CN107835035B - Open-loop demodulation method and device for short-frame burst communication with low signal-to-noise ratio - Google Patents

Abstract

The invention discloses a method and a device for demodulating open loop of low signal-to-noise ratio short frame burst communication, belonging to the technical field of direct sequence spread spectrum communication demodulation. The invention adopts a segmented data processing mode, firstly de-modulates the de-spread data, then carries out zero filling FFT, estimates the frequency deviation change frequency and frequency deviation and carries out compensation. And accurately estimating and compensating each section of frequency deviation by using an exhaustion method, and deblurring the phase by using the polarity of the frame header to finish correct demodulation of data. Compared with the prior art, the invention overcomes the defects of phase ambiguity, low convergence speed and the like of the traditional phase-locked loop coherent demodulation device, can realize the quick search and compensation of frequency deviation and frequency deviation change rate, completes phase identification and phase correction, and ensures that a direct sequence spread spectrum signal can be normally demodulated under the conditions of short time and low signal-to-noise ratio.

Description

Open-loop demodulation method and device for short-frame burst communication with low signal-to-noise ratio

technical field

The invention relates to a carrier phase tracking and coherent demodulation method (hereinafter referred to as a demodulation device), in particular to a low signal-to-noise ratio short frame burst communication open-loop demodulation method and device, which are suitable for high dynamics, low signal-to-noise ratio, etc. The invention relates to open-loop search, tracking and symbol coherent demodulation of carrier frequency offset and frequency offset change rate in a short frame burst direct sequence spread spectrum system under the environment, belonging to the technical field of demodulation in direct sequence spread spectrum communication.

Background technique

The communication between high-speed aircraft and ground equipment or high-speed aircraft has the characteristics of long communication distance, large link attenuation, and low signal-to-noise ratio, so direct sequence spread spectrum communication system is often used; and this kind of communication often requires communication concealment. , With anti-detection capability, it adopts the short frame burst system. Due to the fast movement speed and complex movement characteristics of the aircraft, the relative radial motion between the transmitter and the receiver on the communication link has a first-order or even higher-order acceleration. On the one hand, due to the short frame burst communication system, frame communication The time is very short, so the high-order motion acceleration can be ignored, but on the other hand, there is a large Doppler frequency offset in the received signal of the receiver and the Doppler first-order frequency offset change rate (hereinafter referred to as the frequency offset change rate) cannot be ignored. ; Since the signal-to-noise ratio of the received signal is very low, the coherent demodulation method is used. If the phase-locked loop method is used to realize the tracking of the carrier Doppler frequency offset and the frequency offset change rate, in the short frame burst communication system In this case, the loop cannot be converged in time, resulting in abnormal demodulation of data. Therefore, an open-loop search method with faster search speed and shorter time should be used to complete coherent demodulation.

The usual direct sequence spread spectrum receiver mainly includes acquisition module, despreading module, demodulation module, decoding module and so on. The function of the acquisition module is to obtain the frequency offset and phase information of the received signal and send them to the despreading module; the despreading module uses the captured information to restore the data before spreading; the demodulation module uses the despreading output data to restore the baseband information before modulation ; The decoding module uses the output of the demodulation module to complete the decoding.

It is assumed that the spreading code of the direct sequence spread spectrum communication receiver is c[l], l=0,1,2,...,L-1, and L is the length of the spreading code. c(l) takes a value of 0 or 1, using radio technology conventions, c[l]=0 means logically positive, c[l]=1 means logically negative, and uses root raised cosine pulse shaping, pulse shaping filter impulse function is h(t), then the baseband waveform expression after shaping is:

Among them, b[m] is the baseband data symbol, its period T _s =LT _c , and T _c is the chip period. In the short frame burst communication system, there are generally certain requirements for the frame structure of the burst frame; in order to facilitate capture, the frame At the beginning of communication, a pilot signal must be transmitted first, that is, b[m] is always 0. After the pilot signal ends, b[m] is the M sequence with a fixed length or the frame header of the Gold code, and the data segment after the frame header ends. After modulation, the signal sent by the radio frequency is:

S _srnd (t)=cos(2πf _RF t)S _b,s (t)

In the above formula, f _RF is the radio frequency, and for the convenience of expression, it is assumed that the initial phase of the carrier and the initial phase of the code of the transmitted signal are both 0.

After the Doppler effect and signal delay in the wireless transmission process, it reaches the receiver and performs I/Q two-way quadrature down-conversion on the RF signal, and the complex baseband signal is obtained as:

In the above formula, f _β and f _σ are respectively the Doppler frequency offset change rate and Doppler frequency offset generated by the relative radial acceleration motion of both sides of the communication link; τ ₀ is the code phase delay. The two-dimensional search of Doppler frequency offset and pseudo-code phase is completed through the acquisition module of the receiver. Among them, once the frequency offset change rate is considered, "frequency offset" and "initial frequency offset" are two different concepts. By default, "frequency offset" refers to the instantaneous frequency offset; but for the occasions where there is a frequency offset change rate, the instantaneous frequency offset is a time-varying function, and only the "initial frequency offset" (the instantaneous frequency offset at the beginning of the frame burst communication) is constant. The frequency offset obtained by the capture device is the instantaneous frequency offset.

The despreading module compensates the input data of the despreading module in the frequency domain by using the rough estimation of the frequency offset obtained by the acquisition device, but does not compensate the frequency offset change rate, and the compensation result is:

为捕获模块得到的频偏的粗估计，r[n]为r(t)数字化表达式。对r_{de_In}[n]进行解扩；The above formula

For the rough estimation of the frequency offset obtained by the capture module, r[n] is the digitized expression of r(t). despread r _{de_In} [n];

Using the periodic characteristics of the spreading code to perform a correlation operation on the input data, that is, despreading. Only when r _{de_In} [n] and y[n] are the same sequence and their phases are consistent with each other, the correlation value is the largest at this time, and d[n] is the despreading output result, which is updated according to the symbol rate. Considering the influence of frequency offset change rate and residual frequency offset, the despread data output can be written as:

ω_σ为捕获模块得到频偏粗估计相对于码片速率归一化数字频率，R_c为码片速率，R_c＝R_s·L。代表残余载波初始(数字)频偏，假定初始相位为

b_n∈0/1代表发送的二进制符号(BPSK调制)；解调模块的作用是完成频偏变化率，频偏的搜索、补偿以及相位鉴别，解调后符号记为z(n),其表达式为：In the above formula, β is the digitized frequency offset change rate normalized to the symbol rate;

ω _σ is the frequency offset coarse estimation obtained by the acquisition module, normalized digital frequency relative to the chip rate, R _c is the chip rate, R _c =R _s ·L. represents the initial (digital) frequency offset of the residual carrier, assuming that the initial phase is

b _n ∈ 0/1 represents the transmitted binary symbol (BPSK modulation); the function of the demodulation module is to complete the frequency offset change rate, frequency offset search, compensation and phase discrimination. The demodulated symbol is denoted as z(n), which is The expression is:

make

正确估计并补偿，即

此时解调结果

If β, Δω and

Correctly estimate and compensate, i.e.

The demodulation result at this time

Traditional BPSK coherent demodulation modules usually use phase-locked loops, such as square loops, Costas loops, etc., to achieve carrier tracking, but they cannot effectively solve the "inverted π" problem, that is, the recovered local carrier and the required The coherent carrier wave of the signal may be in phase or in opposite phase. This phase uncertainty will cause the demodulated digital baseband signal to be exactly opposite, and all the digital symbols judged to be wrong.

SUMMARY OF THE INVENTION

The purpose of the device is to overcome the technical defects of the traditional phase-locked loop coherent demodulation device with phase ambiguity and slow convergence speed, and propose a low-signal-to-noise ratio short-frame burst communication open-loop demodulation method and device. Offset and frequency offset change rates are searched and compensated to complete phase discrimination and phase correction, ensuring that the direct sequence spread spectrum signal can be demodulated normally in a short time and under the condition of low signal-to-noise ratio.

The low-signal-to-noise ratio short-frame burst communication open-loop demodulation method and device include a low-signal-to-noise ratio short-frame burst communication open-loop demodulation device, referred to as the device for short; and a low-signal-to-noise ratio short frame burst communication open-loop demodulation device method, referred to as this method.

The device includes a two-dimensional search module, a fine search module 1, a fine search module 2, a fine search module 3 and a frame synchronization module;

Among them, the fine search module 1, the fine search module 2 and the fine search module 3 are multiplexed, and the three fine search modules multiplexed in this device are exactly the same;

The input data of the device is the output of the despreading module, and the output data of the device is the demodulation result, which is sent to the decoding module;

The connection relationship of each module of this device is as follows:

The two-dimensional search module is connected to the fine search module 1; the fine search module 1 is connected to the fine search module 2, the fine search module 2 is connected to the fine search module 3, and the fine search module 3 is connected to the frame synchronization module;

The functions of each module in this device are as follows:

The function of the two-dimensional search module is to carry out two-dimensional search, complete the joint estimation of the initial frequency offset and frequency offset change rate of the received signal, obtain the estimation of the initial frequency offset and frequency offset change rate, and calculate the Doppler frequency offset change rate and the frequency offset change rate. The Doppler frequency offset is compensated; the function of the fine search module is to complete the estimation and compensation of frequency offset, phase estimation and phase compensation; the function of the frame synchronization module is to solve the phase ambiguity.

This method is achieved through the following technical solutions:

The device adopts the method of segmented data processing for input data, including the following steps:

Step 1: Run a two-dimensional search module for Doppler frequency offset change rate and Doppler frequency offset change rate based on the input data of the device to complete the joint estimation of the initial frequency offset and frequency offset change rate of the received signal, and obtain the initial frequency offset and the estimation of the rate of change of frequency offset;

Wherein, the input data of the device is expressed as the following formula (1), also referred to as the first segment of data;

为初始相位；b_n∈0/1代表发送的二进制符号，β是数字化的频偏变化率，简称频偏变化率；紨为指数函数；In the above formula (1), Δω represents the initial frequency offset of the residual carrier, and the initial frequency offset of the residual carrier is the digital frequency offset, also called the frequency offset;

is the initial phase; b _n ∈ 0/1 represents the sent binary symbol, β is the digitized frequency offset change rate, referred to as the frequency offset change rate; 紨 is an exponential function;

和

等效为如下公式(2)中的二维最优化问题的解：Maximum Likelihood Estimation of β and Δω in Eq. (1) in a Gaussian White Noise Environment

and

It is equivalent to the solution of the two-dimensional optimization problem in the following formula (2):

即

按照

对x[n]进行Q次“解线性调频调制”，简称“解线调”，对应的符号x_q[n]表达为如下公式(3)：The approximate solution can be obtained by two-dimensional grid search, and the entire uncertainty range of the frequency offset and frequency offset change rate is divided into a two-dimensional plane grid according to a certain accuracy, and the way to obtain a two-dimensional plane is: for the frequency offset change rate, in its Within the variation range, take Q

which is

according to

Perform Q times of "de-chirp modulation" on x[n], referred to as "de-linear modulation", and the corresponding symbol x _q [n] is expressed as the following formula (3):

和

其中，

表示步骤一对频偏变化率的估计结果，

表示步骤一对初始频偏的估计结果；Fill x _q [n] with zeros until the length is K points, and then perform the K-point FFT operation to obtain x _q [k]; take the modulo of x _q [k], the result is |x _q [k]|; Find the maximum value of all Q×K modulo results |x _q [k]|, and obtain β and Δω estimators according to the position of this maximum value on the two-dimensional grid

and

in,

represents the estimation result of a pair of frequency offset change rates of the step,

represents the estimation result of a pair of initial frequency offsets in the step;

和

来对输入数据的频偏变化率和频偏进行补偿；The purpose of this first piece of data is to obtain an estimate of the rate of change of frequency offset and frequency offset, and there is no need to perform subsequent operations on this piece of data, and then the baseband waveform of the (N+1) symbol enters the despreading module.

and

to compensate the frequency offset change rate and frequency offset of the input data;

Complete the joint estimation of the initial frequency offset and the frequency offset change rate of the received signal, and obtain the estimation of the initial frequency offset and the frequency offset change rate. Starting from the second piece of data, the compensation of the frequency offset change rate uses the frequency obtained from the first piece of data. The estimated result of the offset change rate, and the default frequency offset change rate has been compensated. Even if there is an error in the estimation result of the frequency offset change rate obtained in step 1, that is, there is a residual Doppler frequency offset change rate, as long as the processing length of the second segment of data and the value of Q are appropriate, the residual Doppler frequency offset can be completely ignored. The effect of the rate of change on the frequency offset estimation;

Each piece of data only needs to estimate and compensate the frequency offset to complete the data demodulation;

Step 2: Run the fine search module based on the estimation result of the initial frequency offset obtained from the first piece of data, and perform compensation based on the accurate estimation result of the frequency offset obtained from the second piece of data, thereby completing data demodulation;

The second step is as follows:

Step 2.1 Compensate based on the estimated value of the initial frequency offset in step 1, and obtain the second data of formula (4):

是第二段数据的初始相位、M为本段数据的处理长度；为简化数学推导利用换元法将第二段数据的下标N+1≤n≤N+M改成0≤n≤M紨1；Among them, Δω ₂ is the frequency offset of the second segment of data,

is the initial phase of the second segment of data, and M is the processing length of the segment of data; in order to simplify the mathematical derivation, the subscript N+1≤n≤N+M of the second segment of data is changed to 0≤n≤M by using the substitution method紨1;

Step 2.2 Run the fine search module to obtain an accurate estimate of the frequency offset;

The fine search module of the device is run, and the fine search module uses the exhaustive method to obtain an accurate estimation of the frequency offset, and then obtains the coherent demodulation result of the second segment of data, specifically:

Step 2.21 First, square the second segment of data (4) to modulate, and then do the inner product with the square of the local inner product template. The expression of the local inner product template is as formula (5):

与x²[n]内积，再取内积模值最大者为复正弦波数字频率为-2Δω₂的一个估计，基于此估计得到

用于补偿频偏；相位

按照下式(6)估计：The specific method of doing the inner product with the square of the local inner product template is to divide Δω ₂ into P small intervals at equal intervals, and use the square of P complex sine waves with different digital frequencies.

The inner product with x ² [n], and the maximum value of the inner product modulus is an estimate of the complex sine wave digital frequency of -2Δω _2. Based on this estimate, we get

Used to compensate frequency offset; phase

It is estimated according to the following formula (6):

和

都满足公式(6)的要求，由于

初始值未知，在这里

取

或者

都可以；当数据处理到帧头段时，利用帧头的相关结果极性，来确定绝对相位，不妨先取

得到第二段数据相干解调结果为下式(7)：The "divide by 2" operation in the above equation (6) will result in phase ambiguity, i.e.

and

meet the requirements of formula (6), because

The initial value is unknown, here

Pick

or

can be; when the data is processed to the frame header segment, the absolute phase is determined by using the polarity of the correlation result of the frame header.

The coherent demodulation result of the second segment of data is obtained as the following formula (7):

Step 3: Accurately estimate and compensate the frequency offset of the data in this section, and deblur the phase of the data in this section, that is, the phase of the data in the previous section needs to be consistent;

The third step is as follows:

The third piece of data is:

The data in this section is:

Among them, when step 3 is run for the first time, this segment of data refers to the third segment of data, and so on;

Step 3.1: Run the fine search module to obtain an accurate estimate of the frequency offset; the fine search module uses the exhaustive method to obtain the accurate estimate of the frequency offset, and then obtains the coherent demodulation result of this section of data, specifically:

与x²[n]内积，再取内积模值最大者为复正弦波数字频率为-2Δω₃的一个估计，基于此估计得到

用于补偿频偏；相位

按照下式(9)估计：First, square the data of this segment to modulate, and then do the inner product with the square of the local inner product template. The expression of the local inner product template is as formula (5); the specific method of doing the inner product with the square of the local inner product template is to use Δω ₂ Divide into P sections at equal intervals, and use the square of P complex sine waves with different frequencies

Inner product with x ² [n], and then taking the largest value of the inner product modulus is an estimate of the complex sine wave digital frequency of -2Δω ₃ , based on this estimate, we get

Used to compensate frequency offset; phase

It is estimated according to the following formula (9):

和

都满足公式(9)的要求，即：得到频偏估计

同时得到相位

和

The "divide by 2" operation in the above equation (9) will result in phase ambiguity, i.e.

and

All meet the requirements of formula (9), that is: get the frequency offset estimate

get the phase at the same time

and

Step 3.2: Deblur the phase output in step 3.1, and the deblurred reference phase is constructed according to the results of the frequency offset and initial phase of the previous section, because the final phase of the previous section is the initial phase of this section;

频偏估计结果为

其末段相位应为(9)：For example, the initial phase estimation result of the second segment is:

The frequency offset estimation result is

Its final phase should be (9):

This is because the final phase and the initial phase are separated by N-1 symbols; if the despreading process of the spread spectrum system is considered, the initial phase of the third segment and the final phase of the second segment are slightly different, so (10) :

Since the second segment itself has phase ambiguity, the third segment and subsequent segments only need to strictly follow the phase of the previous segment to maintain a phase continuity; the specific processing process is as follows (11):

和

哪个更接近

若：judge

and

which is closer

like:

作为初相位的估计结果；否则，取

然后根据第三段的数据进行相干解调，并根据

对频偏进行补偿，第三段相干解调输出为：then take

As the estimation result of the initial phase; otherwise, take

Then perform coherent demodulation according to the data of the third segment, and according to

The frequency offset is compensated, and the third-stage coherent demodulation output is:

Step 4: Repeat step 3 based on the subsequent data segment until the pilot segment data processing ends; the obtained phase estimation result can be consistent with the phase of the previous segment of data;

Step 5. After the pilot segment data processing is completed in steps 1 to 4, the data processing of the frame header segment is entered. The specific processing method is to correlate the frame header data to obtain the frame header correlation result;

Among them, the frame header is usually an M sequence or a Gold code sequence, and the frame header is the beginning of the data segment;

The frame header outputs data z[n] through despreading, frequency estimation, compensation and phase estimation, and performs the correlation operation of the following formula (14) with the local frame header template:

The correlation process of this formula (14) is the same as the operation of the despreading process, the difference is that the correlation operation is performed at the chip level in the despreading process, and the despreading is the time length of the correlation length of one symbol; Correlation operation performed at the level, the length of the correlation operation depends on the length of the frame header; r _header [n] is the frame header correlation result; s _{header_tep} [n] is the frame header local template; M is the frame header length;

Step 6: Correlate the frame header correlation result with the local template, determine whether the frame synchronization is successful according to whether the absolute value of the correlation result exceeds a predetermined threshold, and perform corresponding operations, specifically: taking the frame header correlation operation result r _header [n] Modulo operation, if the modulo result exceeds the threshold, it is considered that the frame synchronization is successful, and skips to step 7 to start the decoding device; otherwise, it is considered that the data of this frame fails to receive, and returns to step 1 to wait for the arrival of the next frame of data;

Step 7. When the frame synchronization is successful, the phase is de-blurred according to the polarity of the correlation result; the details are as follows: according to the problem of "inverted π", when the threshold is exceeded, it is necessary to confirm the positive and negative of the correlation result when the threshold is exceeded; if If the polarity is positive, it means that the phase held at this time is 0, and z[n] is directly output into the decoding device; if the polarity of the correlation result is negative, it means that the phase held at this time is π, and z[n] needs to be inverted. , and then to the decoding module; after the decoding is completed, the data of this frame is successfully received, and the device returns to step 1 to wait for the arrival of the next frame of data;

So far, from step 1 to step 7, the open-loop demodulation method for short frame burst communication with low signal-to-noise ratio is completed.

beneficial effect

Compared with the prior art, the method and device for open-loop demodulation of short frame burst communication with low signal-to-noise ratio proposed by the present invention have the following beneficial effects:

1. Accurate estimation of Doppler frequency offset and frequency offset change rate can be completed in a very short time by using methods such as delineation and two-dimensional search;

2. Through the segmentation processing method, the exhaustive method is used to finely estimate and compensate the frequency offset of each segment, so as to realize the tracking of the carrier frequency and complete the phase identification;

3. Use the frame header polarity to correct the phase to ensure that the data information can be demodulated correctly, suitable for short frame burst direct sequence spread spectrum system carrier frequency offset and frequency offset change rate under high dynamic, low signal-to-noise ratio and other environments open-loop search, tracking, and symbol-coherent demodulation.

Description of drawings

Fig. 1 is the principle block diagram of the open-loop demodulation method and device for low-signal-to-noise ratio short frame burst communication according to the present invention and the device in Embodiment 1;

2 is a two-dimensional search plane of intermediate frequency offset and frequency offset change rate in the first embodiment of the open-loop demodulation method and device for low-signal-to-noise ratio short frame burst communication according to the present invention;

Fig. 3 is the demodulation principle in step 1 of the embodiment of the low signal-to-noise ratio short frame burst communication open-loop demodulation method and device of the present invention;

FIG. 4 is a schematic block diagram of the frequency fine search by exhaustive method in the method and apparatus for open-loop demodulation of short frame burst communication with low signal-to-noise ratio according to the present invention.

Detailed ways

The device will be described in further detail below with reference to examples and accompanying drawings.

This example takes the direct sequence spread spectrum BPSK communication receiver as an example, and its baseband digital signal processing part is shown in Figure 1.

It can be seen from Figure 1 that the despread data is entered into the device, and the Doppler frequency offset change rate and Doppler frequency offset two-dimensional search module, fine search module and frame synchronization module of the device are sequentially run to complete the demodulation function. , and output the data to the decoding module of the receiver to complete the decoding.

单位Hz。频偏变化率ΔΔf范围±6，单位kHz/s。Input data parameter index of this device: symbol rate R _s =2.5ksps; initial frequency offset Δf range

The unit is Hz. Frequency deviation rate of change ΔΔf range ±6, unit kHz/s.

对x[n]进行解线调，如图2所示，频偏变化率ΔΔf搜素范围±6kHz/s,

搜索精度250Hz/s(相邻两个q值之间相差250Hz/s)，Q值为

解线调表达式为The despread data is input into the device, and the prior information of the data frequency offset and frequency offset change rate is Δf=-130.9140Hz; ΔΔf=-3253.3Hz/s. The length of the first segment of data is 128 symbols, and all are pilots (the baseband data are all "0"). After entering the Doppler frequency offset change rate and Doppler frequency offset two-dimensional search module, divide the initial frequency offset and the entire uncertainty range of the frequency offset change rate into two-dimensional plane grids according to a certain precision.

Delineate x[n], as shown in Figure 2, the frequency offset change rate ΔΔf search range is ±6kHz/s,

The search accuracy is 250Hz/s (the difference between two adjacent q values is 250Hz/s), and the Q value is

The demodulation expression is

对49组解线调结果x_q[n]补128个零，做256点FFT，得到49组x_q[k]，每组x_q[k]长度256。将全部的49×256的取模结果|x_q[k]|画在网格上，根据最大值对应的坐标，得出此段信号的频偏和频偏变化率。如附图3，得到的频偏为-136.7188Hz,频偏变化率-3000Hz/s，即转化为数字域上得到的频偏估计量

频偏变化率

分别对频偏变化率和频偏进行补偿。In the above formula

49 sets of demodulation results x _q [n] are filled with 128 zeros, and 256-point FFT is performed to obtain 49 sets of x _q [k], each set of x _q [k] has a length of 256. Draw all the 49×256 modulo results |x _q [k]| on the grid, and obtain the frequency offset and frequency offset change rate of this signal according to the coordinates corresponding to the maximum value. As shown in Figure 3, the obtained frequency offset is -136.7188Hz, and the frequency offset change rate is -3000Hz/s, which is converted into the frequency offset estimator obtained in the digital domain

Frequency deviation rate of change

Compensate for frequency offset change rate and frequency offset respectively.

进行补偿，即使在第一段估计出来的

有误差，误差最大为500Hz/s，在64个符号的时间长度，频率积累为

如果本段频偏估计得出的频率超过12.8Hz，认为本段频率估计有错，维持前一次频率估计结果，只对每段数据估计出的频偏进行更新并补偿。由数字信号处理知识可知，第一段数据处理后参与频偏补偿的精度为

所以从第二段数据开始采用穷举法来得到频偏更加精确的估计。Starting from the second segment of data, the length of the data processing is 64, and the default is that the frequency offset change rate has been estimated using the first segment of the frequency offset change rate.

compensated even if the first paragraph estimated

There is an error, the maximum error is 500Hz/s, in the time length of 64 symbols, the frequency accumulation is

If the estimated frequency of this segment exceeds 12.8Hz, it is considered that the estimated frequency of this segment is wrong, the previous frequency estimation result is maintained, and only the estimated frequency offset of each segment of data is updated and compensated. From the knowledge of digital signal processing, it can be known that the accuracy of participating in frequency offset compensation after the first segment of data processing is:

Therefore, the exhaustive method is used to obtain a more accurate estimate of the frequency offset from the second piece of data.

The second piece of data entered into the fine search module is:

不同频率模板平方与x²[n]内积，共得到64组内积结果，内积结果表达式为：As shown in Figure 4, the second segment of data x[n] is squared to de-modulate, and then the exhaustive method is used, that is, 64 complex sine wave templates with different digital frequencies are used, and the frequency interval between two groups of adjacent templates is

The square of different frequency templates and the inner product of x ² [n], a total of 64 sets of inner product results are obtained. The inner product result expression is:

取内积结果模值最大者，估计出复正弦波数字频率为-2Δω₂,根据仿真结果，

的值最大，此时p＝28。所以本段数据的频率估计为

In the above formula

Taking the largest modulus value of the inner product result, the digital frequency of the complex sine wave is estimated to be -2Δω ₂ . According to the simulation results,

The value of is the largest, at this time p=28. So the frequency of this data is estimated to be

相位估计按照：

Phase estimation follows:

即可。用此段数据估计得到的

对频偏进行补偿。此段数据的解调结果为Dividing by 2 will cause phase ambiguity. Since the second segment of data has no data prior information, it is directly selected.

That's it. estimated from this data

Compensate for frequency offset. The demodulation result of this piece of data is

和

估计方法和第二段数据估计方法一样，通过精细搜索模块，得到本段数据的频偏估计

对频偏进行补偿in the third data

and

The estimation method is the same as the estimation method of the second section of data. Through the fine search module, the frequency offset estimation of this section of data is obtained.

Compensate for frequency offset

Starting from the third segment of data, the data needs to be de-fuzzed to ensure that the phase is consistent with the previous segment.

The initial phase of the third segment is

作为初相位的估计结果,然后根据第三段的数据进行相干解调，此段数据的解调结果为：then take

As the estimation result of the initial phase, then perform coherent demodulation according to the data of the third segment. The demodulation result of this segment of data is:

The processing flow of the fourth and subsequent sections is exactly the same as that of the third section. When the frame synchronization is successful, the polarity of the correlation result is negative. After the frame header ends, the demodulation result of the data segment needs to take the z[n] symbol. On the contrary, it is output to the decoding device to complete the decoding.

test results

条件下，仿真得到误码率为0.395，理论误码率为0.375，与理论值基本符合。Statistics on the bit error rate of the demodulation result of this segment of data, 10,000 symbols, signal-to-noise ratio

Under the conditions of simulation, the bit error rate is 0.395 and the theoretical bit error rate is 0.375, which is basically consistent with the theoretical value.

The demodulation device proposed by the present invention can estimate the frequency offset and the frequency offset change rate in a short time under the condition of low signal-to-noise ratio and make real-time compensation, so as to realize the accurate tracking of the frequency offset and the frequency offset change rate of the signal by the receiver. , to complete the phase discrimination and phase correction, to ensure that the demodulated data bit error rate is consistent with the theoretical value.

To sum up, the above are only preferred embodiments of the present invention, and are not intended to limit the protection scope of the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims (5)

1. The low signal-to-noise ratio short frame burst communication open-loop demodulation device is characterized in that: it comprises a Doppler frequency offset change rate and a Doppler frequency offset change rate two-dimensional search module, a fine search module and a frame synchronization module; Among them, the fine search modules are multiplexed, and the multiplexed fine search modules in this device are exactly the same; The input data of the device is the output of the despreading module, and the output data of the device is the demodulation result, which is sent to the decoding module; The connection relationship of each module of this device is as follows: The two-dimensional search module is connected to the fine search module 1; the fine search module 1 is connected to the fine search module 2, the fine search module 2 is connected to the fine search module 3, and the fine search module 3 is connected to the frame synchronization module; The functions of each module in this device are as follows: The function of the two-dimensional search module for Doppler frequency offset change rate and Doppler frequency offset change rate is to perform two-dimensional search, complete the joint estimation of the initial frequency offset and frequency offset change rate of the received signal, and obtain the initial frequency offset and frequency offset. The function of the fine search module is to complete the estimation and compensation of frequency offset, phase estimation and phase compensation; the function of the frame synchronization module is to solve the Phase blur. 2. in the low-signal-to-noise ratio short-frame burst communication open-loop demodulation device as claimed in claim 1, adopt the low-signal-to-noise ratio short frame burst communication open-loop demodulation method, it is characterized in that: comprise the steps: Step 1: Run a two-dimensional search module for Doppler frequency offset change rate and Doppler frequency offset change rate based on the input data of the device to complete the joint estimation of the initial frequency offset and frequency offset change rate of the received signal, and obtain the initial frequency offset and the estimation of the rate of change of frequency offset; Wherein, the input data of the device is also referred to as the first segment of data; Step 2: Based on the estimation result of the initial frequency offset obtained from the first piece of data, run the fine search module 1, fine search module 2 and fine search module 3 to obtain the second piece of data, and obtain an accurate estimation of the frequency offset based on the second piece of data The result and compensation are performed to complete the data demodulation; Step 3: Accurately estimate and compensate the frequency offset of the data in this section, and deblur the phase of the data in this section, that is, the phase of the data in the previous section needs to be consistent; Step 4: Repeat step 3 based on the subsequent data segment until the pilot segment data processing ends; the obtained phase estimation result can be consistent with the phase of the previous segment of data; Step 5. After the pilot segment data processing is completed in steps 1 to 4, the data processing of the frame header segment is entered. The specific processing method is to correlate the frame header data to obtain the frame header correlation result; Among them, the frame header is usually an M sequence or a Gold code sequence, and the frame header is the beginning of the data segment; The frame header outputs data z[n] through despreading, frequency estimation, compensation and phase estimation, and performs the correlation operation of the following formula (14) with the local frame header template:

The correlation process of this formula (14) is the same as the operation of the despreading process, the difference is that the correlation operation is performed at the chip level in the despreading process, and the despreading is the time length of the correlation length of one symbol; Correlation operation performed at the level, the length of the correlation operation depends on the length of the frame header; r _header [n] is the frame header correlation result; s _{header_tep} [n] is the frame header local template; M is the frame header length; Step 6: Correlate the frame header correlation result with the local template, determine whether the frame synchronization is successful according to whether the absolute value of the correlation result exceeds a predetermined threshold, and perform corresponding operations, specifically: taking the frame header correlation operation result r _header [n] Modulo operation, if the modulo result exceeds the threshold, it is considered that the frame synchronization is successful, and skips to step 7 to start the decoding device; otherwise, it is considered that the data of this frame fails to receive, and returns to step 1 to wait for the arrival of the next frame of data; Step 7. When the frame synchronization is successful, the phase is de-blurred according to the polarity of the correlation result; the details are as follows: according to the problem of "inverted π", when the threshold is exceeded, it is necessary to confirm the positive and negative of the correlation result when the threshold is exceeded; if If the polarity is positive, it means that the phase held at this time is 0, and z[n] is directly output into the decoding device; if the polarity of the correlation result is negative, it means that the phase held at this time is π, and z[n] needs to be inverted. , and then to the decoding module; after the decoding is completed, the data of this frame is successfully received, and the device returns to step 1 to wait for the arrival of the next frame of data; So far, from step 1 to step 7, the open-loop demodulation method for short frame burst communication with low signal-to-noise ratio is completed. 3. The low-signal-to-noise ratio short-frame burst communication open-loop demodulation method according to claim 2, characterized in that: in step 1, the input data of this device is expressed as the following formula (1):

In the above formula (1), Δω represents the initial frequency offset of the residual carrier, and the initial frequency offset of the residual carrier is the digital frequency offset, also called the frequency offset;

is the initial phase; b _n ∈ 0/1 represents the sent binary symbol, β is the digitized frequency offset change rate, referred to as the frequency offset change rate; e is an exponential function; Maximum Likelihood Estimation of β and Δω in Eq. (1) in a Gaussian White Noise Environment

and

It is equivalent to the solution of the two-dimensional optimization problem in the following formula (2):

Specifically, the approximate solution is carried out through a two-dimensional grid search, and the entire uncertainty range of the frequency offset and the frequency offset rate of change is divided into a two-dimensional plane grid according to a certain accuracy, and the method of obtaining a two-dimensional plane is: Within the variation range, take Q

which is

according to

Perform Q times of "de-chirp modulation" on x[n], referred to as "de-linear modulation", and the corresponding symbol x _q [n] is expressed as the following formula (3):

Fill x _q [n] with zeros until the length is K points, and then perform the K-point FFT operation to obtain x _q [k]; take the modulo of x _q [k], the result is |x _q [k]|; Find the maximum value of all Q×K modulo results |x _q [k]|, and obtain β and Δω estimators according to the position of this maximum value on the two-dimensional grid

and

in,

represents the estimation result of a pair of frequency offset change rates of the step,

represents the estimation result of a pair of initial frequency offsets in the step; The purpose of this first piece of data is to obtain an estimate of the rate of change of frequency offset and frequency offset, and there is no need to perform subsequent operations on this piece of data, and then the baseband waveform of the (N+1) symbol enters the despreading module.

and

to compensate the frequency offset change rate and frequency offset of the input data; Complete the joint estimation of the initial frequency offset and the frequency offset change rate of the received signal, and obtain the estimation of the initial frequency offset and the frequency offset change rate. Starting from the second piece of data, the compensation of the frequency offset change rate uses the frequency obtained from the first piece of data. The estimation result of the offset change rate, and the default frequency offset change rate has been compensated; even if there is an error in the estimation result of the frequency offset change rate obtained in step 1, that is, there is a residual Doppler frequency offset change rate, as long as the second data processing The value of length and Q is appropriate, and the influence of the residual Doppler frequency offset change rate on the frequency offset estimation can be completely ignored; each piece of data only needs to estimate and compensate the frequency offset to complete the data demodulation. 4. The low-signal-to-noise ratio short-frame burst communication open-loop demodulation method according to claim 3, characterized in that: step 2, specifically: Step 2.1 Compensate based on the estimated value of the initial frequency offset in step 1, and obtain the second data of formula (4):

Among them, Δω ₂ is the frequency offset of the second segment of data,

is the initial phase of the second segment of data, and M is the processing length of the segment of data; in order to simplify the mathematical derivation, the subscript N+1≤n≤N+M of the second segment of data is changed to 0≤n≤M by using the substitution method -1; Step 2.2 Run the fine search module to obtain an accurate estimate of the frequency offset; The fine search module of the device is run, and the fine search module uses the exhaustive method to obtain an accurate estimation of the frequency offset, and then obtains the coherent demodulation result of the second segment of data, specifically: Step 2.21 First, square the second segment of data (4) to modulate, and then do the inner product with the square of the local inner product template. The expression of the local inner product template is as formula (5):

The specific method of doing the inner product with the square of the local inner product template is to divide Δω ₂ into P small intervals at equal intervals, and use the square of P complex sine waves with different digital frequencies.

The inner product with x ² [n], and the maximum value of the inner product modulus is an estimate of the complex sine wave digital frequency of -2Δω _2. Based on this estimate, we get

Used to compensate frequency offset; phase

It is estimated according to the following formula (6):

The "divide by 2" operation in the above equation (6) will result in phase ambiguity, i.e.

and

meet the requirements of formula (6), because

The initial value is unknown, here

Pick

or

can be; when the data is processed to the frame header segment, the absolute phase is determined by using the polarity of the correlation result of the frame header.

The coherent demodulation result of the second segment of data is obtained as the following formula (7):

5. The low-signal-to-noise ratio short-frame burst communication open-loop demodulation method according to claim 4, wherein step 3 is specifically: The data in this section is:

Among them, when step 3 is run for the first time, this segment of data refers to the third segment of data, and so on; Step 3.1: Run the fine search module to obtain an accurate estimate of the frequency offset; the fine search module uses the exhaustive method to obtain the accurate estimate of the frequency offset, and then obtains the coherent demodulation result of this section of data, specifically: First, square the data of this segment to modulate, and then do the inner product with the square of the local inner product template. The expression of the local inner product template is as formula (5); the specific method of doing the inner product with the square of the local inner product template is to use Δω ₂ Divide into P sections at equal intervals, and use the square of P complex sine waves with different frequencies

Inner product with x ² [n], and then taking the largest value of the inner product modulus is an estimate of the complex sine wave digital frequency of -2Δω ₃ , based on this estimate, we get

Used to compensate frequency offset; phase

It is estimated according to the following formula (9):

The "divide by 2" operation in the above equation (9) will result in phase ambiguity, i.e.

and

All meet the requirements of formula (9), that is: get the frequency offset estimate

get the phase at the same time

and

Step 3.2: Deblur the phase output in step 3.1, and the deblurred reference phase is constructed according to the results of the frequency offset and initial phase of the previous section, because the final phase of the previous section is the initial phase of this section; For example, the initial phase estimation result of the second segment is:

The frequency offset estimation result is

Its final phase should be (9):

This is because the final phase and the initial phase are separated by N-1 symbols; if the despreading process of the spread spectrum system is considered, the initial phase of the third segment and the final phase of the second segment are slightly different, so (10) :

Since the second segment itself has phase ambiguity, the third segment and subsequent segments only need to strictly follow the phase of the previous segment to maintain a phase continuity; the specific processing process is as follows (11):

judge

and

which is closer

like:

then take

as the estimation result of the initial phase; otherwise, take

Then perform coherent demodulation according to the data of the third segment, and according to

The frequency offset is compensated, and the third-stage coherent demodulation output is:

Images