrier frequency offset. Through a first-order linear fitting method combined with unscented Kalman filtering technology, long-term tracking of the carrier frequency offset of channel state information is achieved, which is described in detail below.

[0051] Please refer to Figure 1 , Figure 1 A carrier frequency offset tracking method proposed in an embodiment of the present application is shown. The method is applied to an RX receiving end and includes the following steps:

[0052] S100: Receive multiple Wi-Fi frames containing custom flags.

[0053] The RX receiving end corresponds to the Tx transmitting end. The RX receiving end and the Tx transmitting end work under the 2.4 / 5 / 6GHz frequency band, bandwidth bw (20, 40, 80, 160, 320MHz), protocol format (IEEE 802.11a / g / n / ac / ax / be), number of spatial streams s (1-n) and MCS rate supported by all Wi-Fi standards. For example, the transmitting end fc is 6GHz, bw is 160MHz, and the transmission format is IEEE802.11ax protocol.

[0054] The RX receiving end and the Tx transmitting end are in the same Wi-Fi channel and working bandwidth. For example, both Tx and Rx work in Wi-Fi standard channel 36 (carrier frequency 5180MHz) with a bandwidth of 20MHz.

[0055] The Tx transmitter continuously sends a total of N Wi-Fi frames containing flags to the RX receiver. The frame can be divided into a training segment and a tracking segment according to the custom flag stage.

[0056] Combined with the basic principles of wireless communication, first set the signal frequency sent by the Tx transmitter to X, the signal frequency received by the RX receiver to Y, and the frequency response of the transmission channel to H. The mathematical relationship of the signal is Y = HX + N, where the noise term N represents the random interference in the environment. When the receiving end has a known reference signal X, the channel frequency response H, also known as CSI, can be calculated through the channel estimation algorithm, which is the solution to the mathematical relationship between the three.

[0057] S200 : Obtain a phase sequence and a receiving time sequence according to a Wi-Fi frame.

[0058] In a possible embodiment, step S200 further includes the following sub-steps:

[0059] Measuring channel state information of Wi-Fi frames;

[0060] Perform DC subcarrier interpolation on the channel state information to obtain a phase sequence;

[0061] A reception time series is constructed based on the reception time of the Wi-Fi frames.

[0062] In the Wi-Fi protocol (802.11 / a / g / n / ac / ax protocol) based on Orthogonal Frequency Division Multiplexing (OFDM), the channel state information is represented as a 3D matrix Where N SC Represents the number of subcarriers, N STS is the Tx spatial stream, N Rx is the number of Rx antennas. According to the Wi-Fi protocol standard, the subcarrier numbers are: The DC subcarrier is not used, and the bandwidth bw can be 20 / 40 / 80 / 160 / 320MHz. SC The values ​​are 64, 128, 256, 512, and 1024. For example, in the 802.11n protocol with a bandwidth of 20 MHz, the subcarrier number sequence can be expressed as K = [-28, -27, ..., -1, 1, ..., 27, 28].

[0063] The RX receiving end interpolates the measured channel state information with the DC subcarrier. After interpolation, the subcarrier index sequence is expressed as: K +DCIf the Intel AX210 network card is used as the RX receiving end, under the conditions of IEEE 802.11n protocol, 5.2GHz operating frequency, and 40MHz bandwidth, the Tx transmitter uses a single spatial stream to transmit, and the Rx end uses dual antennas to receive, then the size of the measured channel frequency response is 114×1×2, and the first dimension of the channel frequency response is the subcarrier column, whose subcarrier number index is (-57,-56,…,-1,1,…,56,57). After DC bit carrier interpolation, 115 subcarriers are obtained (set DC to 0), and its subcarrier number index becomes (-57,-56,…,-1,0,1,…,56,57), and the channel frequency response matrix becomes 115×1×2.

[0064] After interpolation, take The phase of any spatial stream in Then use Take out the phase value θ of the DC carrier DC Since the DC subcarrier is only affected by the CFO error, the CSI of the DC subcarrier can be expressed as where f cfo is the carrier frequency offset, and T is the unit receiving time interval. After interpolation, the phase sequence of the DC subcarrier contained in each frame of CSI can be obtained:

[0065] S300: Determine a training segment and a tracking segment in a Wi-Fi frame according to a custom flag bit.

[0066] The RX receiving end determines the flag bit, distinguishes the training segment from the tracking segment through the flag bit, and extracts the phase sequence of the corresponding DC subcarrier, that is, in, Considering the situation where there is frame loss during the communication between the transmitter and receiver, but not all frames in a certain segment are lost (such as all training segment frames are lost), the Wi-Fi frame sent by the Tx transmitter contains two consecutive and identical Wi-Fi segment frames, and the DC position subcarrier phase sequence corresponding to each segment frame can be obtained according to the above process. and Similarly, the flag bit is used to distinguish the training segment frame reception time sequence T training The tracking segment frame receiving time series T tracking .

[0067] The training segment and the tracking segment are processed differently. The training segment is used to estimate CFO. The CFO estimation value of the training segment is used to calibrate the CFO estimation value of the tracking segment frame to achieve long-term CFO tracking.

[0068] S400, calculating a phase difference sequence of adjacent frames of a training segment according to a phase sequence, calculating a differential time series according to a time series, and performing a first-order linear fitting on the phase difference sequence and the differential time series to obtain a CFO estimation value of the training segment.

[0069] The calculation formula for calculating the phase difference sequence of adjacent frames in the training segment based on the phase sequence is: The calculation formula for the differential time series of the training segment is calculated based on the time series:

[0070] In a possible implementation, a phase difference sequence of adjacent frames in a training segment is calculated according to a phase sequence, and a differential time series is calculated according to a time series;

[0071] Unwrapping the phase difference sequence to obtain an unwrapped phase difference sequence;

[0072] Perform a first-order linear fit on the differential time series and the phase difference series after unwrapping to obtain the slope value of the fitting function;

[0073] Calculate the CFO estimate for the training segment based on the slope value.

[0074] The phase difference sequence of adjacent frames in the training segment Perform phase unwrapping operation to obtain the phase difference sequence after unwrapping Difference time series As the independent variable T, the phase difference sequence after unwinding A first-order linear fit is performed on the dependent variable Θ(T) to obtain the slope k of the fitting function.

[0075] We can also use the difference time series T i As the independent variable T, the phase difference sequence ΔΘ i,unwrap As the dependent variable, a first-order linear fitting is performed to obtain a fitting function of Y(T)=k·T+b, and the slope k can be obtained, where k is the rate of change of the carrier frequency.

[0076] The estimated value of the carrier frequency offset CFO of the training segment is calculated based on the slope k of the fitting function. k is the slope value of the fitting function.

[0077] S500, initializing an unscented Kalman filter according to the CFO estimate, combining the initialized unscented Kalman filter with the CFO estimate of the training segment to predict and correct the CFO estimate of the tracking segment, and outputting a CFO long-term measurement value sequence.

[0078] The unscented Kalman filter needs to be initialized before predicting the CFO estimate of the corrected tracking segment.

[0079] In a possible embodiment, the step of outputting a CFO long-term measurement value sequence includes:

[0080] The state transfer function is constructed based on the CFO estimation value and phase difference sequence of the training segment and the differential time series of the tracking segment;

[0081] Generate Sigma points based on the initial values ​​of the covariance matrix and the initial values ​​of the state vector;

[0082] The state transfer function is used to predict the state of the Sigma point and obtain the k+1-step prior state vector;

[0083] Calculate the posterior state and the covariance of the posterior state using the k+1 step prior state vector;

[0084] The CFO estimate of the posterior state is used as the CFO estimate of the next cycle to predict and correct the CFO estimate of the tracking segment until all tracking segments are executed and the CFO long-term measurement value sequence is output.

[0085] First, construct the tracking segment differential time series Δt based on the time series T received by the receiving end tracking :Δt tracking (k)=T(k+1)-T(k), where k represents the kth frame belonging to the tracking segment received by the receiving end.

[0086] The state transition function includes the CFO change rate ω cfo 、CFO current value f cfo , the phase accumulation value θ without unwrapping wrap , the phase accumulation of the unwrapped unwrap , construct the state vector matrix X according to the state transfer function:

[0087]

[0088] Among them, N represents the number of Wi-Fi frames received by the receiver, and the four dimensions of the state vector matrix X represent the state information selected for the CFO long-term tracking problem.

[0089] The initial values ​​of the four dimensions of the state variables are: cfo (0)=0, ( Indicates the phase sequence of the DC subcarrier tracking segment The first phase value in express Result after unwinding).

[0090] According to the process noise variance Q of each state variable i =[50,500,0,0], construct the process noise function w i : Among them, i represents the subscript corresponding to the different dimensions of the state variable, RANDOM represents the generation of a random number that obeys the standard normal distribution, and then constructs the observation noise function v: R is the observed variance.

[0091] Construct the state transfer equation F based on the transfer relationship between state vectors:

[0092]

[0093] Among them, X i represents the predicted value designed for the i-th dimension of the state vector, w i (i=1,2,3,4) represents the process noise of the i-th dimension of the state vector, k represents the state of the k-th frame of the corresponding state information, and unwrap() represents unwrapping the phase.

[0094] The unwrapped phase accumulation θ in the state vector wrap The measured phase sequence Θ DC Observe and construct the corresponding observation sequence Z(k): Z(k)=H[Θ DC (k),v]=Θ DC (k)+v, where Θ DC (k) is the measured phase of subcarrier zero, v represents the observation noise, and the observation noise R and observation function G are also constructed. The number of Wi-Fi frames received by the receiver is N.

[0095] According to the initialized system state X(k|k) and covariance P(k|k), 2n+1 Sigma points are generated (n represents the dimension of the state variable X):

[0096]

[0097] In the formula, λ=3-n is usually taken, and λ is an empirical formula: represents the i-th column of the covariance matrix, where P refers to the covariance of the current state.

[0098] Arrange it into: Among them, X(k|k) represents the last 2n+1 Sigma points, represents the mean of the initial state, P(k|k) represents the covariance of the initial state, and About This point is symmetrical, and λ can determine the distance between the surrounding 2n Sigma points and the center X(k|k).

[0099] According to the Sigma point and the state transition equation F, calculate the k+1 step prior state X of these Sigma points (i) (k+1|k) and its mean and variance.

[0100] First calculate X (i) The corresponding weight ω (i) :

[0101]

[0102] Take λ=3-n, where λ is an empirical formula.

[0103] According to ω (i) The mean and covariance matrix of the k+1-step prior state of the system state quantity are obtained, and the calculation formula is as follows:

[0104]

[0105] In a possible embodiment, the step of calculating the a posteriori state and the covariance of the a posteriori state using the k+1-step priori state vector includes:

[0106] The predicted state of the Sigma point is calculated using the k+1-step prior state vector and the observation function, and the mean, variance and cross-correlation equation of the predicted state are calculated;

[0107] The Kalman gain is calculated according to the cross-correlation equation and the covariance matrix, and the initial value of the covariance matrix and the time series of the training segment are corrected by the Kalman gain to obtain the posterior state and the covariance of the posterior state.

[0108] According to the previous step, we get the mean of the k+1 step prior state And the covariance matrix P(k+1|k), again using the unscented transformation, generates new 2n+1 Sigma points, the calculation formula is as follows:

[0109]

[0110] Substitute the Sigma point set into the observation equation and calculate the k+1-step predicted state Z of 2n+1 Sigma points. (i) (k+1|k): Z (i) (k+1|k)=H[X (i) (k+1|k)]. According to the predicted state Z (i) (k+1|k), using the weight ω (i) Calculate the mean of the observed value k+1 step forward The covariance matrix S(k+1|k) is used to calculate the cross-correlation equation J(k+1|k). The calculation formula is:

[0111]

[0112] The Kalman gain K(k+1) is calculated based on the cross-correlation function J(k+1|k) and the covariance matrix S(k+1|k): K(k+1)=J(k+1|k)·S -1 (k+1|k), and use the Kalman gain K(k+1) to get the posterior state X(k+1|k+1) and covariance P(k+1|k+1):

[0113]

[0114] At this time, the second dimension in the posterior state X(k+1|k+1) is used as the CFO estimate (CFO current tracking result). Repeat the above steps after constructing the process noise function until all tracking segments are completed. At this time, the tracking sequence of the state variable CFO by the unscented Kalman filter is the target CFO tracking sequence.

[0115] For a better understanding of this application, please refer to Figure 2 , Figure 2 Another flow chart of the carrier frequency offset tracking method proposed in the embodiment of the present application is shown, comprising the following steps:

[0116] S1. The transmitter continuously sends a Wi-Fi frame sequence containing a stage flag. The frame sequence is divided into a training segment and a tracking segment.

[0117] S2. The receiving end obtains the CSI, receiving time and phase information of each frame.

[0118] S3. Determine the stage flag. If the stage flag indicates a training segment, execute S4. If the stage flag indicates a tracking segment, execute S7.

[0119] S4. Perform a first-order linear fitting based on the DC subcarrier phase and the receiving time to estimate the CFO.

[0120] S5. Determine whether the training segment is finished. If finished, execute S6; if not finished, execute S4.

[0121] S6. Get the final CFO estimate of the training segment As the initial state of the unscented Kalman filter.

[0122] S7. Initialize the unscented Kalman filter according to the CFO estimation value of the training segment, and use the receiving time series T to calculate the time difference between the current receiving time and the previous receiving time.

[0123] S8. Construct a state vector and a state transfer function including the CFO change rate, CFO and measured phase according to the CFO estimated value, subcarrier phase No. 0 and time difference of the training segment frame.

[0124] S9. Generate a Sigma point set using the initial state vector and covariance matrix.

[0125] S10. For each Sigma point, the state is predicted by the state transfer function F to obtain the prior state and calculate the corresponding mean and covariance.

[0126] S11. Use the prior states and observation functions of these Sigma points to calculate the predicted state of each Sigma point and calculate the mean and covariance of the predicted state.

[0127] S12. Calculate the cross-correlation function J(k+1|k), obtain the Kalman gain, and obtain the posterior state X(k+1|k+1) and covariance P(k+1|k+1).

[0128] S13, determine whether the tracking segment is finished, if the tracking segment is not finished, execute S9, if it is finished, execute S14.

[0129] S14. Obtain the CFO long-term tracking sequence.

[0130] Figure 3 The residual diagram between the predicted CFO and the measured CFO for different tracking methods proposed in the embodiment of the present application is shown. The CFO tracking is predicted by constant CFO tracking and the CFO tracking is predicted by UKF. The CFO residuals between the two predicted CFOs and the measured CFOs are compared. UKF has better robustness than EKF.

[0131] Compared with the prior art, the embodiments of the present application have the following beneficial effects:

[0132] First, compared with the Extended Kalman Filter (EKF), UKF has higher estimation accuracy. EKF relies on the first-order Taylor expansion of the system, which introduces linearization errors. UKF approximates the nonlinear system by selecting a set of carefully selected sampling points (called Sigma points), which makes the estimation accuracy of UKF relatively higher.

[0133] Second, UKF has better robustness than EKF. Because EKF relies on the Jacobian matrix of the system, when these matrices are close to singularity, the performance of EKF will be affected. UKF does not need to calculate the Jacobian matrix, so it is more stable in this regard.

[0134] Third, the implementation of UKF is relatively simple, and there is no need to solve the Jacobian matrix, only the traceless transformation needs to be calculated. This makes UKF easier to implement and debug in practical applications.

[0135] A possible implementation of a carrier frequency offset tracking device is provided below, which is used to execute the various execution steps and corresponding technical effects of the carrier frequency offset tracking method shown in the above embodiment and possible implementation. The device is applied to an RX receiving end, including:

[0136] A receiving module, used for receiving multiple Wi-Fi frames containing custom flags;

[0137] A data processing module, used for obtaining a phase sequence and a receiving time sequence according to a Wi-Fi frame;

[0138] A training-tracking segment determination module, used to determine the training segment and tracking segment in the Wi-Fi frame according to a custom flag bit;

[0139] An estimation value calculation module is used to calculate the phase difference sequence of adjacent frames of the training segment according to the phase sequence, calculate the differential time series of the training segment according to the time series, and perform first-order linear fitting on the phase difference sequence and the differential time series to obtain the CFO estimation value of the training segment;

[0140] The output module is used to initialize the unscented Kalman filter according to the CFO estimation value of the training segment, predict and correct the CFO estimation value of the tracking segment through the initialized unscented Kalman filter and the CFO estimation value of the training segment, and output the CFO long-term measurement value sequence.

[0141] This preferred embodiment provides a computer device, which can implement the steps in any embodiment of the carrier frequency offset tracking method provided in the embodiments of the present application. Therefore, the beneficial effects of the carrier frequency offset tracking method provided in the embodiments of the present application can be achieved. Please refer to the previous embodiments for details and will not be repeated here.

[0142] A person skilled in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructions, or by controlling related hardware through instructions, and the instructions can be stored in a computer-readable storage medium and loaded and executed by a processor. To this end, an embodiment of the present application provides a storage medium, in which a plurality of instructions are stored, and the instructions can be loaded by a processor to execute the steps of any embodiment of the carrier frequency offset tracking method provided in the embodiment of the present application.

[0143] The storage medium may include: a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0144] Since the instructions stored in the storage medium can execute the steps in any carrier frequency offset tracking method embodiment provided in the embodiments of the present application, the beneficial effects that can be achieved by any carrier frequency offset tracking method provided in the embodiments of the present application can be achieved. Please refer to the previous embodiments for details and will not be repeated here.

[0145] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present application should be included in the protection scope of the present application.

Claims

1. A method for tracking a carrier frequency offset, characterized in that: The method is applied to an RX receiving end, and the method comprises: Receive multiple Wi-Fi frames containing custom flags; Obtaining a phase sequence and a receiving time sequence according to the Wi-Fi frame; Determine the training segment and the tracking segment in the Wi-Fi frame according to the custom flag bit; The phase difference sequence of adjacent frames of the training segment is calculated according to the phase sequence, and the differential time series of the training segment is calculated according to the time series; Unwrapping the phase difference sequence to obtain an unwrapped phase difference sequence; Perform a first-order linear fit on the differential time series of the training segment and the phase difference series after unwrapping to obtain the slope value of the fitting function; Calculating a CFO estimate value of the training segment according to the slope value; Initialize the unscented Kalman filter according to the CFO estimation value of the training segment, and construct the state transfer function according to the CFO estimation value of the training segment, the phase difference sequence and the differential time series of the tracking segment; Generate Sigma points based on the initial values ​​of the covariance matrix and the initial values ​​of the state vector; Using the state transfer function to predict the state of the Sigma point, a k+1-step priori state vector is obtained; The predicted state of the Sigma point is calculated using the k+1-step prior state vector and the observation function, and the mean, variance and cross-correlation equation of the predicted state are calculated; Calculating the Kalman gain according to the cross-correlation equation and the covariance matrix, and correcting the initial value of the covariance matrix and the time series of the training segment by the Kalman gain to obtain the posterior state and the covariance of the posterior state; The CFO estimate of the posterior state is used as the CFO estimate of the next cycle to predict and correct the CFO estimate of the tracking segment until all tracking segments are executed and the CFO long-term measurement value sequence is output.

2. The carrier frequency offset tracking method according to claim 1, wherein: CFO Estimates The calculation formula is: k is the slope value of the fitting function.

3. The carrier frequency offset tracking method according to claim 1, wherein: The state transfer function includes a CFO change rate, a CFO current value, an unwrapped phase accumulation amount, and an unwrapped phase accumulation amount.

4. The carrier frequency offset tracking method according to claim 1, wherein: The step of obtaining a phase sequence and a receiving time sequence according to the Wi-Fi frame includes: Measuring channel state information of the Wi-Fi frame; Perform DC subcarrier interpolation on the channel state information to obtain a phase sequence; A reception time series is constructed based on the reception time of the Wi-Fi frames.

5. A carrier frequency offset tracking device, characterized in that: The device is applied to an RX receiving end, and includes: A receiving module, used for receiving multiple Wi-Fi frames containing custom flags; A data processing module, used to obtain a phase sequence and a receiving time sequence according to the Wi-Fi frame; A training-tracking segment determination module, used to determine a training segment and a tracking segment in a Wi-Fi frame according to the custom flag bit; An estimation value calculation module is used to calculate a phase difference sequence of adjacent frames of a training segment according to a phase sequence, and to calculate a differential time sequence of the training segment according to a time sequence; Unwrapping the phase difference sequence to obtain an unwrapped phase difference sequence; Perform a first-order linear fit on the differential time series of the training segment and the phase difference series after unwrapping to obtain the slope value of the fitting function; Calculating a CFO estimate value of the training segment according to the slope value; An output module is used to initialize the unscented Kalman filter according to the CFO estimation value of the training segment, and to construct a state transfer function according to the CFO estimation value of the training segment, the phase difference sequence, and the differential time series of the tracking segment; Generate Sigma points based on the initial values ​​of the covariance matrix and the initial values ​​of the state vector; Using the state transfer function to predict the state of the Sigma point, a k+1-step priori state vector is obtained; The predicted state of the Sigma point is calculated using the k+1-step prior state vector and the observation function, and the mean, variance and cross-correlation equation of the predicted state are calculated; Calculating the Kalman gain according to the cross-correlation equation and the covariance matrix, and correcting the initial value of the covariance matrix and the time series of the training segment by the Kalman gain to obtain the posterior state and the covariance of the posterior state; The CFO estimate of the posterior state is used as the CFO estimate of the next cycle to predict and correct the CFO estimate of the tracking segment until all tracking segments are executed and the CFO long-term measurement value sequence is output.

6. A computer device, characterized in that: The computer device includes a processor and a memory, wherein a computer program is stored in the memory, and the computer program is loaded and executed by the processor to implement the carrier frequency offset tracking method as described in any one of claims 1-4.

7. A computer-readable storage medium, characterized in that: The storage medium stores a computer program, and the computer program is loaded and executed by a processor to implement the carrier frequency offset tracking method as described in any one of claims 1-4.

Images