Frequency Fast Recovery Method for Wireless Internet of Things Communication and Sensing

Abstract

The present invention discloses a method for rapid frequency recovery for wireless Internet of Things communication and sensing, which mainly solves the problems in the prior art of poor frequency offset estimation accuracy under short PN, difficulty in balancing the estimation range and accuracy, and high complexity. The implementation scheme is as follows: perform coarse frequency offset estimation and coarse frequency offset compensation on the QPSK complex baseband preamble sequence in sequence; use the complex baseband preamble sequence after coarse frequency offset compensation to perform fine frequency offset estimation; calculate the total estimated value using the coarse frequency offset estimated value and the fine frequency offset estimated value; use the total estimated value to perform frequency offset compensation on the signal data part to obtain the recovered signal. When performing frequency offset estimation, the present invention adopts two rounds of estimation, namely coarse estimation and fine estimation, to improve the frequency offset estimation accuracy of short PN. By adopting different estimation intervals for the two rounds of estimation, the maximum frequency estimation range is ensured. By adopting a combination of phase decision and rotation, a non-linear operation structure is avoided, and the complexity of frequency recovery is reduced. It can be used for high-speed and low-latency Internet of Things communication.

Description

Technical Field

[0001] The present invention belongs to the technical field of wireless communication, and particularly relates to a method for rapid frequency recovery, which can be used for high-speed and reliable communication in the Internet of Things (IoT) scenario. Background Art

[0002] With the booming development of wireless technology and the Internet of Things, wireless IoT application scenarios such as smart home, smart healthcare, and intelligent transportation have emerged. These application scenarios connect a large number of wireless devices and achieve intelligent communication and perception through real-time signal transmission and interaction between devices. Therefore, wireless signals are required to be quickly captured for time-frequency synchronization and access to meet the requirements of low latency and high density in IoT applications. At the same time, in order to ensure the correct reception and parsing of wireless signals, the preamble PN header is used in the frame structure of the wireless communication system to achieve synchronization and carrier frequency recovery.

[0003] Accurate carrier frequency recovery is an important prerequisite for ensuring the correct synchronization and access of wireless signals. During the transmission of wireless signals, due to unstable factors such as crystal oscillator errors and Doppler frequency shift at the transceiver ends, there will be a large deviation between the carrier frequencies of the signals at the receiving end and the transmitting end. In order to correctly demodulate the received signal, it is necessary to estimate and compensate the frequency offset using the PN header in the frame structure. In short-distance and high-density communication scenarios, in order to accelerate the synchronization speed, reduce resource overhead, and improve spectrum utilization, a shorter PN header is often used to complete the recovery of the carrier frequency offset, which poses higher requirements for the rate and accuracy of frequency offset estimation.

[0004] Traditional wireless signal receivers often use phase-locked loops to achieve carrier frequency synchronization. Once the phase-locked loop is locked, high-precision frequency synchronization can be obtained. However, the synchronization time of the phase-locked loop is long and the complexity is high, which cannot meet the low-latency requirements of Internet of Things communication and sensing. To meet the requirements of low latency and high-speed communication, an open-loop carrier synchronization method based on feedforward is often used. This synchronization method has the advantages of simple implementation, low computational complexity, and suitability for high-speed communication. The open-loop carrier synchronization method based on feedforward is divided into data-aided DA estimation algorithms and non-data-aided NDA estimation algorithms. The non-data-aided algorithm has a high complexity and cannot quickly capture the carrier frequency offset. Therefore, it is rarely used in the high-speed communication scenarios of the Internet of Things. The data-aided algorithm uses the autocorrelation of the preamble structure to achieve frequency offset estimation. At present, the data-aided estimation methods are divided into frequency-domain algorithms and time-domain algorithms. The frequency-domain algorithm needs to perform a fast Fourier transform FFT on the data and then perform autocorrelation. Compared with the time-domain algorithm, the complexity is higher but the performance is not improved. Therefore, the time-domain algorithm is often used. Some classic time-domain data-aided frequency offset estimation algorithms such as Kay, Fitz, L&R, etc. all use the PN header in the frame structure to obtain the estimated value based on the maximum likelihood estimation criterion. Among them, although the algorithm proposed by Kay has high estimation accuracy, its accuracy only approaches the Cramer-Rao bound CRB under high signal-to-noise ratio conditions, and the estimation accuracy is poor under low signal-to-noise ratio conditions; the algorithms proposed by Fitz and L&R can obtain better estimation accuracy, but the frequency range they can estimate is smaller.

[0005] Kim et al. proposed a frequency offset estimation method based on MPSK symbols in "An improved non-data-aided carrier frequency offset estimator for PSK packets," Proceedings of GLOBECOM'96. 1996 IEEE Global Telecommunications Conference, London, UK, 1996, pp. 1345-1347 vol. 2, doi: 10.1109 / GLOCOM.1996.587665. First, it performs a non-linear transformation by taking the M-th power of the received complex baseband MPSK signal. Second, it reduces the threshold value by taking the mean to improve the estimation accuracy under low signal-to-noise ratio. Then, it multiplies the adjacent symbols conjugately to obtain the phase difference information between symbols. Finally, it takes the magnitude of the phase difference and uses the weighted accumulation method to obtain the final frequency offset estimation value. Due to the influence of the length of the symbol sequence participating in the weighted accumulation on the final accuracy of the frequency offset estimation, the estimation accuracy is greatly reduced during high-speed Internet of Things communication when the length of the available PN sequence for accumulation is short, resulting in its performance deviating seriously from the Cramer-Rao bound (CRB) and being unable to meet the communication requirements of short distance and high density with short PN headers. At the same time, since it is difficult to balance the estimation accuracy and the estimation range of this method, it cannot meet the requirement of achieving high-precision estimation while covering the estimation range. In addition, due to the high structural complexity of the M-th power non-linear transformation used to remove the modulation symbol information in this method, it cannot meet the low-latency requirement of fast access in the Internet of Things. Summary of the Invention

[0006] The object of the present invention is to propose a fast frequency recovery method for wireless Internet of Things communication and sensing in view of the deficiencies of the prior art, so as to improve the frequency offset estimation accuracy in the short PN header data frame mode on the basis of covering the maximum frequency offset estimation range, reduce the complexity of demodulation removal, ensure the efficient access of data frames, and meet the high-density and low-latency requirements of Internet of Things communication and sensing.

[0007] To achieve the above object, the technical solutions adopted by the present invention include the following steps:

[0008] (1) Extract the coarse phase difference information from the QPSK complex baseband preamble PN sequence r(k) with a length of N collected in wireless Internet of Things communication, and obtain the real part value I of the coarse phase difference according to the coarse phase difference information di and the imaginary part value Q di , and then use the real part value I of the coarse phase difference di and the imaginary part value Q di to obtain the real part value θ of the coarse rotation amount real1_i and the imaginary part value θ of the coarse rotation amount imag1_i;

[0009] (2) Utilize the real part value I of the coarse phase difference information di and the imaginary part value Q di , the real part value θ of the coarse rotation amount real1_i and the imaginary part value θ of the coarse rotation amount imag1_i to roughly estimate the carrier frequency offset of r(k) and obtain the rough frequency offset estimation value Δf coarse :

[0010] (2a) According to the real part value I of the coarse phase difference information di and the imaginary part value Q di , the real part value θ of the coarse rotation amount real1_i and the imaginary part value θ imag1_i , calculate N - 1 preliminary demodulation information values D(i):

[0011]

[0012] where N is the length of the r(k) sequence and j represents the imaginary part symbol.

[0013] (2b) Utilize the N - 1 preliminary demodulation information values D(i) to calculate and obtain the rough frequency offset estimation value Δf coarse :

[0014]

[0015] where T represents the period of a QPSK symbol, arctan(·) represents the arctangent function, imag(·) represents taking the imaginary part of the data, real(·) represents taking the real part of the data, and ∑ represents summation.

[0016] (3) Utilize the rough frequency offset estimation value Δf coarse to perform rough frequency offset compensation on the QPSK complex baseband preamble PN sequence r(k) and obtain the compensated QPSK complex baseband preamble PN sequence r c (k);

[0017] (4) Utilize the compensated QPSK complex baseband preamble PN sequence r c (k) to extract the fine phase difference information of the symbol interval m, and obtain the real part value and the imaginary part value of the fine phase difference. Then, utilize the real part value and the imaginary part value of the fine phase difference to obtain the real part value θ real2_i and the imaginary part value θ imag2_i of the fine rotation amount;

[0018] (5) According to the real part value and the imaginary part value Real part value θ of the fine rotation amount real2_i and imaginary part value θ of the fine rotation amount imag2_i perform fine estimation on the carrier frequency offset of the compensated QPSK complex baseband preamble PN sequence r c (k) to obtain the fine frequency offset estimation value Δf fine :

[0019] (5a) According to the real part value and imaginary part value of the fine phase difference information, the real part value θ real2_i and imaginary part value θ imag2_i of the fine rotation amount, calculate N - m quadratic demodulation information values D c (i):

[0020]

[0021] where N represents the length of the QPSK complex baseband PN sequence, m is the symbol interval, and j represents the imaginary part symbol.

[0022] (5b) Use the N - m quadratic demodulation information values D c (i) to obtain the fine frequency offset estimation value Δf fine :

[0023]

[0024] where T represents the period of a QPSK symbol, arctan(·) represents the arctangent function, imag(·) represents taking the imaginary part of the data, real(·) represents taking the real part of the data, and ∑ represents summation.

[0025] (6) According to the coarse frequency offset estimation value Δf coarse and the fine frequency offset estimation value Δf fine obtain the overall carrier frequency offset estimation value Δf all ;

[0026] (7) Use the overall carrier frequency offset estimation value Δf all to perform carrier frequency offset compensation on the data part signal s(t) after r(k) to obtain the restored data signal s c (t).

[0027] Compared with the prior art, the present invention has the following advantages:

[0028] Since the present invention estimates the carrier frequency offset of the QPSK complex baseband preamble short PN sequence by combining two - layer estimation of coarse frequency offset estimation and fine frequency offset estimation, it can improve the frequency offset estimation accuracy in the short PN - header data frame mode on the basis of covering the maximum frequency offset estimation range;

[0029] Due to the adoption of a rotating structure to remove the modulation information carried by constellation points, the present invention reduces the complexity of demodulation and improves the access speed of data frames, meeting the requirements of high density and low latency for Internet of Things communication and perception. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 is the overall flowchart of the implementation of the present invention;

[0031] Figure 2 is the sub - flowchart for rough frequency offset estimation using the rough phase difference and rough rotation amount in the present invention;

[0032] Figure 3 is the sub - flowchart for fine frequency offset estimation using the fine phase difference and fine rotation amount in the present invention;

[0033] Figure 4 is the comparison chart of frequency offset estimation accuracy between the present invention and existing methods;

[0034] Figure 5 is the comparison chart of frequency offset estimation range between the present invention and existing methods. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0035] The following further describes the embodiments of the present invention in detail with reference to the drawings.

[0036] Refer to Figure 1 , the implementation steps of this example are as follows:

[0037] Step 1, extract the rough phase difference information and rough rotation amount of the QPSK complex baseband preamble PN sequence r(k).

[0038] 1.1) Represent the QPSK complex baseband preamble PN sequence r(k) collected in wireless Internet of Things communication as:

[0039] r(k) = e j(2πkΔfT+Φ+θ) + n(k) k = 1,... N

[0040] where N represents the sequence length, T represents a QPSK symbol period, Φ represents the data information, that is, the QPSK modulation phase, Φ = 2πn / 4, n = 0, 1, 2, 3, θ is the determined but unknown carrier start phase difference between the transmitter and receiver, Δf is the carrier frequency offset, and n(k) is a complex zero - mean Gaussian white noise sequence, whose real and imaginary parts are independently and identically distributed with variance σ 2 = N0 / (2E s ), N0 is the one - sided power spectral density of Gaussian white noise, and E sis the symbol energy; Since the r(k) sequence contains the carrier frequency offset Δf and the modulation phase Φ, it is impossible to distinguish whether the phase information in r(k) is the modulation phase Φ carried by the data itself or the phase information accumulated by Δf. Therefore, if we want to estimate Δf, we need to remove the influence of the modulation information Φ;

[0041] 1.2) To eliminate the influence of Φ, conjugate multiply every two adjacent symbol data r(k) and r(k - 1) of the QPSK complex baseband preamble PN sequence and rotate by π / 4 to obtain N - 1 coarse phase difference information d(i) and the phase information φ(i) of d(i):

[0042] d(i) = r(k) × r * (k - 1) × exp(jπ / 4)

[0043] φ(i) = arg(d(i)) = 2πΔfT + Φ i -Φ i-1 +N i -N i-1

[0044] where k = 2,..., N, i = 1,..., N - 1, * denotes conjugate, j denotes the imaginary part symbol, N i -N i-1 is the phase difference of Gaussian white noise, Φ i -Φ i-1 is the phase difference of the modulation symbol. Since PN uses QPSK modulation, Φ i -Φ i-1 = mπ / 4, m = ±1 or ±3. If the influence of the noise N i -N i-1 and 2πΔfT on φ(i) does not exceed π / 4, then we can determine the value of Φ by judging which quadrant the constellation point corresponding to d(i) falls into i -Φ i-1 and obtain the value of Φ i -Φ i-1 After obtaining the value of Φ, we can eliminate the phase difference of this modulation symbol by rotation:

[0045] When Φ i -Φ i-1 = π / 4, d(i) falls into the first quadrant, its real part is positive, and its imaginary part is positive;

[0046] When Φ i -Φ i-1 = 3π / 4, d(i) falls into the second quadrant, its real part is negative, and its imaginary part is positive;

[0047] When Φ i -Φ i-1When Φ = -3π / 4, d(i) lies in the third quadrant, its real part is negative, and its imaginary part is negative;

[0048] When Φ i -Φ i-1 = -π / 4, d(i) lies in the fourth quadrant, its real part is positive, and its imaginary part is negative;

[0049] Therefore, the quadrant area where Φ - Φ is located can be obtained by judging the signs of the real part and the imaginary part of d(i), and d(i) is expressed as: d(i) = I i -Φ i-1 , where I di +jQ di , where I di represents the real part of d(i), and Q di represents the imaginary part of d(i);

[0050] If it is judged that d(i) is in the first quadrant, it means that Φ i -Φ i-1 = π / 4, then the influence of the modulation phase can be eliminated by rotating d(i) clockwise by π / 4 to make Φ i -Φ i-1 = 0;

[0051] Similarly, when d(i) is in the second, third, and fourth quadrants, the influence of the modulation phase can be eliminated by rotating d(i) clockwise by 3π / 4, 5π / 4, and 7π / 4;

[0052] 1.3) Take the real part and the imaginary part of each coarse phase difference information d(i) respectively to obtain the real part value I di of the coarse phase difference and the imaginary part value Q di :

[0053] I di = real(d(i))

[0054] Q di = imag(d(i))

[0055] where real(·) represents taking the real part of the data, and imag(·) represents taking the imaginary part of the data.

[0056] 1.4) In order to obtain the complex expression of the quantity to be rotated, according to the magnitudes of the real part value I di of the coarse phase difference and the imaginary part value Q di , obtain the real part value θ real1_i of the coarse rotation quantity and the imaginary part value θ imag1_i of the coarse rotation quantity:

[0057]

[0058] Step 2, use the real part value I of the coarse phase difference informationdi and the imaginary part value Q di , the real part value θ of the coarse rotation amount real1_i and the imaginary part value θ of the coarse rotation amount imag1_i Coarsely estimate the carrier frequency offset of r(k) to obtain a coarse frequency offset estimate value Δf coarse , and perform coarse frequency offset compensation on it.

[0059] Refer to Figure 2 , the specific implementation of this step is as follows:

[0060] 2.1) According to the real part value I of the coarse phase difference information di and the imaginary part value Q di , the real part value θ of the coarse rotation amount real1_i and the imaginary part value θ imag1_i , calculate N - 1 preliminary demodulation information values D(i):

[0061]

[0062] where N is the length of the QPSK complex baseband preamble PN sequence r(k), and j represents the imaginary part symbol;

[0063] 2.2) According to the preliminary demodulation information value D(i), obtain its phase information φ d (i):

[0064] φ d (i) = arg(D(i)) = 2πΔfT + N i -N i-1

[0065] where Δf is the carrier frequency deviation, T represents a QPSK symbol period, and N i -N i-1 is the phase difference of Gaussian white noise;

[0066] 2.3) According to the relationship between the phase information φ d (i) and the frequency, obtain the single - time frequency estimate value Δf i :

[0067]

[0068] 2.4) To reduce the influence of noise, take the average value of N - 1 times as the coarse frequency offset estimate value Δf coarse :

[0069]

[0070] where T represents the period of a QPSK symbol, imag(·) represents taking the imaginary part of the data, and real(·) represents taking the real part of the data, Characterizes the argument value of the constellation points. arctan(·) represents the arctangent function. After taking the arctangent of the argument value, it is the phase difference caused by the frequency offset of the constellation point. Multiple accumulations and averaging can reduce the impact of noise on the frequency offset estimation.

[0071] Since calculating the argument is a non - linear process modulo 2π, to avoid the phase ambiguity problem, Δf coarse should satisfy:

[0072] |2πΔf coarse T| < π / 4

[0073] This expression indicates the maximum estimation range of the coarse frequency offset estimation.

[0074] 2.5) Use the coarse frequency offset estimation value Δf coarse to perform coarse frequency offset compensation on the QPSK complex baseband preamble PN sequence r(k), and obtain the compensated QPSK complex baseband preamble PN sequence r c (k).

[0075]

[0076] where r(k) represents the complex baseband preamble PN sequence, Δf coarse represents the coarse frequency offset estimation value, T represents the period of a QPSK symbol, N represents the length of the QPSK complex baseband PN sequence, and j represents the imaginary part symbol.

[0077] Step 3, use the compensated QPSK complex baseband preamble PN sequence r c (k) to extract the fine phase difference information and fine rotation amount of symbol interval m.

[0078] The above - mentioned coarse frequency offset estimation uses adjacent symbols to estimate the frequency offset value within the maximum estimation range. However, the cumulative frequency offset of adjacent symbol intervals is small, and the estimation accuracy is low. Therefore, the residual frequency offset amount after coarse frequency offset compensation is large. To improve the estimation accuracy, a fine frequency offset estimation needs to be performed on the preamble sequence after coarse frequency offset compensation. The specific implementation is as follows:

[0079] 3.1) Take two data r c (k) with an interval of m from the sequence r c (k + m) and r c (k) for conjugate multiplication and rotate by π / 4 to obtain N - m fine phase difference information d c (i):

[0080] d c (i) = r c (k + m) × r c * (k) × exp(jπ / 4) i = 1,..., N - m,

[0081] where N represents the length of the QPSK complex baseband PN sequence, * represents conjugation, j represents the imaginary part symbol, and the cumulative frequency offset when the symbol interval is m is m times the cumulative frequency offset of adjacent symbols, which can more finely estimate the residual frequency offset and reduce the frequency offset estimation error;

[0082] 3.2) For the fine phase difference information d c (i) Take the real part and the imaginary part to obtain the real part value of the fine phase difference and the imaginary part value

[0083]

[0084]

[0085] where real(·) represents taking the real part of the data, and imag(·) represents taking the imaginary part of the data;

[0086] 3.3) To obtain the complex expression of the quantity to be rotated, according to the real part value and the imaginary part value of the fine phase difference, obtain the real part value θ real2_i of the fine rotation quantity and the imaginary part value θ imag2_i of the fine rotation quantity, and the expressions are as follows respectively:

[0087]

[0088] Step 4, use the real part value and the imaginary part value of the fine phase difference information, the real part value θ real2_i and the imaginary part value θ imag2_i of the fine rotation quantity to finely estimate the carrier frequency offset of r(k) and obtain the fine frequency offset estimation value Δf fine .

[0089] Refer to Figure 3 for the specific implementation of this step as follows:

[0090] 4.1) According to the real part value and the imaginary part value of the fine phase difference information, the real part value θ real2_i and the imaginary part value θ imag2_i of the fine rotation quantity, calculate N - m quadratic demodulation information values D c (i):

[0091]

[0092] where N represents the length of the QPSK complex baseband PN sequence, m is the symbol interval, and j represents the imaginary part symbol;

[0093] 4.2) Utilize N - m quadratic de - modulation information values D c (i) Obtain the fine frequency offset estimation value Δf fine :

[0094]

[0095] where T represents the period of a QPSK symbol.

[0096] Similar to the coarse frequency offset estimation, the method of taking the average of N - m estimations is adopted to reduce the influence of noise on the estimation accuracy. Since calculating the argument of a complex number is a non - linear process modulo 2π, to avoid the phase ambiguity problem, Δf fine should satisfy:

[0097] |2πmΔf fine T| < π / 4

[0098] This expression indicates the maximum estimation range of the fine frequency offset estimation.

[0099] Step 5, Calculate the overall carrier frequency offset estimation value Δf coarse based on the coarse frequency offset estimation value Δf fine and the fine frequency offset estimation value Δf all , and use Δf all to perform frequency offset compensation on the data part of the received signal.

[0100] 5.1) Calculate the overall carrier frequency offset estimation value Δf coarse based on the coarse frequency offset estimation value Δf fine and the fine frequency offset estimation value Δf all :

[0101] Since the overall carrier frequency offset estimation is divided into two rounds of coarse frequency offset estimation and fine frequency offset estimation, and the fine frequency offset estimation value is obtained on the basis of coarse frequency offset compensation, the overall carrier frequency offset estimation value can be obtained by summing the coarse frequency offset estimation value and the fine frequency offset estimation value.

[0102] Δf all = Δf coarse + Δf fine

[0103] where Δf coarse represents the coarse frequency offset estimation value, and Δf fine represents the fine frequency offset estimation value;

[0104] 5.2) Use the overall carrier frequency offset estimation value Δf all to perform frequency offset compensation on the data part s(t) of the received signal, and obtain the restored data signal s c (t):

[0105]

[0106] where Δf all represents the overall carrier frequency offset estimation value, s(t) represents the received signal of the data part, and j is the imaginary part symbol.

[0107] The effects of the present invention are further described below through simulation experiments.

[0108] 1. Simulation conditions:

[0109] The simulation uses an Inter Core i9-10900F CPU with a main frequency of 2.80 GHz, 32.0 GB of memory, a 64-bit operating system, and Microsoft Windows 10 Professional Edition, and is verified on the MATLAB 2022a simulation software.

[0110] The length of the complex baseband short preamble PN sequence modulated by QPSK is N = 56, and the length of the complex baseband long preamble PN sequence modulated by QPSK is 128. It is assumed that symbol synchronization has been completed before frequency offset estimation to obtain a complete preamble sequence modulated by QPSK.

[0111] 2. Simulation content and result analysis:

[0112] Simulation 1: Set the frequency offset to 500 Hz at the baseband symbol level, and the signal-to-noise ratio increases from -5 dB to 15 dB at intervals of 1 dB each time. At each signal-to-noise ratio, 5000 Monte Carlo simulations are performed using the present invention and the prior art respectively. Among them, the prior art is simulated separately in the cases of long PN and short PN, and the present invention is simulated in the case of short PN. The absolute value of the amplitude difference between the estimated frequency and the added frequency offset is statistically counted as the frequency offset estimation error. The results are as Figure 4 shown.

[0113] From Figure 4 it can be seen that when using the prior art, at the same signal-to-noise ratio, the estimation accuracy of the short PN header is lower than that of the long PN header. The combination of coarse frequency offset estimation and fine frequency offset estimation adopted by the present invention improves the estimation accuracy of the short PN header. At a signal-to-noise ratio of 0 dB, the frequency offset estimation error of the prior art is 134 Hz, and the frequency offset estimation error of the present invention technology is 87 Hz, and the estimation error is reduced by about 35.07%, improving the frequency offset estimation accuracy.

[0114] Simulation 2: With the signal-to-noise ratio set at 10 dB, the normalized frequency offset is set at the baseband symbol level, and the offset value increases from -0.25 to 0.25 at intervals of 0.05. At each frequency offset parameter setting, 5000 Monte Carlo simulations are performed using the present invention and the prior art respectively. Among them, the prior art is simulated separately in the cases of long PN and short PN, and the present invention is simulated in the case of short PN. The absolute value of the amplitude difference between the estimated frequency and the applied frequency offset is respectively counted as the frequency offset estimation error. The results are as Figure 5 shown.

[0115] From the simulation Figure 5 it can be seen that, under the condition of ensuring the same estimation range as that of the long PN, the present invention reduces the frequency offset estimation error under the short PN and improves the accuracy of the frequency offset estimation under the short PN, which is superior to the existing algorithms.

[0116] The above simulation shows that the present invention can complete the frequency offset estimation on the premise of the short PN header preamble sequence and improve the accuracy of the frequency offset estimation on the basis of ensuring the estimation range.

[0117] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, and improvements made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A fast frequency recovery method for wireless Internet of Things communication and sensing, characterized in that Including the following steps: (1) Extract the coarse phase difference information by using the QPSK complex baseband preamble PN sequence r(k) with length N collected in the wireless IoT communication, and obtain the real part value I of the coarse phase difference according to the coarse phase difference information. di and the imaginary part value Q di , and then use the real part value I of the coarse phase difference di and the imaginary part value Q di to obtain the real part value θ of the coarse rotation amount real1_i and the imaginary part value θ of the coarse rotation amount imag1_i ; (2) Utilize the real part value I of the coarse phase difference information di and the imaginary part value Q di , the real part value θ of the coarse rotation amount real1_i and the imaginary part value θ of the coarse rotation amount imag1_i to roughly estimate the carrier frequency offset of r(k) and obtain the rough frequency offset estimation value Δf coarse : (2a) According to the real part value I of the coarse phase difference information di and the imaginary part value Q di , the real part value θ of the coarse rotation amount real1_i and the imaginary part value θ imag1_i , calculate N - 1 preliminary demodulation information values D(i): Where N is the length of the r(k) sequence, and j represents the imaginary part symbol; (2b) Using N - 1 preliminary demodulated information values D(i), calculate the coarse frequency offset estimation value Δf coarse : Where T represents the period of a QPSK symbol, arctan(·) represents the arctangent function, imag(·) represents taking the imaginary part of the data, real(·) represents taking the real part of the data, and ∑ represents summation; (3) Utilize the coarse frequency offset estimation value Δf coarse Perform coarse frequency offset compensation on the QPSK complex baseband preamble PN sequence r(k) to obtain the compensated QPSK complex baseband preamble PN sequence r c (k); (4) Utilize the compensated QPSK complex baseband preamble PN sequence r c (k) Extract the fine phase difference information of symbol interval m, and obtain the real part value of the fine phase difference according to the fine phase difference information and the imaginary part value Then utilize the real part value of the fine phase difference and the imaginary part value to obtain the real part value θ of the fine rotation amount real2_i and the imaginary part value θ of the fine rotation amount imag2_i ; (5)Based on the real part value of the fine phase difference information and the imaginary part value The real part value θ of the fine rotation amount real2_i and the imaginary part value θ of the fine rotation amount imag2_i Perform fine estimation on the carrier frequency offset of the compensated QPSK complex baseband preamble PN sequence r c (k) to obtain the fine frequency offset estimation value Δf fine : (5a)According to the real part value of the fine phase difference information and the imaginary part value The real part value θ of the fine rotation amount real2_i and the imaginary part value θ imag2_i , calculate N-m quadratic demodulation information values D c (i): Where N represents the length of the QPSK complex baseband PN sequence, m is the symbol interval, and j represents the imaginary part symbol; (5b) Utilize N - m quadratic de - modulation information values D c (i) Obtain the fine frequency offset estimation value Δf fine : Where T represents the period of a QPSK symbol, arctan(·) represents the arctangent function, imag(·) represents taking the imaginary part of the data, real(·) represents taking the real part of the data, and ∑ represents summation; (6) According to the coarse frequency offset estimation value Δf coarse and the fine frequency offset estimation value Δf fine obtain the overall carrier frequency offset estimation value Δf all ; (7) Using the estimated value of the overall carrier frequency offset Δf all Perform carrier frequency offset compensation on the data part signal s(t) after r(k) to obtain the restored data signal s c (t).

2. The method according to claim 1, wherein In step (1), the real part value I of the phase difference is obtained based on the coarse phase difference information di and the imaginary part value Q di , which is achieved as follows: (1a) Perform conjugate multiplication on two adjacent symbol data r(k) and r(k - 1) of the QPSK complex baseband preamble PN sequence and rotate by π / 4 to obtain N - 1 pieces of coarse phase difference information d(i): d(i) = r(k) × r * (k - 1) × exp(jπ / 4) where k = 2,..., N, i = 1,..., N - 1, * denotes conjugate, N denotes the length of the QPSK complex baseband PN sequence, and j denotes the imaginary part symbol; (1b) Take the real part and the imaginary part of each coarse phase difference information d(i) respectively to obtain the real part value I of the coarse phase difference di and the imaginary part value Q di : I di = real(d(i)) Q di = imag(d(i)) Where real(·) represents taking the real part of the data, and imag(·) represents taking the imaginary part of the data.

3. The method according to claim 1, wherein The real part value θ of the rough rotation amount obtained in step (1) real1_i and the imaginary part value θ of the rough rotation amount imag1_i are respectively expressed as follows: Among them, I di and Q di represent the real part value and the imaginary part value of the coarse phase difference, and N represents the length of the QPSK complex baseband PN sequence.

4. The method according to claim 1, wherein The compensated QPSK complex baseband preamble PN sequence r c (k) obtained in step (3) is expressed as follows: where r(k) represents the complex baseband preamble PN sequence, Δf coarse represents the coarse frequency offset estimation value, T represents the period of a QPSK symbol, N represents the length of the QPSK complex baseband PN sequence, and j represents the imaginary part symbol.

5. The method according to claim 1, wherein In step (4), the real part value is obtained based on the fine phase difference information and the imaginary part value The implementation is as follows: (4a) After compensating for the coarse frequency offset, the complex baseband preamble PN sequence r c (k) Two symbol data r separated by m c (k + m) and r c (k) are conjugated and multiplied, and rotated by π / 4 to obtain N - m fine phase difference information d c (i): d c (i) = r c (k + m)×r c * (k)×exp(jπ / 4) i = 1,..., N - m where N represents the length of the QPSK complex baseband PN sequence, * denotes conjugation, and j denotes the imaginary part symbol; (4b) For the fine phase difference information d c (i) Take the real part and the imaginary part to obtain the real part value of the fine phase difference and the imaginary part value Where real(·) represents taking the real part of the data, and imag(·) represents taking the imaginary part of the data.

6. The method according to claim 1, wherein The real part values θ of the N-m coarse rotation amounts obtained in step (4) real2_i and the imaginary part values θ of the coarse rotation amounts imag2_i are expressed as follows: Among them, and represent the real part value and the imaginary part value of the fine phase difference, m represents the symbol interval, and N represents the length of the QPSK complex baseband PN sequence.

7. The method according to claim 1, wherein The overall carrier frequency offset estimation value Δf obtained in step (6) all , is expressed as follows: Δf all = Δf coarse + Δf fine Among them, Δf coarse represents the coarse frequency offset estimation value, and Δf fine represents the fine frequency offset estimation value.

8. The method according to claim 1, wherein The restored data signal s c (t) obtained in step (7) is expressed as follows: Among them, Δf all represents the overall carrier frequency offset estimation value, s(t) represents the received signal of the data part, and j is the imaginary part symbol.

Images