Large-scale distributed ultrahigh-precision synchronization method and system

Abstract

The invention discloses a large-scale distributed ultrahigh-precision synchronization method and system, and the method is based on a main node system and a plurality of node subsystems, and comprises the steps: S1, enabling the node subsystems to obtain reference clock mark information Ck sent by the main node system in each local oscillation period, estimating uniform local frequency marking information Flu by adopting a weighted cyclic filtering algorithm; s2, based on reference clock mark information Ck in each local oscillation period and the unified local frequency mark information Flu, estimating unified local phase mark offset information # imgabs0 #; S3, further optimizing the unified local phase mark offset information # imgabs1 # so as to estimate local precise clock information Clu, and if the local precise clock information Clu is greater than the local precise clock information Clu, estimating the local precise clock information # imgabs0 #; therefore, the main node system controls the node subsystems to perform accurate synchronization. According to the large-scale distributed ultrahigh-precision synchronization method and system provided by the invention, subnanosecond synchronous acquisition can be carried out on node subsystems distributed at different geographic positions.

Description

Technical Field

[0001] The present invention relates to the field of computer network technology, and in particular to a large-scale distributed ultra-high precision synchronization method and system. Background Art

[0002] The application of large-scale distributed ultra-high precision transmission technology is very extensive. The most typical application scenarios include military linear array sonar wet-end distributed hydrophone transmission, ultra-large-scale multi-cable and multi-detector transmission in offshore oil exploration, and deep-sea thin cable sonar acquisition, transmission and detection systems.

[0003] The defects of traditional and existing distributed transmission systems are summarized as follows: First, traditional transmission technology often uses RS485 bus, ATM transmission, custom LVDS transmission, Ethernet plus RS485, Ethernet plus LVDS architecture. These architectures are difficult to implement in engineering applications, have backward performance, and poor reliability; second, the synchronization method is backward. The existing synchronization method mostly adopts an additional communication protocol and uses redundant cables to transmit a common system master clock node and synchronization pulse, or uses serial data clock recovery with command delay compensation to achieve small-scale synchronization acquisition. The synchronization accuracy deteriorates as the system scale increases. It does not have the scalability of large-scale systems, and it is relatively complex to implement. It requires additional software and hardware coordination, which has an adverse impact on system reliability. It is limited in the application of military sonar and seismic oil exploration fields with extremely strict reliability requirements; third, the existing technology does not have the data transmission capacity of ultra-large-scale systems, and the data throughput and transmission distance are limited; fourth, the existing technology business data, synchronization signals, clocks, and power supplies must be transmitted using separate cables. There are many cables in the entire system, the connection relationship is complex, and the process implementation is complex, which has an adverse impact on system reliability.

[0004] The technology described in this patent is a new technology proposed and implemented based on the previous technical accumulation of the patentee. The previous technology involves two invention patents, "CN201910140614.2-A cascade synchronous data acquisition and transmission system based on PTP protocol" and "CN201811089362.7-A sensor data acquisition and transmission system and method based on PTP synchronization". The above patents all use the PTP protocol for synchronization between distributed systems, and the performance reaches sub-microsecond level, which can meet the synchronization requirements of small-scale distributed systems. However, in the application scenarios of large-scale distributed ultra-high precision systems, the synchronization accuracy of the above technology will deteriorate. Therefore, there is an urgent need for a distributed method and system with high throughput and high synchronization accuracy that can meet large-scale expansion. Summary of the invention

[0005] In view of this, the present invention proposes a high-precision distributed synchronization method and system, which can perform sub-nanosecond synchronous acquisition of sensor data distributed in different geographical locations.

[0006] In order to achieve the above object, the technical solution adopted by the present invention is:

[0007] A large-scale distributed ultra-high precision synchronization method, characterized in that it is based on a master node system and multiple node subsystems, the master node system is used to control the node subsystems to synchronize, and receive and store data collected from the node subsystems, including:

[0008] S1: The node subsystem obtains the reference clock mark information C sent by the master node system in each local oscillation cycle k , using the weighted cyclic filtering algorithm to estimate the unified local frequency label information F lu ;

[0009] S2: Based on the reference clock mark information C in each local oscillator cycle k and the unified local frequency label information F lu , estimate the unified local phase mark offset information

[0010] S3: Mark the offset information of the unified local phase Further optimization is performed to estimate the local precise clock information C lu , so as to achieve synchronization of the master node system control node subsystem.

[0011] Furthermore, step S1 includes the following steps:

[0012] (1) The master node system sends synchronization information to the information receiving and processing module in the node subsystem in each clock, and obtains the reference clock mark information C in each local oscillation cycle through decoding. k ;

[0013] (2) The information receiving and processing module receives the reference clock mark information C in each local oscillation cycle. k Sending to the synchronization processing module in the node subsystem;

[0014] (3) In the synchronization processing module, the reference clock mark information C in each local oscillation cycle is used. k Get the reference frequency mark information F within each local oscillator cycle k , obtain local frequency mark information through the local oscillator in the node subsystem And adopt weighted cyclic filtering algorithm to measure the unified local frequency mark information F lu .

[0015] Furthermore, the step (3) uses a weighted cyclic filtering algorithm to measure the unified local frequency marker information F lu , the specific steps are as follows:

[0016] a. Set the estimated error to the difference between the reference frequency marker information in the nth local oscillator cycle and the local frequency marker information in the nth system local oscillator cycle: System error ε m is the median frequency error of the local oscillator in the node subsystem as indicated in the product manual; according to the weighted cyclic filtering algorithm, the cyclic weight can be obtained as

[0017] Therefore, the initial value of the estimated error is the difference between the reference frequency mark information in the first local oscillator cycle and the local frequency mark information in the first local oscillator cycle: The initial value of the cycle weight is

[0018] b. According to the weighted cyclic filtering algorithm, the local frequency tag information estimation value obtained within the nth local oscillator period is: And update the local frequency mark information in the n+1th local oscillator cycle to the estimated value of the local frequency mark information in the nth local oscillator cycle:

[0019] Therefore, the estimated value of the local frequency marker information in the second local oscillator cycle is: The local frequency mark information in the second local oscillator cycle is updated to

[0020] c. Repeat steps a and b until When the synchronization accuracy is less than the set threshold, the operation stops and the unified local frequency mark information F is obtained. lu =F m , where F m This is the reference frequency mark information when the operation is stopped.

[0021] Further, step S2 includes the following steps:

[0022] (1) In the synchronization processing module, the reference clock mark information C in each local oscillation cycle is used. k Obtain the reference time stamp information T in all the local oscillation cycles k ;

[0023] (2) The synchronization processing module records and receives all the reference clock mark information C in each local oscillation cycle. k The local time stamp of the moment

[0024] (3) Construct a two-dimensional matrix M k , the matrix has 2 rows and N columns, and the vector T k and Fill in the matrix M respectively k The first and second rows of k Split into A 2x2 matrix cluster, represented as:

[0025]

[0026] (4) Perform the same operation on each matrix in the matrix cluster: Get timestamp offset sequence The time stamp offset sequence is deducted from the singular value and the centroid value is taken to obtain the unified local time stamp offset information ΔT lu ;

[0027] (5) Combined with the unified local frequency label information F lu , through the formula Obtain the unified local phase mark offset information

[0028] Further, step S3 includes the following steps:

[0029] (1) The synchronization processing module uses the unified local frequency mark information F lu and the local phase mark offset information Generate local clock mark information Where, t is the local clock sequence;

[0030] (2) The reference clock mark information C within any local oscillation cycle i It can be expressed as in, is the reference phase mark offset information within the first local oscillator cycle;

[0031] (3) The synchronization processing module uses the unified local frequency mark information to generate reference clock mark information through the local oscillator of the node subsystem in Marking offset information for reference phase;

[0032] (4) respectively mark the reference clock information C r The local clock information and the reference clock mark information C in any local oscillation cycle 1 After frequency conversion and desharpening, we can obtain the following results:

[0033]

[0034] Therefore, the unified local phase mark offset information estimated in step S2 can be obtained by comparing the phases of the above two functions. The accuracy of the phase mark offset information is further optimized to obtain the phase mark error of the unified local phase mark offset information and the reference phase mark offset information.

[0035] (5) The synchronization processing module uses the unified local frequency mark information F lu , unify the local phase mark offset information and the phase mark error Obtain the local precise clock information C lu

[0036]

[0037] The present invention also provides a large-scale distributed ultra-high precision synchronization system, characterized in that it is based on a master node system and multiple node subsystems, the master node system is used to control the node subsystems to synchronize, and receive and store data collected from the node subsystems, including:

[0038] The master node system includes a master clock module and a data transmission module;

[0039] The node subsystem includes an information receiving and processing module, a synchronization processing module and a local clock module;

[0040] The information receiving and processing module is used for the node subsystem to obtain the command information and synchronization information sent by the master node system in each local oscillation cycle, and obtain the command by decoding. k ; and mark the reference clock information C k Sending to the synchronization processing module;

[0041] The synchronization processing module is used to mark the reference clock information C k Perform three-level synchronization processing to estimate the local precise clock information C lu , so as to achieve synchronization of the master node system control node subsystem.

[0042] The local clock module provides clock information for the node subsystem;

[0043] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the large-scale distributed ultra-high precision synchronization method as described above when executing the program.

[0044] The beneficial effects of the present invention are: the present invention can solve the problem of low synchronization accuracy in the current application scenarios of large-scale distributed ultra-high precision synchronization systems, and the specific beneficial effects are:

[0045] (1) The weighted cyclic filtering algorithm can improve the frequency synchronization accuracy of the acquisition and transmission nodes in different geographical locations in the distributed system;

[0046] (2) The cascade synchronization method is used to further improve the time phase synchronization accuracy, so that the final accuracy reaches the sub-nanosecond level.

[0047] (3) The transmission of synchronization information between the master node and all nodes shares the same channel with other data transmission, which reduces deployment difficulty and costs.

[0048] (4) This system can achieve sub-nanosecond synchronization and can guarantee the synchronization of subsystems in large-scale, long-distance distributed systems, and has great application potential. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0050] Figure 1 It is a flow chart of a large-scale distributed ultra-high precision synchronization method provided by the present invention;

[0051] Figure 2 It is a flowchart of step S1 of the large-scale distributed ultra-high precision synchronization method provided by the present invention;

[0052] Figure 3 It is a flow chart of the weighted cyclic algorithm in step S1 of the large-scale distributed ultra-high precision synchronization method provided by the present invention;

[0053] Figure 4 is a flow chart of step S2 of the large-scale distributed ultra-high precision synchronization method provided by the present invention;

[0054] Figure 5 is a flow chart of step S3 of the large-scale distributed ultra-high precision synchronization method provided by the present invention;

[0055] Figure 6 It is a schematic diagram of the distributed architecture of the master node system and multiple node subsystems of the large-scale distributed ultra-high precision synchronization method of the present invention;

[0056] Figure 7It is a schematic diagram of the large-scale distributed ultra-high precision synchronization system of the present invention. DETAILED DESCRIPTION

[0057] In order to better understand the present invention, the content of the present invention is further explained below in conjunction with the examples, but the content of the present invention is not limited to the following examples. Those skilled in the art can make various changes or modifications to the present invention, and these equivalent forms are also within the scope of the claims listed in the application.

[0058] Combine the following Figure 1-Figure 5 The present invention describes a large-scale distributed ultra-high precision synchronization method and system.

[0059] Figure 1 is a flow chart of a large-scale distributed ultra-high precision synchronization method provided by the present invention. In an embodiment of the present invention, based on a master node system and multiple node subsystems, such as Figure 6 As shown, the master node system is used to control the synchronization of the node subsystems and receive and store data collected from the node subsystems. The method includes:

[0060] S1: The node subsystem obtains the reference clock mark information C sent by the master node system in each local oscillation cycle k , using the weighted cyclic filtering algorithm to estimate the unified local frequency label information F lu ;

[0061] S2: Based on the reference clock mark information C in each local oscillator cycle k and the unified local frequency label information F lu , estimate the unified local phase mark offset information

[0062] S3: Mark the offset information of the unified local phase Further optimization is performed to estimate the local precise clock information C lu , so as to achieve synchronization of the master node system control node subsystem.

[0063] Specifically, the master node system described in the embodiment of the present invention is the master, and the node subsystem is the slave, each of which has a local oscillator clock. The system local oscillator refers to the local oscillator of the master node system, and the local local oscillator refers to the local oscillator of the node subsystem. The goal of the present invention is to synchronize the local oscillators of all node subsystems with the local oscillator of the master node system.

[0064] It can be understood that the unified local frequency and unified local phase described in the embodiment of the present invention may be any system of the multi-node subsystem.

[0065] The prior art usually uses the precise clock synchronization algorithm PTP to synchronize the time between the master node system and all node subsystems. The specific method is that the master node system sends synchronization information to all node subsystems, and the node subsystems compensate for the time deviation of the master and slave systems by calculating the timestamp. However, the estimation accuracy of the time frequency and time phase of this method is slightly insufficient, and cannot meet the synchronization accuracy requirements of large-scale distributed ultra-high precision systems.

[0066] In the embodiment of the present invention, by using a weighted cyclic filtering algorithm to perform multiple error estimates on the reference frequency sent by the master node system and the local frequency of the node subsystem, and simultaneously correcting the local frequency, the offset of the reference frequency and the local frequency can be greatly eliminated, and the synchronization of time and frequency information can be achieved. The cyclic weight of the weighted cyclic filtering algorithm can be set according to the synchronization convergence time requirement and the actual system scale.

[0067] In the embodiment of the present invention, the reference time mark information of all system local oscillation cycles of the master node and the local time mark information of the node subsystem receiving the reference time mark information are recorded, and group calculation is performed to obtain the master-slave node time error of each system local oscillation cycle. Due to the system software and hardware errors, it is necessary to remove singular points, and the weighted average algorithm or Kmeans clustering algorithm can be used to achieve singular point removal; further, the centroid of the remaining time error cluster is taken to obtain more accurate time error information; finally, the phase error information of the master and slave nodes is obtained through the phase calculation formula to achieve synchronization of time phase information.

[0068] In the embodiment of the present invention, third-party clock information is introduced as reference clock information, and the reference clock mark information is used as the master clock information to further optimize the above-mentioned synchronized phase information to obtain accurate time phase synchronization. The reference clock information can use the reference clock mark information of a certain local oscillation period, or can use the local local oscillator clock.

[0069] Figure 2 is a flow chart of step S1 of the distributed synchronization method provided by the present invention; Figure 2 shown.

[0070] (1) The master clock module of the master node system provides the clock information required for the operation of the master node system. In this embodiment, the master clock frequency can be 1 MHz, so each clock cycle is T p =1us.

[0071] The master node system is in each T pThe trigger rising edge sends synchronization information to all downstream node subsystems directly connected to the master node system through the network port using a 10 Gigabit Ethernet cable or a Gigabit Ethernet cable. The directly connected downstream node subsystem then sends the synchronization information to the serially connected downstream node subsystems in sequence until the last serially connected downstream node subsystem. In this embodiment, the master node system has 4 network ports and can be connected to 4 node subsystems; the synchronization information contains the clock information of the master node system; the synchronization information is encoded using a convolutional code to improve confidentiality and robustness.

[0072] After receiving the encoded synchronization information, the information receiving and processing module in the node subsystem can decode it using the Viterbi method to obtain the reference clock mark information C k .

[0073] (2) To the synchronization processing module in the node subsystem; in this embodiment, Gigabit / 100M Ethernet can be used for communication between the information receiving processing module and the synchronization processing module.

[0074] (3) The synchronization processing module is p The reference clock mark information C is received within k , which includes the reference frequency marker information F k , k is the index number of the local oscillator period.

[0075] The synchronous processing module is in each T p Get the local frequency mark information generated by the sub-clock module of the node subsystem In theory, the local clock generated by the node subsystem should be consistent with the reference clock generated by the master node system. However, in actual system operation, the corresponding clock frequency information has a certain deviation. The farther the distance between the master node system and the node subsystem or the more complex the routing, the greater the deviation, resulting in serious asynchrony.

[0076] The reference frequency mark information F provided by the master node system k As a benchmark, by sequentially p The reference frequency mark information and the local frequency mark information are subjected to a weighted cyclic filtering algorithm within a time interval, and an updated After that, the clock information of the sub-clock module of the node subsystem is corrected; after several iterations, until the error between the two is less than the set threshold, the weighted cyclic filtering process is stopped; the last calculated Meet the synchronization requirements and mark as unified local frequency mark information F lu .

[0077] Based on the content of the above embodiment, step S1 includes the following steps:

[0078] (1) The master node system sends synchronization information to the information receiving and processing module in the node subsystem in each clock, and obtains the reference clock mark information C in each local oscillation cycle through decoding. k ;

[0079] (2) The information receiving and processing module receives the reference clock mark information C in each local oscillation cycle. k Sending to the synchronization processing module in the node subsystem;

[0080] (3) In the synchronization processing module, the reference clock mark information C in each local oscillation cycle is used. k Get the reference frequency mark information F within each local oscillator cycle k , obtain local frequency mark information through the local oscillator in the node subsystem And adopt weighted cyclic filtering algorithm to measure the unified local frequency mark information F lu .

[0081] Figure 3 FIG. 1 is a flow chart of a weighted cyclic filtering algorithm of step S1 of the distributed synchronization method provided by the present invention. Figure 3 As shown, the step (3) uses a weighted cyclic filtering algorithm to measure the unified local frequency mark information F lu , the specific steps are as follows:

[0082] a. Set the estimated error to the difference between the reference frequency marker information in the nth local oscillator cycle and the local frequency marker information in the nth system local oscillator cycle: System error ε m is the median frequency error of the local oscillator in the node subsystem as indicated in the product manual; according to the weighted cyclic filtering algorithm, the cyclic weight can be obtained as

[0083] Therefore, the initial value of the estimated error is the difference between the reference frequency mark information in the first local oscillation cycle and the local frequency mark information in the first local oscillation cycle: The initial value of the cycle weight is

[0084] b. According to the weighted cyclic filtering algorithm, the local frequency tag information estimation value obtained within the nth local oscillator period is: And update the local frequency mark information in the n+1th local oscillator cycle to the estimated value of the local frequency mark information in the nth local oscillator cycle:

[0085] Therefore, the estimated value of the local frequency mark information in the second local oscillation cycle is: The local frequency mark information in the second local oscillation cycle is updated as follows:

[0086] c. Repeat steps a and b until When the synchronization accuracy is less than the set threshold, the operation stops and the unified local frequency mark information F is obtained. lu =F m , where F m This is the reference frequency mark information when the operation is stopped.

[0087] The synchronization accuracy threshold described in the embodiment of the present invention refers to a preset deviation threshold, and according to performance requirements, the value should not be higher than 100ppt.

[0088] The specific threshold is set in advance based on the convergence speed requirements and the number of node subsystems. When the required convergence speed is faster, the threshold is set larger; when the number of node subsystems is larger, the synchronization accuracy of the remote node subsystem is systematically reduced due to the increased routing complexity. Therefore, for the remote node system, it is necessary to consider both the guarantee of synchronization accuracy and the improvement of computing efficiency to determine the threshold value.

[0089] In this embodiment, the specific value of the cyclic weight can be determined in an adaptive manner according to the formula It can be obtained automatically or determined in advance based on the convergence speed requirements and the number of node subsystems. When the required convergence speed is faster, the loop weight is set smaller; when the number of node subsystems is more, the synchronization accuracy of the remote node subsystem is systematically reduced due to the increased routing complexity. Therefore, for the remote node system, it is necessary to consider both the guarantee of synchronization accuracy and the improvement of computing efficiency to determine the loop weight value.

[0090] Figure 4 is a flow chart of step S2 of the distributed synchronization method provided by the present invention. Figure 4 As shown, step S2 includes the following steps:

[0091] (1) In the synchronization processing module, the reference clock mark information C in each local oscillation cycle is used. k Obtain the reference time stamp information T in all the local oscillation cycles k ;

[0092] (2) The synchronization processing module records and receives all the reference clock mark information C in each local oscillation cycle. k The local time stamp of the moment

[0093] (3) Construct a two-dimensional matrix M k, the matrix has 2 rows and N columns, and the vector T k and Fill in the matrix M respectively k The first and second rows of k Split into A 2x2 matrix cluster, represented as:

[0094]

[0095] (4) Perform the same operation on each matrix in the matrix cluster: Get timestamp offset sequence The time stamp offset sequence is deducted from the singular value and the centroid value is taken to obtain the unified local time stamp offset information ΔT lu ;

[0096] (5) Combined with the unified local frequency label information F lu , through the formula Obtain the unified local phase mark offset information

[0097] In this embodiment, the time mark offset sequence may be constructed by not completely aligning all local oscillation periods T p The local time stamp information and the reference time stamp information are acquired, and sampling is performed only at intervals. The sampling rate may be considered to be 1:3 to improve processing efficiency.

[0098] Since the distance and route between each node subsystem and the main node system are different, and each T p The benchmark time stamp information and local time stamp information within the time interval are unbalanced, resulting in certain differences in the time stamp offsets and non-uniform distribution, which easily generates unclean data; therefore, it is necessary to remove the singular values ​​in the time stamp offset sequence. In this embodiment, a weighted average algorithm or a Kmeans clustering algorithm can be used to process the singular values ​​in the time stamp offset sequence; the weighted average algorithm is easy to operate and has a fast speed, but when the singular value error is large, the processing effect is not good; the Kmeans clustering algorithm needs to study the data in advance to determine the initial value, and the speed is slow, but it can remove the singular values ​​to a great extent and has a good processing effect.

[0099] In this embodiment, the centroid can be solved by the following formula:

[0100]

[0101] Among them, x i The time stamp offset ΔT is the i-th time stamp i , K is the number of time edit offset values ​​after removing singular values, a iis the corresponding weight value; in this embodiment, the weight value can be negatively correlated with the routing complexity, specifically, a i Can be used

[0102] Figure 5 is a schematic flow chart of step S3 provided by the present invention. Figure 5 As shown, step S3 includes the following steps:

[0103] (1) The synchronization processing module uses the unified local frequency mark information E lu and the local phase mark offset information Generate local clock mark information Where, t is the local clock sequence;

[0104] (2) The reference clock mark information C within any local oscillation cycle i It can be expressed as in, The reference phase mark offset information within any local oscillator cycle;

[0105] (3) The synchronization processing module uses the unified local frequency mark information to generate reference clock mark information through the local oscillator of the node subsystem in Marking offset information for reference phase;

[0106] (4) respectively mark the reference clock information C r The local clock information and the reference clock mark information C in any local oscillation cycle i After frequency conversion and desharpening, we can obtain the following results:

[0107]

[0108] Therefore, the unified local phase mark offset information estimated in step S2 can be obtained by comparing the phases of the above two functions. The accuracy is further optimized to obtain the phase mark error of the unified local phase mark offset information and the reference phase mark offset information.

[0109] (5) The synchronization processing module uses the unified local frequency mark information F lu , unify the local phase mark offset information and the phase mark error Obtain the local precise clock information C lu

[0110]

[0111] The synchronization module obtains the unified local frequency mark information F lu and the local phase mark offset information After that, the sub-clock module in the node subsystem generates updated clock information C l At this time, C l The synchronization accuracy of the reference clock information generated by the master node system can reach sub-millisecond level, which meets the synchronization requirements of small-scale distributed systems; however, the synchronization requirements of large-scale distributed ultra-high precision systems are sub-nanosecond level, so it is necessary to l Further optimization and correction are performed.

[0112] The reference clock mark information can be generated by the corrected sub-clock module of the node subsystem, or by the reference clock information generated by the master node system. The first method is adopted in this embodiment to reduce the delay and bandwidth overhead caused by information transmission.

[0113] In this embodiment, the reference clock marking information C r The local clock information and the reference clock mark information C in any local oscillation cycle i FPGA digital logic multiplication operation can be used for frequency conversion processing.

[0114] The desharpening process in this embodiment can be implemented by using a Butterworth digital low-pass filtering algorithm implemented by FPGA logic programming; the cutoff frequency can be 1 Hz.

[0115] In this embodiment, the comparison between F1 and F2 is used to obtain the phase mark error. The operation is as follows:

[0116] 1. Mix F1 and F2 to obtain the mixed signal:

[0117] 2. The time interval counting method can be used to indirectly measure the phase difference by calculating the time difference between the rising edges of ΔF.

[0118] The large-scale distributed ultra-high precision synchronization system provided by the present invention is described below. The large-scale distributed ultra-high precision synchronization system described below and the large-scale distributed ultra-high precision synchronization method described above can refer to each other.

[0119] Figure 7 is a schematic diagram of the structure of a large-scale distributed ultra-high precision synchronization system provided by the present invention. The distributed synchronization system provided by the present invention is based on a master node system and multiple node subsystems, such as Figure 6As shown, a distributed synchronization system is characterized in that it is based on a master node system and multiple node subsystems, the master node system is used to control the node subsystems to synchronize, and receive and store data collected from the node subsystems, including:

[0120] The master node system includes a master clock module and a data transmission module;

[0121] The node subsystem includes an information receiving and processing module, a synchronization processing module and a local clock module.

[0122] The information receiving and processing module is used for the node subsystem to obtain the command information and synchronization information sent by the master node system in each local oscillation cycle, and obtain the reference clock mark information C in each local oscillation cycle through decoding. k ; and mark the reference clock information C k Sending to the synchronization processing module;

[0123] The synchronization processing module is used to mark the reference clock information C k Perform three-level synchronization processing to estimate the local precise clock information C lu , so as to achieve synchronization of the master node system control node subsystem.

[0124] The distributed synchronization system described in this embodiment can be used to execute the above-mentioned distributed synchronization method embodiment. Its principles and technical effects are similar and will not be repeated here.

[0125] The basic principle of the large-scale distributed ultra-high precision synchronization system of the present invention is to place the system in a scene where distributed large-scale signal collection is required, such as air, underwater, chemical medium, etc. The master node system is directly connected to multiple node subsystems through a 10 Gigabit Ethernet cable or a Gigabit Ethernet cable at the network port, and the subnodes are connected in series with the downstream subnodes. There is no specific limit on the number of node subsystems, and the maximum value can be calculated based on their sampling frequency and system throughput. The master node system includes a master clock module to provide reference clock information for the entire system.

[0126] The data transmission module is used for the master node system to send synchronization information to the node subsystem. The module modulates and encodes the synchronization information using phase PSK and then sends it out. In this embodiment, the QPSK modulation method and the 1 / 2 convolution encoding method can be used.

[0127] The transmission method of synchronization information can adopt wavelength division multiplexing or time division multiplexing.

[0128] The node subsystem includes an information receiving and processing module and a synchronization module which communicate using a serial port. In this embodiment, an RS485 serial port can be used for connection.

[0129] After receiving the modulated and encoded synchronization information, the information receiving and processing module performs demodulation and decoding. In this embodiment, LLR soft demodulation and Viterbi decoding can be used respectively.

[0130] The local clock module is used to provide local clock information for the node subsystem.

[0131] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the large-scale distributed ultra-high precision synchronization method as described above when executing the program.

[0132] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0133] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0134] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A large-scale distributed ultra-high precision synchronization method, characterized in that: Based on a master node system and multiple node subsystems, the master node system is used to control the node subsystems for synchronization, and receive and store data collected from the node subsystems, including: S1: The node subsystem obtains the reference clock mark information C sent by the master node system in each local oscillation cycle k , using the weighted cyclic filtering algorithm to estimate the unified local frequency label information F lu ; S2: Based on the reference clock mark information C in each local oscillator cycle k and the unified local frequency label information F lu , estimate the unified local phase mark offset information S3: Mark the offset information of the unified local phase Further optimization is performed to estimate the local precise clock information C lu , so as to achieve precise synchronization of the master node system control node subsystem.

2. The large-scale distributed ultra-high precision synchronization method according to claim 1, characterized in that: Step S1 includes the following steps: (1) The master node system sends synchronization information to the information receiving and processing module in the node subsystem in each clock, and obtains the reference clock mark information C in each local oscillation cycle through decoding. k ; (2) The information receiving and processing module receives the reference clock mark information C in each local oscillation cycle. k Sending to the synchronization processing module in the node subsystem; (3) In the synchronization processing module, the reference clock mark information C in each local oscillation cycle is used. k Get the reference frequency mark information F within each local oscillator cycle k , obtain local frequency mark information through the local oscillator in the node subsystem And adopt weighted cyclic filtering algorithm to measure the unified local frequency mark information F lu .

3. The large-scale distributed ultra-high precision synchronization method according to claim 2, characterized in that: The step (3) uses a weighted cyclic filtering algorithm to measure the unified local frequency label information F lu , the specific steps are as follows: a. Set the estimated error to the difference between the reference frequency marker information in the nth local oscillator cycle and the local frequency marker information in the nth system local oscillator cycle: System error ε m is the median frequency error of the local oscillator in the node subsystem as indicated in the product manual; according to the weighted cyclic filtering algorithm, the cyclic weight can be obtained as Therefore, the initial value of the estimated error is the difference between the reference frequency mark information in the first local oscillator cycle and the local frequency mark information in the first local oscillator cycle: The initial value of the cycle weight is b. According to the weighted cyclic filtering algorithm, the estimated value of the local frequency mark information within the nth local oscillator period is obtained as follows: And update the local frequency mark information in the n+1th local oscillator cycle to the estimated value of the local frequency mark information in the nth local oscillator cycle: Therefore, the estimated value of the local frequency marker information in the second local oscillator cycle is: The local frequency mark information in the second local oscillator cycle is updated to c. Repeat steps a and b until When the synchronization accuracy is less than the set threshold, the operation stops and the unified local frequency mark information F is obtained. lu =F m , where F m This is the reference frequency mark information when the operation is stopped.

4. The large-scale distributed ultra-high precision synchronization method according to claim 3, characterized in that: Step S2 includes the following steps: (1) In the synchronization processing module, the reference clock mark information C in each local oscillation cycle is used. k Obtain the reference time stamp information T in all the local oscillation cycles k ; (2) The synchronization processing module records and receives all the reference clock mark information C in each local oscillation cycle. k The local time stamp of the moment (3) Construct a two-dimensional matrix M k , the matrix has 2 rows and N columns, and the vector T k and Fill in the matrix M respectively k The first and second rows of k Split into A 2x2 matrix cluster, represented as: (4) Perform the same operation on each matrix in the matrix cluster: Get timestamp offset sequence The time stamp offset sequence is deducted from the singular value and the centroid value is taken to obtain the unified local time stamp offset information ΔT lu ; (5) Combined with the unified local frequency label information F lu , through the formula Obtain the unified local phase mark offset information 5. The large-scale distributed ultra-high precision synchronization method according to claim 4, characterized in that: Step S3 includes the following steps: (1) The synchronization processing module uses the unified local frequency mark information F lu and the local phase mark offset information Generate local clock mark information Where, t is the local clock sequence; (2) The reference clock mark information C within any local oscillation cycle i It can be expressed as in, The reference phase mark offset information within any local oscillator cycle; (3) The synchronization processing module uses the unified local frequency mark information to generate reference clock mark information through the local oscillator of the node subsystem in Marking offset information for reference phase; (4) respectively mark the reference clock information C r The local clock information and the reference clock mark information C in any local oscillation cycle 1 After frequency conversion and desharpening, we can obtain the following results: Therefore, the unified local phase mark offset information estimated in step S2 can be obtained by comparing the phases of the above two functions. The accuracy of the phase mark offset information is further optimized to obtain the phase mark error of the unified local phase mark offset information and the reference phase mark offset information. (5) The synchronization processing module uses the unified local frequency mark information F lu , unify the local phase mark offset information and the phase mark error Obtain the local precise clock information C lu 6. A large-scale distributed ultra-high precision synchronization system, characterized in that: Based on a master node system and multiple node subsystems, the master node system is used to control the node subsystems for synchronization, and receive and store data collected from the node subsystems, including: The master node system includes a master clock module and a data transmission module; The node subsystem includes an information receiving and processing module, a synchronization processing module and a local clock module; The information receiving and processing module is used for the node subsystem to obtain the command information and synchronization information sent by the master node system in each local oscillation cycle, and obtain the command by decoding. The reference clock mark information C in each local oscillation cycle k ; and mark the reference clock information C k Sending to the synchronization processing module; The synchronization processing module is used to mark the reference clock information C k Perform three-level synchronization processing to estimate the local precise clock information C lu , so as to achieve synchronization of the master node system control node subsystem; The local clock module provides clock information for the node subsystem.

7. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the large-scale distributed ultra-high precision synchronization method as described in any one of claims 1 to 5 is implemented.

Images