Disclosure of Invention
The invention provides a frequency estimation and automatic frequency deviation control method suitable for an OFDM receiver aiming at the technical problems in the prior art, which can realize accurate frequency deviation estimation and automatic frequency control in a larger frequency deviation range, and the frequency deviation of a received signal processed by the technical system can be controlled in a minimum range, thereby ensuring that the subsequent processing can be normally carried out.
In order to achieve the above object, the technical solution of the present invention is as follows, a fast digital automatic frequency for an orthogonal frequency division multiplexing receiverA method of rate control, the method comprising the steps of: 1) frequency offset estimation, 2) frequency offset judgment and 3) frequency offset control, wherein 1) the frequency offset estimation step sets training sequences for frequency offset estimation at a sending end, and the training sequences are sequentially received at a receiving end, and the specific method comprises the following steps: 11) obtaining the negative value of the rear half part of the received time domain preamble training sequence
And fine synchronization is carried out in the time domain to obtain the fractional frequency multiplication deviation estimated value of the subcarrier interval
;
The fine synchronization is to select a 64-point long correlation window, and perform correlation operation on the conjugate of the first 64-point sample of the received preamble training sequence and the 64-point sample delayed by 128 points to obtain a correlation calculation result
Thereby obtaining an estimated value of the fractional carrier frequency offset
;
2) The frequency offset judgment step judges whether the frequency offset estimation value obtained in the step 1) needs to be subjected to frequency offset control, and the specific method comprises the following steps:
21) obtaining the frequency deviation estimated value obtained in the step 1), and averaging M points to obtain a frequency deviation mean value
,
22) Setting the frequency deviation control target range of the receiver working normally according to the design requirement of the communication system
Wherein
Representing the maximum frequency deviation value required by the system, and judging the mean value of the frequency deviation
If the target range is met, no frequency offset control is carried out if the target range is met, otherwise, frequency offset control is carried out;
3) the frequency control step adaptively adjusts the control voltage of the temperature compensated voltage controlled oscillator TCXO so as to correct the frequency offset, and simultaneously stores the control coefficient s and the control voltageThe method is used for increasing the efficiency of automatic control after the system is powered off, and comprises the following specific steps:
31) calculating a control coefficient initial value s: frequency offset obtained due to the frequency offset estimation step
Is the deviation of corresponding radio frequency carrier, if the carrier frequency is m times of TCXO frequency, the frequency deviation mean value is converted into the frequency difference of TCXO frequency
The voltage-controlled oscillator control voltage v is proportional to the oscillation frequency f, i.e.
Where k and b are constants, to obtain
To obtain
,
32) For the above step 2) decisionObtaining the frequency deviation mean value needing to be subjected to frequency deviation control to carry out frequency deviation adjustment, and carrying out the frequency deviation adjustment according to formula 31)
Calculating the difference value of the voltage-controlled oscillator control voltage v to be adjusted
Adjusting the control voltage of the voltage controlled oscillator to
Wherein
In order to adjust the value of the control voltage,
33) judging the effect after the control of the step 32), wherein the frequency deviation after the control meets the requirement
Then the control voltage is stored
And if the frequency deviation is not satisfied after the control, repeating 32) after correcting the control coefficient s.
As an improvement of the present invention, said method for fast digital automatic frequency control of an orthogonal frequency division multiplexing receiver is characterized in that said step 1 of frequency offset estimation further comprises the following step, 12) pair
Performing fractional frequency offset compensation to obtain training sequence after fractional frequency offset compensation
The frequency offset compensation is to obtain a fractional frequency multiplication offset estimation value
Then, the estimated value is used
And performing frequency compensation on the training sequence in the time domain as follows:
13) in the frequency domain
Coarse synchronization is carried out to obtain an integral multiple frequency offset estimation value of subcarrier spacing
,
The coarse synchronization of the frequency domain is firstly aligned
And locally transmitting training sequences
Performing Fourier transform FFT to obtain frequency domain data thereof
And
separately for frequency domain data
And
performing difference operation to obtain difference operation result
And
,
then, the difference sequence is paired
And
performing sliding cross-correlation calculation, wherein when the cross-correlation value reaches the maximum value, the shift element is the size of the integer frequency offset, and the estimated value of the integer frequency offset
Calculated as follows:
wherein
Which is related to the spacing of the non-zero points of the training sequence in the frequency domain, the range of the integer frequency offset estimation is
Exceeding will cause erroneous integer frequency offset estimation.
As a modification of the present invention, the step 11) sends the preamble training sequence to transmit data on frequency domain using an integer multiple of 4 subcarriers, and other subcarriers are null, so that it includes 4 64 short training sequences in time domain, where the first short training sequence is denoted by a, the second short training sequence is a repetition of the first short training sequence, and the third and fourth short training sequences are negative values of the first short training sequence, and are denoted by A, A, -a, respectively.
As an improvement of the present invention, in the step 12), the correlation operation is defined as:
wherein,
it is shown that the sequence is taken as the conjugate,
represents the time-domain training sequence that is transmitted,
representing additive white Gaussian noise with zero mean and variance
,
Representing carrier frequency offset normalized by subcarrier spacing
,
。
As a refinement of the present invention, the step 13) defines the difference operation as:
andrespectively representing frequency domain dataAndand (5) carrying out data after difference operation.
As an improvement of the present invention, the frequency offset M point obtained in the frequency offset estimation step in step 21) is averaged, and the larger the value of M is, the more accurate the frequency offset estimation is, the more accurate the control effect is, but the longer the time is, the time effectiveness is reduced. And the value M is selected in compromise by the requirements of system frequency offset control precision and timeliness. In the method, the frequency deviation M point obtained in the frequency deviation estimation step is taken as an average value, and M is 1.
As an improvement of the invention, in the step 31), the voltage-controlled oscillator control voltage v is proportional to the oscillation frequency f
Wherein k and b are constants. k. The value of b is given by a manufacturer data manual of the voltage-controlled oscillator, and different components and parts have differences; control coefficient
And m is a frequency multiplication factor, and the k value drifts to some extent due to different environmental factors and individual components, so that s needs to be corrected in a self-adaptive manner.
As an improvement of the present invention, in the step 33), the frequency offset control coefficient s is calibrated, and the method includes:
wherein
In order to control the difference in the voltage,
is the average value of the frequency offsets before frequency offset control,
is the frequency deviation average value after the frequency deviation control.
As an improvement of the present invention, in the step 33), for the control voltage meeting the system frequency offset requirementAnd a majority judgment method is adopted for storage, the data are repeatedly stored in a plurality of addresses, the voltage value of the position is read after the system is restarted, and the value with the largest occurrence ratio is used as the initial value of the control voltage.
Compared with the prior art, the invention has the following advantages: the method only utilizes the good autocorrelation characteristic of a short training sequence A in a leading training sequence, and uses the conjugation of the first 64-point sample value and the data with less 64-point sample values with 128 points delayed to carry out correlation operation to accurately obtain fractional multiple frequency offset; meanwhile, a differential operation of a preamble training sequence frequency domain is introduced, the estimation of the integer frequency offset is completed by utilizing the good autocorrelation characteristic of the preamble training sequence frequency domain, the robustness of the estimation of the integer frequency offset is improved, the influence of the timing offset on the algorithm is eliminated, and simulation shows that the integer frequency offset can be correctly estimated under different timing offsets; the phenomenon that the frequency deviation fluctuates continuously outside a target range due to environmental factors, individual differences of components and time drift is overcome by adaptively adjusting the frequency deviation control coefficient; the real-time updating of the initial value of the control voltage avoids the fact that a standard signal source is required to be used for calibration before the system is used due to different environments, and meanwhile, the system efficiency is improved.
Detailed Description
For a better understanding and appreciation of the invention, the invention will be further illustrated and described below in connection with the accompanying drawings and detailed description.
Example 1:
referring to fig. 6, a fast digital automatic frequency control method for an orthogonal frequency division multiplexing receiver, the method comprising the steps of: 1) frequency offset estimation, 2) frequency offset judgment and 3) frequency offset control, wherein in the frequency offset estimation step, training sequences for frequency offset estimation are set at a sending end, and the training sequences are sequentially received at a receiving end, referring to fig. 1, a preamble training sequence structure for frequency offset estimation is used for sending the preamble training sequence to transmit data by using subcarriers of integral multiple of 4 in a frequency domain, and other subcarriers are null, so that the preamble training sequence comprises 4 64-point short training sequences in a time domain, a first short training sequence is represented by a, and the preamble training sequence can be represented as [ a, -a ], as shown in fig. 1.
Referring to fig. 2, fractional frequency offset estimation and frequency offset compensation: the frequency offset estimation occurs after timing synchronization, when a synchronization point is detected
Then, the following N (N is 256) point data is extracted, and the data is corrected
The dot symbol bit is inverted to obtain a time domain preamble training sequence
:
Using the periodicity and correlation of the leading training sequence, selecting a correlation window with the length of 64 points, and performing correlation operation on the conjugate of the first 64 point sample of the leading training sequence and the 64 point sample value delayed by 128 points to define:
wherein,
it is shown that the sequence is taken as the conjugate,
represents the time-domain training sequence that is transmitted,
representing additive white Gaussian noise with zero mean and variance
,
Denotes carrier frequency offset normalized by subcarrier spacing (
),
Based on the above formula when
The fractional octave bias can be estimated as:
estimating to obtain decimal frequency offset
Then, the estimated value is used
And performing frequency offset compensation on the signal in a time domain as follows:
wherein,
for the purpose of frequency offset compensation of the pre-signal,
is a frequency offset compensated signal.
Referring to fig. 3, integer-multiple frequency offset estimation:
after time domain decimal frequency offset estimation and compensation, ICI caused by carrier frequency offset is basically eliminated, but integral frequency offset may still exist in the system
And timing deviation
The integer frequency offset estimation is a frequency domain training sequence to be received in the frequency domain
And transmitted training sequence
And completing the shifting cross correlation, wherein the shifting is the size of the integral frequency offset when the cross correlation value reaches the maximum value. Timing deviation
Then frequency domain data will be received
And transmitting the frequency domain training sequence
The difference is made separately to eliminate the influence of the timing deviation.
Firstly, to
And
performing Fourier transform FFT to obtain frequency domain data thereof
And
separately for frequency domain data
And
performing a difference operation, defining a difference operation:
and
respectively representing frequency domain data
And
and (5) carrying out data after difference operation.
Then the difference sequence is
And
cross correlation for slidingOperation, when the cross-correlation value reaches the maximum, the shift value is the size of the integral frequency deviation, the estimated value of the integral frequency deviation is
Calculated as follows:
wherein
Which is related to the spacing of the non-zero points of the training sequence in the frequency domain, the range of the integer frequency offset estimation is
Exceeding will cause erroneous integer frequency offset estimation.
Referring to fig. 5, fig. 5 is a simulation diagram of integer frequency offset estimation without timing frequency offset and with different timing offsets, and it can be seen that, due to the use of differential operation, the integer frequency offset can be correctly estimated, and the influence of the timing frequency offset is completely eliminated.
Synthesis, available frequency offset
Total estimate
Comprises the following steps:
for the M frequency deviation estimated values
Averaging to obtain a frequency spectrum average value
The M value is selected in a compromise mode of timeliness and accuracy, and the larger the M value is, the more accurate the frequency offset control is, and meanwhile, the poorer the control timeliness is. The embodiment is a burst system, in order to ensure that M is taken as 1 for the time period of frequency offset control, i.e. frequency offset judgment and control are carried out immediately after frequency offset is estimated,
。
referring to fig. 4, the frequency offset estimation value is determined
Whether or not to satisfy
Wherein
The maximum allowable positive frequency offset meeting the design requirement of the communication system is set by the overall design of the system. If it satisfies
The communication system can normally work without adjusting the frequency offset, so that the frequency offset is not adjusted. If not satisfied with
If the communication system is not working normally, the frequency offset needs to be adjusted, and the frequency offset is reported to the frequency offset control step to adjust the frequency offset of the system.
The TCXO of the temperature compensated vco used in this embodiment may control its control voltage by using a register, so as to adjust its own oscillation frequency. Chip data manual acquisition type
Where f is the oscillation frequency, v is the control voltage, and k and b are constants. The frequency multiplier of the system frequency multiplier is 175 to obtain the frequency of the radio frequency carrier
,
Wherein
As a difference in the frequency of the radio frequency carriers, i.e. the above-mentioned frequency offset
,
And calculating to obtain the initial frequency offset control coefficient of the system
. And calculating the initial control voltage v of the system according to the carrier frequency, and writing the initial control voltage v into a temperature compensation voltage-controlled oscillator register.
Calculating the difference value of the control voltage v
Changing the control voltage v to
And changing the oscillation frequency of the local temperature compensation voltage-controlled oscillator so as to correct the radio frequency carrier frequency and reduce the frequency deviation. After the control, judging the frequency deviation estimated value obtained in the frequency deviation estimation step, and if the frequency deviation after the control meets the requirement
Then the frequency deviation obtained by the above calculation is controlledThe coefficient s is accurate in the application environment, the control coefficient s and the control voltage at that time are stored
. Otherwise s must be corrected. The correction method of the frequency deviation control coefficient s comprises
Wherein
In order to control the difference in the voltage,
is the average value of the frequency offsets before frequency offset control,
is the frequency deviation average value after the frequency deviation control.
And after correcting the frequency deviation control coefficient s, repeating the frequency deviation judgment and control process, judging whether the frequency deviation meets the requirement of the system frequency deviation, if so, storing the control voltage and the control coefficient, and if not, correcting the control coefficient s again. Iteratively correcting s until the control frequency offset satisfiesAnd at the moment, the control voltage v and the control coefficient s are the control voltage and the control coefficient of the correct temperature compensation voltage-controlled oscillator of the system in the current working environment, and are stored, so that the system can be conveniently and directly called next time after being restarted. The storage method uses majority judgment, simultaneously stores the majority judgment in 10 addresses, reads the value of the address when calling next time, and the most value is the control voltage and the control coefficient, thereby avoiding the problem that the system cannot work due to overlarge frequency deviation after restarting because of the storage error caused by accident.
The automatic frequency offset control is continuously executed in the system operation process, and frequency drift caused by the change of each parameter of the system along with the drift of time and the operation state is effectively avoided.
It should be noted that the above-mentioned embodiments illustrate only preferred embodiments of the invention, and are not intended to limit the scope of the invention, so that equivalents and substitutions made on the above-mentioned embodiments are within the scope of the invention as defined by the claims.