A time domain O&M timing synchronization method for serial high-efficiency communication

Through the serial efficient communication time domain O&M timing synchronization method, the timing error is estimated using IQ path parallelism and DFT operations, and the rapid synchronization and error compensation of high-speed communication are achieved, solving the resource limitation of traditional algorithms in high-speed large bandwidth transmission and The problem of extending synchronization time improves signal-to-noise ratio and hardware resource utilization efficiency.

CN115996162BActive Publication Date: 2024-11-26THE 54TH RESEARCH INSTITUTE OF CHINA ELECTRONICS TECHNOLOGY GROUP CORPORATION
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202211474592.1
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-11-23
Publication Date
2024-11-26
Estimated Expiration
2042-11-23

AI Technical Summary

Technical Problem

The traditional O&M timing synchronization algorithm cannot meet the requirements of high-speed, large-bandwidth data transmission, and the ADC cannot sample at the best sampling time of the symbol, resulting in signal deviation and synchronization time extended.

Method used

Using serial efficient communication time domain O&M timing synchronization method, through IQ path parallelism and symbol non-integer multiple sampling design, DFT operation estimates timing errors, and dynamically controls the position of filter coefficient taps in the time domain to complete timing error compensation and reduce Hardware resource occupancy and synchronization time.

Benefits of technology

Under the GiHz ultra-large bandwidth and ultra-high-speed modem and demodulation system, fast timing synchronization is achieved, signal-to-noise ratio is improved, hardware resource use is reduced, real-time and efficient of high-speed communication is met, and algorithm structure is simplified.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN115996162B_ABST
    Figure CN115996162B_ABST
Patent Text Reader

Abstract

The present invention discloses a time domain O&M timing synchronization method for serial efficient communication, which belongs to the field of terahertz millimeter wave high-speed communication. This method uses ADC to complete signal sampling, extracts complex sequence after data capture to the timing synchronization module for processing; calculates the modulus square of the pilot sequence, that is, the envelope amplitude, and sends the result to the IQ two-way DFT operation respectively; uses the real part and the imaginary part after the DFT operation to calculate the timing error estimate; calculates the filter tap position based on the timing error calculation result, which is used to index the multi-phase filter coefficient, and the multi-phase filter completes the timing error compensation and signal data rate conversion in the time domain. The present invention adopts an open-loop IQ two-way parallel and symbol non-integer multiple sampling design, and each IQ channel processes the sample data serially, and performs timing error compensation in the time domain, avoiding DFT and IDFT operations while further reducing the occupation of hardware resources.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the field of terahertz millimeter wave communication, and in particular to a serial high-efficiency time-domain O&M timing synchronization method for communication bit synchronization. Background Art

[0002] In the field of software radio, the existence of transmission delay and clock offset at the transmitting and receiving ends will cause deviations in the received signal when the signal passes through the channel. The data after sampling and down-conversion at the receiving end is a sequence of code elements. Only by adjusting the sampling clock at the receiving end to make it consistent with the phase of the received code element frequency can the sampling decision be made at the best sampling time of the received symbol.

[0003] Bit synchronization can be divided into two categories according to whether there is data assistance or not: data-assisted and non-data-assisted. The latter does not require additional auxiliary sequences, has higher spectrum utilization than data-assisted methods, and is more widely used in engineering. The more representative non-data-assisted bit synchronization algorithms are the Gardner algorithm and the O&M algorithm. The O&M algorithm was proposed by OerderM and MeyrH in 1988 to estimate timing errors using square-law nonlinear transformations.

[0004] However, the traditional O&M timing synchronization algorithm does not meet the requirements of high-speed and large-bandwidth data transmission. Summary of the invention

[0005] The purpose of the present invention is to provide a serial high-efficiency communication time domain O&M timing synchronization method, which can compress the use of hardware logic resources from the operation architecture, reduce resource consumption while ensuring ultra-high-speed modulation rate transmission, and shorten the timing synchronization time.

[0006] The technical solution adopted by the present invention is as follows:

[0007] A serial high-efficiency communication time domain O&M timing synchronization method, comprising the following steps:

[0008] Step 1, sampling the signal, completing the frame synchronization of the sampled data, performing timing synchronization processing on the complex sequence after frame synchronization, and obtaining a complex timing synchronization sequence;

[0009] Step 2, perform I and Q serial input timing processing on the complex timing synchronization sequence, multiply each bit by bit in the form of binary symbols and accumulate and sum to obtain a square value, and then sum the results obtained from the I and Q paths to obtain the received signal module square;

[0010] Step 3: Send the signal envelope value to the I and Q channels respectively, and each channel signal is respectively correlated with the corresponding rotation factor W nI and W nq Multiply and sum to complete the DFT operation;

[0011] Step 4, using the real part and imaginary part of the signal after DFT operation, the timing error is estimated to calculate the frequency domain phase offset, that is, the timing error estimation result;

[0012] Step 5: perform loop filtering and smoothing based on the timing error estimation result, calculate the timing synchronization time domain delay, and correspond to the filter tap position for subsequent processing;

[0013] Step 6, the filter coefficients corresponding to the different timing error estimation results are stored in the ROM, and the filter tap position outputted in step 5 is used as the ROM index address to dynamically control the ROM to output the filter coefficients;

[0014] Step 7, input the filter coefficients output by the ROM into the polyphase filter module to complete the sampling signal timing error compensation.

[0015] Furthermore, the specific method of step 1 is to use an analog-to-digital conversion module connected to the FPGA baseband board to sample the signal, the sampled signal completes frame synchronization in the baseband board, and the output result includes a pilot sequence extracted from the received sampled data, and the sequence is used for timing error estimation.

[0016] Furthermore, in the DFT operation of step 3, a sampling rate that satisfies the Nyquist sampling theorem is used, the signal is in complex form, and the DFT operation is expressed as:

[0017]

[0018] Among them, x k is the receiving signal model; X m For x k The Fourier transform of I and Q is represented by W respectively. nI =Im[e -j2kπ / N ] and W nq =Re[e -j2kπ / N ]; N is the number of DFT points.

[0019] Furthermore, in step 4, X m The real part of corresponds to the sampling point modulus and the rotation factor W nI The product of m The imaginary part corresponds to the sampling point modulus and the rotation factor W nq The symbol accumulation length is L, and the I and Q paths are processed serially respectively; X m The real and imaginary parts of are expressed as:

[0020]

[0021]

[0022] Among them, Re[X m ] and Im[X m ] represent X m The real and imaginary parts, I k and Q k are the real and imaginary parts of the received sample signal respectively;

[0023] The DFT operation is represented by the real and imaginary signal operation results of the received signal. The timing error estimate of L symbols is

[0024]

[0025] Among them, the error estimate is an unbiased estimate of ε.

[0026] Furthermore, in step 5, the calculation method of the filter coefficient tap position for timing error compensation is:

[0027]

[0028] in, is the estimated value of the timing error, f us is the root cosine filter symbol upsampling multiple.

[0029] Furthermore, the specific method of step 7 is to convolve the filter coefficients with the signal samples in the time domain. The filter adopts a multi-phase structure design, and uses the different extraction and interpolation multiples of the multi-phase filter to achieve signal timing error compensation and complete data rate conversion at the same time.

[0030] The beneficial effects of the present invention are:

[0031] 1. In the gigahertz ultra-large bandwidth and ultra-high-speed modulation and demodulation system, the traditional square-law timing synchronization algorithm is no longer applicable. The present invention adopts IQ path parallel and symbol non-integer multiple sampling design. Under the limited hardware sampling rate, it not only meets the transmission rate requirements, but also avoids the problem of too many symbol operation points; the open-loop timing structure has the advantages of short timing synchronization time and simple algorithm implementation compared to the closed-loop structure.

[0032] 2. The present invention solves the problem that the ADC cannot sample at the optimal symbol sampling time, and improves the signal-to-noise ratio of the received signal. Compared with the existing algorithm, the structure with control signal feedback is bound to increase the time required for synchronization, slow down the convergence speed, and is not suitable for the demodulation of burst signals. The present invention only needs to extract the pilot sequence of part of the received data for timing synchronization algorithm processing, which meets the real-time and high efficiency of high-speed communication; the IQ path adopts a serial processing architecture respectively, which not only meets the transmission rate but also reduces the use of hardware internal logic resources, ensuring the stability of the algorithm performance. In addition, the present invention proposes a new error correction method, which estimates the filter coefficient tap position according to the timing error estimate and the filter upsampling multiple, and performs timing error estimation compensation in the time domain, while also eliminating the DFT and IDFT modules, effectively reducing the occupation of hardware resources and shortening the time required for the timing synchronization algorithm. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 is a flow chart of the method of the present invention.

[0034] Figure 2 It is a schematic diagram of the principle of the method of the present invention.

[0035] Figure 3 It is a spectrum diagram of the input signal module in the processing process of the present invention.

[0036] Figure 4 It is a comparison diagram of the QPSK modulation constellation points of the pilot sequence before and after the processing of the present invention. DETAILED DESCRIPTION

[0037] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in combination with specific implementation examples and with reference to the accompanying drawings.

[0038] A time domain O&M timing synchronization method for serial high-efficiency communication, the method uses ADC to complete signal sampling, extracts complex sequence after data capture to a timing synchronization module for processing; calculates the modulus square of the pilot sequence, i.e., the envelope amplitude, and sends the result to IQ two paths to complete DFT operation respectively; calculates the timing error estimation value by using the real part and the imaginary part after the DFT operation; obtains the filter tap position according to the timing error calculation result, and uses it to index the multiphase filter coefficient, and the multiphase filter completes the timing error compensation and signal data rate conversion in the time domain.

[0039] Specifically, the method comprises the following steps:

[0040] Step 1: Use the ADC analog-to-digital converter connected to the FPGA baseband board to complete signal sampling. The sampled signal is first synchronized for data capture. The captured signal determines the frame header, extracts the complex pilot sequence and outputs it to the timing synchronization module for timing error estimation.

[0041] Step 2: The sample obtained after the received signal r(t) is upsampled N times by the receiving end is:

[0042]

[0043] Among them, a n is the transmitted data symbol, g(t) is the total response of the shaping filter, channel and receiving filter, and ε is the timing error. After the signal is sampled and captured, the extracted pilot sequence can be expressed as an orthogonal complex vector r k =I k +jQ k , where k = 0, 1, 2, ..., N, the extracted pilot sequence data is used for timing error estimation.

[0044] The complex pilot sequence is serially input into the timing module through the I and Q channels. The binary symbols of each channel are multiplied bit by bit and accumulated to obtain their square values. The sum of the results obtained by the I and Q channels is the received signal modulus. The envelope value can be used express.

[0045] Step 3: Send the obtained signal envelope square to I and Q channels respectively, and each signal is respectively rotated with the corresponding rotation factor W nI and W nq Multiply and accumulate to complete the DFT operation. The sampling rate of this method satisfies the Nyquist sampling theorem. The signal is in the form of a complex signal, and its DFT transform can be expressed as:

[0046]

[0047] Where W nI =Im[e -j2kπ / N ] and W nq =Re[e -j2kπ / N ] are the rotation factors, and N is the number of DFT points.

[0048] Step 4: Modulate the signal x k Calculate the DFT to get X m The real and imaginary parts, X m The real part of corresponds to the sampling point modulus and the rotation factor W nI The product of m The imaginary part corresponds to the sampling point modulus and the rotation factor W nq The symbol accumulation length is L, and I and Q are processed serially. Then X m The real and imaginary parts of

[0049]

[0050]

[0051] At this point, the DFT operation can be represented by the real and imaginary signal operation results of the received signal, and the estimated timing error of L symbols can be calculated as Calculate the frequency domain phase offset and obtain the timing error estimation result.

[0052] Step 5: perform loop filtering and smoothing based on the timing error estimation result and calculate the timing synchronization time domain delay. The calculation formula for the filter coefficient tap position for timing error compensation is: in is the estimated value of the timing error, f us is the root cosine filter symbol upsampling multiple, the calculated result corresponds to the storage address of the coefficient in the ROM, and the index output filter coefficient is used for subsequent compensation processing.

[0053] Step 6: The filter coefficients corresponding to the different timing error estimation results are stored in the ROM, and the filter tap position outputted in step 5 is used as the ROM index address to dynamically control the ROM to output the filter coefficients.

[0054] Step 7, input the filter coefficients obtained by the index ROM into the multi-phase filter module, and convolve with the received signal sampling result in the time domain, so as to achieve data rate conversion while completing the sampling signal timing error compensation.

[0055] This method fully considers the limitation of hardware sampling rate. Under the premise of meeting the data transmission rate, it adopts the design of open-loop IQ parallel and symbol non-integer multiple sampling. The O&M algorithm of serial processing samples of each IQ channel is conducive to saving hardware internal resources and achieving efficient and stable data transmission. The O&M timing error result is sent to the filter for time domain timing error compensation. Compared with the frequency domain implementation, the time domain compensation omits the DFT and IDFT steps, further reducing the number of resources required to implement the algorithm.

[0056] Here is a more specific example:

[0057] like Figure 1 As shown, a serial efficient communication time domain O&M timing synchronization method, the specific implementation steps are as follows:

[0058] Step 1: Use the ADC analog-to-digital converter connected to the FPGA baseband board to complete signal sampling. The sampled signal is first synchronized for data capture. The captured signal determines the frame header, extracts the complex pilot sequence and outputs it to the timing synchronization module for timing error estimation.

[0059] Step 2: The sample obtained after the received signal r(t) is upsampled N times by the receiving end is:

[0060]

[0061] Among them, a n is the data symbol to be sent, g(t) is the total response of the shaping filter, channel and receiving filter, and ε is the timing error. After the signal is sampled and captured, the pilot sequence can be expressed as an orthogonal complex vector r k =I k +jQ k , where k = 0, 1, 2, ..., N, and N = 56, extract part of the pilot sequence for timing error estimation.

[0062] like Figure 2 As shown, the complex pilot sequence is serially input into the timing module through the I and Q channels respectively. Each channel is bit-wise multiplied in the form of binary symbols and the cumulative sum is used to obtain its square value. The sum of the results obtained by the I and Q channels is the square of the received signal. Its envelope value can be used express.

[0063] Step 3: Send the obtained signal envelope value to I and Q channels respectively, and each signal is respectively correlated with the corresponding rotation factor W nI and W nq Multiply and sum to complete the DFT operation.

[0064] The sampling rate of this method satisfies the Nyquist sampling theorem, and the signal is in the form of a complex signal, where the symbol rate f B =4.2GHz, signal bandwidth B = 2.1GHz, sampling rate f s =4.8GHz, the sampling multiple at the receiving end is 8 / 7 times sampling, then its DFT transform can be expressed as:

[0065]

[0066] Where W nI =Im[e -j2kπ / N ] and W nq =Re[e -j2kπ / N ] are the rotation factors, and N is the number of DFT points.

[0067] Step 4: Modulate the signal x k Calculate the DFT to get X m The real and imaginary parts, X m The real part of corresponds to the sampling point modulus and the rotation factor W nI The product of m The imaginary part corresponds to the sampling point modulus and the rotation factor W nq The symbol accumulation length is L = 56, and the number of FFT points in the simulation is 64 points. The result is as follows Figure 3 The power spectrum is shown; I and Q are processed serially, then X m The real and imaginary parts of

[0068]

[0069]

[0070] At this point, the DFT operation can be represented by the real and imaginary signal processing results of the received signal, and the timing error estimate of L symbols can be calculated as The frequency domain phase offset is calculated to obtain the timing error estimation result.

[0071] Step 5: perform loop filtering and smoothing based on the timing error estimation result and calculate the timing synchronization time domain delay. The calculation formula for the filter coefficient tap position of the timing error compensation is: in is the estimated value of the timing error, f us is the root cosine filter symbol upsampling multiple, the calculated result corresponds to the ROM storage address, and the filter coefficient output by the index is used for subsequent compensation processing.

[0072] Step 6: The filter coefficients corresponding to different timing error estimation values ​​are stored in ROM. The coefficients are binary numbers with a number N=64 and a bit width of 8. The filter tap position output in step 5 is used as the ROM index address to dynamically control the filter coefficients output by the ROM.

[0073] Step 7, input the filter coefficient obtained by the index ROM into the polyphase filter module, and convolve it with the received signal sampling result in the time domain, and realize data rate conversion while completing the sampling signal timing error compensation. The results before and after the timing error compensation are as follows: Figure 4 shown.

[0074] The following are the simulation results of this method:

[0075] 1. Conditions

[0076] Signal modulation: QPSK

[0077] FFT points: 64

[0078] Root raised cosine filter roll-off factor: 0.25

[0079] Root cosine filter symbol upsampling multiple: 112

[0080] 2. Results Analysis

[0081] This method completes the timing synchronization error estimation and compensation under the above conditions. The results are as follows Figure 4 The left figure is the constellation diagram before the timing synchronization error compensation, and the right figure is the constellation diagram after the timing synchronization error compensation. It can be clearly seen that this method reduces the error generated in the signal transmission process to a certain extent, and the signal sampling effect has been improved to a certain extent.

[0082] In summary, the present invention estimates the timing error of high-speed communication transmission data, calculates the phase delay error, and sends it to the filter, thereby completing the compensation of the time domain timing error. In view of the problem that the traditional O&M timing synchronization algorithm is limited in hardware sampling in high-speed and large-bandwidth transmission systems, the present invention adopts an open-loop IQ two-way parallel and symbol non-integer multiple sampling design, and each IQ channel processes the sample data serially, and performs timing error compensation in the time domain, avoiding DFT and IDFT operations while further reducing the occupation of hardware resources.

[0083] The above embodiments are only for illustrating the technical concept and features of the present invention, and their purpose is to enable people familiar with the technology to understand the content of the present invention and implement it accordingly, and they cannot be used to limit the protection scope of the present invention. Any equivalent changes or modifications made according to the spirit of the present invention should be included in the protection scope of the present invention.

Claims

1. A time domain O&M timing synchronization method for serial high-efficiency communication, It is characterized in that The following steps are involved: Step 1, sampling the signal, completing the frame synchronization of the sampled data, performing timing synchronization processing on the complex sequence after frame synchronization, and obtaining a complex timing synchronization sequence; the specific method is to use an analog-to-digital conversion module connected to the FPGA baseband card to sample the signal, the sampled signal completes frame synchronization in the baseband card, and the output result includes a pilot sequence extracted from the received sampled data, and the sequence is used for timing error estimation; Step 2, perform I and Q serial input timing processing on the complex timing synchronization sequence, multiply each bit by bit in the form of binary symbols and accumulate and sum to obtain a square value, and then sum the results obtained from the I and Q paths to obtain the received signal module square; Step 3: Send the signal envelope value to the I and Q channels respectively, and each channel signal is respectively correlated with the corresponding rotation factor W nI and W nq Multiply and sum to complete the DFT operation; in the DFT operation, a sampling rate that satisfies the Nyquist sampling theorem is used, the signal is in complex form, and the DFT operation is expressed as: Among them, x k is the receiving signal model; X m For x k The Fourier transform of I and Q is represented by W respectively. nI =Im[[e -j2kπ / N ]] and W nq =Re[[e -j2kπ / N ]]; N is the number of DFT points; Step 4: Use the real and imaginary parts of the signal after DFT operation to estimate the timing error and calculate the frequency domain phase offset, that is, the timing error estimation result; where X m The real part of corresponds to the sampling point modulus and the rotation factor W nI The product of m The imaginary part corresponds to the sampling point modulus and the rotation factor W nq The symbol accumulation length is L, and the I and Q paths are processed serially respectively; X m The real and imaginary parts of are expressed as: Among them, Re[X m ] and Im[X m ] represent X m The real and imaginary parts, I k and Q k are the real and imaginary parts of the received sample signal respectively; The DFT operation is represented by the real and imaginary signal operation results of the received signal. The timing error estimate of L symbols is Among them, the error estimate is an unbiased estimate of e; Step 5: perform loop filtering and smoothing based on the timing error estimation result, calculate the timing synchronization time domain delay, and correspond to the filter tap position for subsequent processing; wherein, the calculation method of the filter coefficient tap position for timing error compensation is: in, is the estimated value of the timing error, f us is the root cosine filter symbol upsampling multiple; Step 6, the filter coefficients corresponding to the different timing error estimation results are stored in the ROM, and the filter tap position outputted in step 5 is used as the ROM index address to dynamically control the ROM to output the filter coefficients; Step 7, input the filter coefficients output by the ROM into the multi-phase filter module to complete the sampling signal timing error compensation; the specific method is to convolve the filter coefficients with the signal samples in the time domain, the filter adopts a multi-phase structure design, and uses the different extraction and interpolation multiples of the multi-phase filter to achieve signal timing error compensation, and complete the data rate conversion at the same time.