A multi-channel synchronous acquisition method of electromagnetic environment data

Abstract

The application discloses a kind of multi-channel synchronous acquisition methods of electromagnetic environment data, it is related to data acquisition technical field, including being deployed in the same physical area with multiple acquisition channels, each channel is independently operated local clock;Pre-allocate a start time offset for each channel, so that each channel starts acquisition in turn according to the preset time sequence, and the difference between the start time offset of any two adjacent channels is greater than the maximum relative drift accumulation of the local clock of all channels within the expected longest continuous working time length;After each channel starts, continuous acquisition is independently carried out according to the local clock of each channel, and the collected data packet is sent to the center node in real time, the application finds the hidden law that is ignored, the order is unchanged and can be aligned, and the order can be guaranteed unchanged by a large enough start offset, so that the processing mode of traditional confrontation drift is changed.

Description

Technical Field

[0001] This invention belongs to the field of data acquisition technology, specifically a multi-channel synchronous acquisition method for electromagnetic environment data. Background Technology

[0002] In distributed electromagnetic environment data acquisition scenarios, multi-channel data synchronization is a crucial prerequisite for achieving subsequent high-precision signal processing such as time difference direction finding and coherent accumulation. It is generally accepted in this field that distributed acquisition synchronization must rely on precise time alignment, i.e., strict synchronization of clocks across all channels, or precise measurement and compensation of time delay differences between channels. Related technologies all revolve around time acquisition and calibration.

[0003] Existing solutions mainly fall into two categories. One relies on external synchronization signals to achieve clock unification, requiring dedicated hardware and high-precision clock devices, resulting in high deployment costs. In indoor, tunnel, and electromagnetically shielded environments, synchronization is easily interrupted due to external signal failure. The other uses a local clock plus drift compensation method, where each channel carries a timestamp, and the central node calculates and corrects drift in real time. This method involves a large amount of computation, and long-term drift accumulation will reduce synchronization accuracy. At the same time, the additional hardware increases the size and power consumption of the device, making it unsuitable for resource-constrained portable devices.

[0004] The aforementioned solutions all aim to eliminate clock drift, relying on external references, additional hardware, or real-time compensation. This technical approach has become the established understanding in the field. However, in engineering scenarios where there is no external time synchronization and hardware cost and power consumption are strictly limited, existing solutions have poor applicability. Furthermore, those skilled in the art have long been unaware that reliable alignment of multi-channel data can be achieved through other paths, rather than the traditional framework of time synchronization. They have also failed to consider using data order to replace time parameters, resulting in the long-term inability to meet the stable synchronization requirements of high-precision signal processing algorithms for input data. Summary of the Invention

[0005] The purpose of this invention is to provide a multi-channel synchronous acquisition method for electromagnetic environment data to solve the problems mentioned in the background art.

[0006] A method for multi-channel synchronous acquisition of electromagnetic environment data includes: Multiple acquisition channels are deployed in the same physical area, with each channel operating its local clock independently. A start time offset is pre-assigned to each channel so that each channel starts acquisition sequentially according to a preset time order, and the difference between the start time offsets of any two adjacent channels is greater than the maximum relative drift accumulation of the local clocks of all channels within the expected longest continuous working time. After each channel is started, it independently performs continuous data acquisition according to its local clock and sends the acquired data packets to the central node in real time. After receiving the data packets, the central node sorts them directly according to the order in which they were received, and outputs the sorted data as synchronized, multi-channel data. Since the difference in the start time offset is greater than the maximum relative drift accumulation, the receiving order of data packets in each channel is always consistent with the start time order, and no timestamp correction or synchronization calibration is required during the acquisition process.

[0007] By setting a preset start time offset and ensuring the difference is greater than the maximum relative drift accumulation, the receiving order and start order of data packets for each channel remain unchanged over a long period. The central node only needs to sort the data according to the receiving order to complete the multi-channel data alignment, without the need for any timestamp correction or synchronization calibration during the acquisition process.

[0008] In some possible implementations, the maximum relative drift accumulation is determined by the following steps: The frequency drift characteristics of the local clock of each acquisition channel within the expected operating temperature range are obtained, and the frequency drift characteristics include the monotonic correspondence between the drift direction and temperature change; Based on the principle that all channels within the same physical region experience the same temperature change trend, it is determined that the relative drift of any two channels shall not exceed the sum of their absolute drifts, and the sum of the absolute values ​​shall be taken as the upper limit of safety. The upper bound of the maximum relative drift accumulation is obtained by multiplying the sum of the largest and second largest absolute drift in each channel, or by taking twice the maximum absolute drift when all channels are of the same type, by the expected longest continuous operating time.

[0009] By acquiring the frequency drift characteristics of the local clock of each channel and taking advantage of the fact that each channel in the same physical area experiences the same temperature change trend, the safe upper limit of the relative drift is determined to be the sum of the maximum absolute drift and the second largest absolute drift (or twice the maximum absolute drift when all channels are of the same type), thereby providing reliable redundancy in the case of unknown drift direction.

[0010] In some possible implementations, a step of performing a drift over-limit self-check using the receiving order is also included: The central node records the acquisition channel identifier corresponding to each received data packet; Compare the currently received channel identifier sequence with the preset startup sequence bit by bit; When more than M sequence misalignments occur in N consecutive comparisons, the system is determined to have exceeded the preset range of maximum relative drift accumulation. Trigger an automatic reinitialization process to reallocate the startup time offsets for each channel.

[0011] This system utilizes the receiving sequence for drift over-limit self-checking. It compares the actual received channel identifier sequence bit by bit with a preset startup sequence, and triggers automatic re-initialization when a sequence misalignment exceeds a set number of consecutive comparisons. This solution can monitor drift beyond safe limits in real time without additional sensors, enabling system self-healing and ensuring long-term operational reliability.

[0012] In some possible implementations, the data packets of each acquisition channel contain the start time offset encoding of that channel, and the central node performs implicit synchronization verification through the following steps: Extract the start time offset code of each channel from the received data packets; Based on the size relationship of the encoded values, a desired receiving order is generated; The actual receiving order is compared with the expected receiving order. If they match, the current order is confirmed to be valid; otherwise, the current data packet is discarded and the system waits for the next round of data packets.

[0013] Each channel's data packets contain a start time offset encoding. The central node generates the expected reception order based on the relationship between the encoded values ​​and compares it with the actual reception order to complete implicit synchronization verification. This verification process does not rely on timestamps or external references; it only needs to compare the order to confirm the validity of the sequence, thus improving the system's robustness in complex communication environments.

[0014] In some possible implementations, the difference in startup time offsets is greater than a set maximum relative drift accumulation, specifically including the following steps: Pre-measure the extreme frequency drift rate of the local clock of each acquisition channel at the extreme operating temperature; Multiply the extreme frequency drift rate by the expected longest continuous operating time, and then multiply by a temperature change rate correction factor to obtain the extreme cumulative drift of a single channel. Add the extreme cumulative drift values ​​of any two channels to obtain the upper bound of the maximum relative cumulative drift value; Set the difference in startup time offsets to an integer multiple of this upper bound, where the multiple is greater than or equal to 2.

[0015] In some possible implementations, the following steps are also included to eliminate the interference of communication link out-of-order delivery on the receiving order: When each acquisition channel sends a data packet, it embeds the local acquisition round counter value and the sending sequence number within the same round into the data packet; After receiving the data packet, the central node first groups it according to the collection round counter value, and then sorts it according to the sending sequence number within the same round to restore the correct collection order. Data packets from different rounds are not sorted across rounds to avoid cross-round order errors caused by fluctuations in communication latency.

[0016] In some possible implementations, after the central nodes are sorted according to the receiving order, time coordinates that can be used by subsequent algorithms are generated through the following steps: The first arriving data packet is taken as the zero point of time; Based on the preset nominal sampling interval, virtual sampling time is allocated sequentially and incrementally to the data packets arranged in the same sampling round. The virtual acquisition time is written into the header of the output data stream as a standardized timestamp for the data packet; The virtual acquisition time allocation process does not read the actual local clock value of any acquisition channel.

[0017] In some possible implementations, the following steps are also included to address situations where the number of channels exceeds the concurrent capacity of the communication link: Multiple acquisition channels are grouped according to the order of start time offset, and the number of channels in each group does not exceed the maximum number of conflict-free concurrency of the communication link; Each channel within the same group is started sequentially according to the rule that the difference in the start time offset is greater than the maximum relative drift accumulation. A protection interval greater than the maximum cumulative drift within a group is set between different groups to prevent the order of groups from being interleaved. The maximum number of conflict-free concurrent connections can be determined by the inflection point of the packet loss rate of the measured communication link under saturation, or by taking the recommended number of concurrent connections of a commonly used wireless protocol (such as LoRa).

[0018] In some possible implementations, the following steps are also included to dynamically adjust the expected maximum continuous operating time: The central node records the timestamp of the data packets received by each acquisition channel and calculates the deviation between the time interval of receiving two adjacent data packets and the nominal sampling interval. Based on the cumulative trend of the aforementioned deviation, the relative drift rate of the local clock of each channel is estimated in real time. When the estimated relative drift rate is lower than a preset threshold, the expected maximum continuous working time is extended; when the estimated relative drift rate is higher than the preset threshold, the expected maximum continuous working time is shortened and a reconfiguration warning is issued; wherein the preset threshold is determined based on the ratio of the upper bound of the maximum relative drift accumulation to the initial expected maximum continuous working time. Within the shortened working time, the cumulative drift amount is recalculated based on the currently estimated relative drift rate. If the cumulative drift amount is still less than the difference between the start time offset and the current drift amount, the operation continues; otherwise, a re-initialization is triggered immediately.

[0019] The adjustment of the expected longest continuous working time does not change the start time offset of each channel, but is only used to update the validity judgment of the sequential invariance in subsequent working periods.

[0020] In some possible implementations, the following steps are also included to leverage order invariance for lossy compressed transmission of multichannel data: The central node buffers data packets from all channels received in the same collection round, and concatenates the data from each channel into a super-long data frame in the order of reception, wherein the reception order is consistent with the start time order of each channel; The ultra-long data frame does not contain any channel identifier or timestamp information, and only implicitly distinguishes the data of different channels by a fixed order position; The spliced ​​ultra-long data frame is used as the output of this acquisition round for subsequent processing.

[0021] The technical solutions provided by the embodiments of this disclosure have at least the following beneficial effects: This invention pre-allocates a start-up time offset to each acquisition channel, ensuring that the difference in offsets between adjacent channels exceeds the maximum relative drift accumulation of all channels' local clocks within the expected longest continuous operating time. This guarantees that the data packet reception order always matches the start-up order. Based on this, the central node only needs to sort the data according to the reception order to achieve synchronous alignment of multi-channel data, without any external synchronization signals, additional sensors, or real-time timestamp correction or synchronization calibration. This method overcomes the technical bias of traditional distributed acquisition relying on time synchronization, simplifying the complex synchronization problem into a sequence guarantee problem. It reduces hardware costs and computing power overhead, improves adaptability in scenarios without external time synchronization and with limited resources, and the aligned data can be directly used for high-precision signal processing algorithms such as time difference direction finding and coherent accumulation. Attached Figure Description

[0022] Figure 1 This is a schematic diagram of the method structure of the present invention. Detailed Implementation

[0023] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0024] Please see Figure 1 This application provides a multi-channel synchronous acquisition method for electromagnetic environment data, including: Multiple acquisition channels are deployed in the same physical area, and each channel runs its local clock independently.

[0025] The same physical area refers to a region where all data acquisition channels are deployed within a horizontal distance of no more than 50 meters and a vertical height difference of no more than 5 meters, and where there are no significant heat or cold sources, such as air conditioning vents, heat dissipation equipment, or direct sunlight, that could cause local temperature differences exceeding ±2℃. If the actual deployment environment cannot meet the above conditions, monitoring should be carried out by adding thermal barriers between channels or deploying temperature sensors simultaneously to ensure that the temperature change trends experienced by each channel are basically consistent.

[0026] This step is used to build a distributed data acquisition hardware architecture, providing basic support for subsequent orderly data acquisition, without requiring additional sensors or external synchronization signals. The acquisition channels use electromagnetic sensor modules with independent acquisition and transmission capabilities, requiring no additional synchronization interfaces and only a standard local clock oscillator. During deployment, ensure that all channels cover the monitoring area and that the installation height and orientation are consistent to avoid signal deviations caused by location differences.

[0027] Each channel's local clock does not require pre-calibration and is allowed to have initial deviations. An industrial-grade clock with a frequency stability of ±50ppm is sufficient, eliminating the need for high-precision atomic clocks. This helps reduce hardware procurement and deployment costs and adapts to the engineering requirements of portable, low-power acquisition devices. This step simplifies the hardware configuration and deployment process, improves the system's adaptability in scenarios without external synchronization signals, avoids interference risks from a global clock source, and ensures operational stability in complex environments.

[0028] The difference in startup time offset is greater than the maximum relative drift accumulation setting, specifically including the following steps: The extreme frequency drift rate of the local clock on each acquisition channel under extreme operating temperatures is measured in advance. It should be noted that this step aims to obtain the upper limit data of clock offset under extreme conditions, providing a reliable boundary for subsequent drift calculations and serving as a prerequisite for reverse utilization of drift constraints. Conventional techniques in this field mostly offset drift effects through real-time compensation, rarely actively quantifying the maximum cumulative drift rate, and even less often using this quantization result to construct sequential safety boundaries, which is significantly different from traditional drift-countermeasures.

[0029] The key understanding of this invention is that although the local clock of each channel has drift, the drift has a slow accumulation characteristic. Within a limited working period, the relative drift amount has a definite upper bound. Although this law objectively exists, it has not been used to solve the synchronization problem for a long time. However, this invention can control the drift effect without relying on a real-time temperature compensation circuit.

[0030] The extreme operating temperature range covers the extreme temperatures that may occur in practical applications, with a low temperature limit of -40℃ and a high temperature limit of 85℃. The specific measurement method is as follows: Measurement equipment: Use a frequency meter (accuracy not less than ±0.1ppm) to measure the channel clock output frequency.

[0031] Sample size: At least 5 acquisition channels of the same model are randomly selected as samples for each temperature point.

[0032] Test procedure: Place the sample channel in a constant temperature chamber and stabilize it at three temperature points: -40℃, 25℃, and 85℃. After holding at each temperature point for 2 hours, start the test and continuously collect the output frequency of the channel clock for 12 hours. Record the frequency value every hour and compare it with the nominal frequency to obtain the frequency drift value (unit: ppm) for that hour.

[0033] Data processing: For each temperature point, the maximum hourly frequency drift value within 12 hours is taken as the extreme drift rate of that channel at that temperature point; then, the average extreme drift rate of all sample channels at the same temperature point is taken as the average extreme drift rate of that temperature point; finally, the maximum average extreme drift rate among the three temperature points is taken as the final extreme frequency drift rate.

[0034] Error range: Repeat the above complete measurement process 3 times, with an interval of 24 hours between each time, and calculate the relative standard deviation (RSD) of the 3 measurement results. If RSD ≤ 5%, take the average of the 3 measurements as the final result; if RSD > 5%, increase the sample size to 10 and remeasure.

[0035] Recording requirements: All raw measurement data should be recorded in the test report for future reference.

[0036] This step locks the drift upper limit by measuring extreme values, avoiding safety hazards caused by insufficient estimation, providing real data support for redundant design, ensuring the adaptability of the solution in a wide temperature range, and enabling the initial offset to stably cover the maximum drift accumulation.

[0037] Multiply the extreme frequency drift rate by the expected longest continuous operating time, and then multiply by the temperature change rate correction factor to obtain the extreme cumulative drift of a single channel.

[0038] Furthermore, this step converts the extreme drift per unit time into the maximum possible offset within the entire working cycle, providing a quantitative basis for setting the starting offset.

[0039] Specifically, the expected maximum continuous working time is preset according to task requirements, such as 24 hours, or 86,400 seconds, to adapt to engineering scenarios involving long-term uninterrupted data collection. The temperature change rate correction factor is used to compensate for the difference between dynamic temperature changes in actual engineering scenarios and static test conditions in the laboratory. In this embodiment, an empirical value of 1.2 is used, which is applicable to conventional outdoor environments (daily temperature difference ≤ 20℃, temperature change rate ≤ 10℃ / h).

[0040] If the application scenario involves more drastic temperature changes (such as deserts or polar regions), the calibration should be recalibrated according to the actual environment: continuously monitor the temperature change curve at the target site for no less than 7 days, measure the clock drift rate, calculate the statistical ratio of the drift rate to the temperature change rate, and take 1.1 times this ratio as the correction factor.

[0041] The specific calculation method is as follows: Single-channel extreme cumulative drift = extreme frequency drift rate × expected longest continuous working time × temperature change rate correction factor. In this embodiment, the extreme frequency drift rate is ±50ppm / h, the expected longest continuous working time is 24h, and the temperature change rate correction factor is 1.2. The calculated single-channel extreme cumulative drift is 5.184 seconds. This step makes the calculation results closer to the actual dynamic working conditions, ensuring coverage of more severe drift scenarios, providing a quantitative basis for sufficiently large misalignment amounts, and ensuring the core objective of maintaining the sequence unchanged during long-term operation.

[0042] The upper bound of the maximum relative cumulative drift is obtained by summing the extreme cumulative drift amounts of any two channels. It should be understood that the core focus of the system is the relative drift between channels, not the absolute drift; this is the essential difference between this invention and traditional solutions. Traditional solutions attempt to eliminate all drift, requiring complex compensation algorithms and hardware support, while this invention only needs to control relative drift to prevent data order disruption. This shift from absolute control to relative constraint breaks the conventional understanding of drift handling in the field.

[0043] Within the same area, the environments of each channel are similar. The maximum relative drift accumulation is the sum of the extreme cumulative drift of any two channels. The specific calculation method is as follows: select the largest extreme cumulative drift among all channels and add it to the second largest extreme cumulative drift to obtain the upper bound of the maximum relative drift accumulation. If all channels are of the same model and from the same batch, then take twice the extreme cumulative drift of a single channel as the upper limit of the maximum relative cumulative drift.

[0044] The reason for using 2 times instead of a smaller factor (such as √2 times) is that the frequency drift directions of different individual crystal oscillators of the same model may be the same or opposite, and the drift direction may reverse with temperature changes. To ensure that the order is not disrupted under any combination of drift directions, the worst-case scenario must be considered—the drift amounts of the two channels are superimposed in the same direction, that is, the relative drift amount is the sum of their absolute values. Therefore, taking 2 times the extreme cumulative drift amount of a single channel as the safety upper bound is the minimum redundancy design that satisfies "absolute safety". If it is possible to ensure that the drift direction of all channels is consistent through pre-screening in engineering (e.g., all drifting in the positive direction), a smaller factor can be used, but this method is a general solution, so a conservative value of 2 times is used.

[0045] In this embodiment, all channels are of the same model, and the extreme cumulative drift of a single channel is 5.184 seconds. Therefore, the upper limit of the maximum relative cumulative drift is 10.368 seconds. This step determines the safety boundary by superposition, ensuring that subsequent offset settings can cover all channel combinations and adapt to different engineering requirements from 2 channels to multiple channels. From a quantitative perspective, it ensures that the drift order cannot be reversed.

[0046] Set the difference in startup time offsets to an integer multiple of this upper bound, where the multiple is greater than or equal to 2.

[0047] In one implementation, to ensure that the data order remains stable, the starting offset difference needs to retain sufficient redundancy. This is because conventional technical approaches tend to synchronize the clocks, reduce drift, or compensate for drift. This is a closed-loop, real-time design that depends on external factors and is easily constrained by the environment and hardware. The present invention makes the clock misalignment large enough and adopts an open-loop, front-end, and autonomous design approach. Its core logic is that it does not directly solve the time synchronization problem, but makes the time synchronization problem irrelevant by keeping the order unchanged.

[0048] Specifically, the value should be a multiple of 2 or higher; in this embodiment, it is 2. The specific calculation method is as follows: The difference in start-up time offset = upper limit of maximum relative drift accumulation × multiple, i.e., 10.368 seconds × 2 ≈ 21 seconds; The selection principle of the multiplier is: set according to the system reliability requirements, the higher the requirements, the larger the multiplier, and the minimum is not less than 2, so as to ensure that even if extreme drift superposition occurs, the order will not be destroyed; The allocation method for the start time offset of adjacent channels is as follows: they are allocated sequentially according to the channel number. The start time of the first channel is 0 seconds, the start time of the second channel is 21 seconds, the start time of the third channel is 42 seconds, and so on, so that the offset difference between adjacent channels remains consistent. The configuration is simple and easy to operate.

[0049] This step constructs a safety boundary that drift cannot cross through redundant design, reducing the possibility of drift causing order changes and providing an underlying guarantee for the central node to achieve synchronization based on sequence. Synchronization effectiveness can be guaranteed without complex calibration algorithms. Because the offset is greater than the maximum relative drift, the order of data packets in each channel remains stable regardless of subsequent clock drift. This is a crucial prerequisite for achieving synchronization without time alignment, relying solely on sequential sorting, ensuring that the aligned data can be directly used in high-precision signal processing algorithms.

[0050] The difference in start-up time offset between adjacent channels is set to 21 seconds, which is greater than the upper limit of the maximum relative drift accumulation. The start-up time offset allocation can be flexibly adjusted as follows: If the number of channels changes or the environmental drift characteristics change in actual application, the extreme frequency drift rate can be remeasured, the upper limit of the maximum relative drift accumulation can be updated according to the above calculation steps, and the start time offset can be redistributed according to the set multiple. As long as the adjacent difference meets the constraint condition of "greater than the upper limit of the maximum relative drift accumulation", sufficient offset redundancy can be reserved to ensure that the acquisition window does not overlap due to drift throughout the entire cycle, adapt to the engineering requirements of different acquisition cycles, and further consolidate the core foundation of unchanged sequence.

[0051] The system also includes the following steps to eliminate the interference of out-of-order communication links on the receiving order: When sending data packets, each acquisition channel embeds the local acquisition round counter value and the sending sequence number within the same round into the data packet.

[0052] Considering the susceptibility of wireless links to out-of-order transmission, a common problem in wireless transmission scenarios, and the core foundation of this invention being the preservation of order—meaning the relative order of data packets in each channel matches the startup order—link out-of-order transmission undermines this core premise. Traditional methods for handling out-of-order transmission often employ timestamp sorting or retransmission mechanisms. This solution, however, restores the correct order using only simple rounds and sequence identifiers, making it more concise and efficient.

[0053] This step involves attaching identification information to each data packet that can be used to restore the order. The specific implementation method is as follows: Each acquisition channel has a built-in 8-bit round counter and a 16-bit sequence counter. The round counter is initially set to 0. After each set of data packets is acquired (the preset number is 1024 sets, which can be adjusted according to the transmission bandwidth), the round counter is automatically incremented by 1. After overflowing, it starts counting again from 0. The sequence counter is initially set to 0. After each data packet is acquired, the sequence counter is automatically incremented by 1. After each round, it is reset to 0. Within the same round, each channel uses its own channel number as a preset fixed number, which is combined with the sequence counter value to form a transmission sequence number. The format is channel number (8 bits) + sequence counter value (16 bits). The round counter value and transmission sequence number are embedded in the reserved field of the data packet. The embedding position is the first 4 bytes of the data packet header, with the first byte being the round counter value and the last 3 bytes being the transmission sequence number, requiring no additional communication overhead. This step, by adding a dual identifier of round and sequence to the data packet, enables the central node to restore the correct order without relying on the clock and arrival time, solving the synchronization failure problem caused by wireless link out-of-order delivery, and ensuring that the core logic with unchanged order is not disrupted by link interference.

[0054] After receiving the data packet, the central node first groups it according to the collection round counter value, and then sorts it according to the sending sequence number within the same round to restore the correct collection order.

[0055] Furthermore, the central node uses the aforementioned two fields to perform out-of-order rearrangement, which is the core manifestation of the central node's handling of degradation in this invention. Traditional synchronization schemes often require the central node to execute complex time alignment algorithms, while this scheme simplifies the synchronization process to a sorting operation based on a preset order, reducing processing complexity and computational overhead. Traditional schemes require complex time alignment algorithms, while this invention only requires simple sorting to achieve synchronization, greatly simplifying the processing logic. Moreover, the sorting is based on a preset sending sequence number, rather than the data packet reception time or timestamp, completely eliminating dependence on time parameters and avoiding synchronization deviations caused by timestamp errors.

[0056] The specific implementation method is as follows: After receiving the data packet, the central node first parses the first 4 bytes of the data packet header and extracts the round counter value and the sending sequence number; Create a round-based cache queue, with each round corresponding to an independent cache area. Store data packets into the corresponding cache area according to the round counter value. If the data for that round already exists in the cache area, append it directly; otherwise, create a new cache area. When the number of data packets received in the buffer in a certain round reaches the preset number (consistent with the preset number of the acquisition channel, which is 1024 groups), the data packets in the buffer are sorted according to the following rules: First, sort the channels in ascending order according to the transmission sequence number. Then, sort the channels within the same channel in ascending order according to the sequence counter value to obtain a standard sequence that is consistent with the start order of each channel. After sorting, the data for that round is output to the subsequent processing module, and the buffer space for that round is released to avoid consuming too much storage resources. Each time the central node receives a data packet, it performs the above parsing and grouping operations. The entire sorting process relies solely on the round number and sequence number, without considering the actual arrival time of the data packets. The calculation logic is simple, consumes minimal computing resources, and is suitable for engineering needs of low-performance central nodes.

[0057] This step reduces the impact of out-of-order communication before entering the synchronization logic, ensuring that subsequent synchronization verification and time coordinate generation use the correct order. This further consolidates the unchanged order and avoids link out-of-order disruption of the inherent order established by the start offset, ensuring synchronization stability in complex communication environments and allowing the sorted data stream to be directly input into high-precision signal processing algorithms.

[0058] Data packets from different rounds are not sorted across rounds to avoid cross-round order errors caused by communication latency fluctuations. It should be understood that cross-round aliasing can disrupt timing integrity, thus affecting core logic that maintains its order. In long-term, multi-round data collection scenarios, this can easily lead to data corruption; therefore, strict isolation is necessary. The specific isolation method is as follows: The central node sets an independent timestamp threshold for each round of buffering. The threshold is the time of receiving the first data packet in that round plus a preset timeout (the preset timeout is 1 second and can be adjusted according to the link latency characteristics). If the buffer in a certain round fails to collect the preset number of data packets within the timeout period, it is judged as packet loss. Only the received data packets are sorted according to the above rules and output, and a packet loss log is recorded at the same time. When processing data from a certain round, the central node does not mix and sort data packets from other rounds. Even if data packets from subsequent rounds arrive early, they are first stored in the corresponding round's buffer. Subsequent rounds are processed only after the current round is completed, ensuring the independence of data from each round.

[0059] This step avoids data overlap across cycles through rule design, ensures the temporal independence and integrity of data in each round, adapts to engineering requirements for continuous multi-round acquisition, improves the reliability of data processing, and ensures that the sorting results of each round can be directly used for high-precision signal processing.

[0060] After each acquisition channel is started, it independently and continuously acquires data according to its local clock and sends the data packets to the central node in real time. Acquisition automatically resumes when a channel reaches a preset offset. The specific startup method is as follows: Each channel has a built-in start timer with an initial value of the corresponding assigned start time offset (e.g., 0 seconds for channel 1, 21 seconds for channel 2, etc.). The timer starts counting down after the system is powered on, and the acquisition module is triggered to start acquisition after the countdown ends. The sampling rate is set to 100Hz, which is 100 data sets per second, to meet the rate requirements of conventional electromagnetic data acquisition. The data packet contains the channel identifier, acquisition sequence number, start offset code, acquisition round counter value, transmission sequence number and raw data, but does not contain complex timestamps. The channel identifier here is only used for initial configuration and verification, and is not used for data alignment. This is fundamentally different from the traditional design that relies on the matching and alignment of the identifier and timestamp. It simplifies the data packet structure, reduces the transmission bandwidth usage, and eliminates the need for additional processing related to timestamps. The transmission method can be wireless or wired, such as LoRa or 5G, and can be selected according to the actual distance and real-time requirements to adapt to the engineering needs of different transmission scenarios.

[0061] Since the starting offset difference is greater than the maximum relative drift accumulation, slight jitter during transmission will not affect the sorting logic, ensuring synchronization stability in complex transmission environments.

[0062] This step simplifies the data packet structure, reduces hardware power consumption and bandwidth usage, and ensures continuous and stable transmission.

[0063] After receiving data packets, the central node first groups them by collection round, and then sorts them by transmission sequence number within the same round. The restored order is used as the multi-channel data order for synchronization and alignment. The central node's multi-channel receiving module first performs out-of-order rearrangement, and then uses the corrected order for subsequent processing. Because the starting offset difference is sufficient, the original sending order of data packets can be kept consistent with the starting order. After rearrangement, the correct timing can be restored, eliminating the dependence on traditional timestamp calibration. It only relies on reliable order to support subsequent synchronization logic, reducing the computing power consumption of the central node. This is the core implementation path of the present invention that can synchronize without time alignment. It is suitable for the engineering needs of low computing power central nodes. Moreover, the time relationship of the aligned data stream is stable and can be directly input into high-precision signal processing algorithms such as time difference direction finding and coherent accumulation.

[0064] The central node implements implicit synchronization verification through the following steps: Extract the start time offset code for each channel from the received data packet. In one implementation of this step, the offset code uses a fixed field format, mapping one-to-one with the offset. The specific implementation is as follows: The startup time offset of each channel is written to the channel's local storage module during system initialization. The encoding rule is to convert the startup time offset (unit: seconds) into a 32-bit binary number as the startup time offset encoding. A 4-byte field is reserved in the data packet to store this encoding, located in bytes 5-8 of the data packet header. After receiving the data packet, the central node parses bytes 5-8 of the header, converts the 32-bit binary number into a decimal value, and obtains the start time offset encoding of the channel. No additional parsing or calculation is required, making the operation simple and efficient and reducing the processing pressure on the central node.

[0065] This step ensures that the code extraction process is simple and reliable, reducing additional errors. The code here is only used to verify whether the order is consistent with the startup order, rather than to calculate the time difference. This reflects the design philosophy of prioritizing order over time, avoiding the accumulation of errors caused by time calculations, and ensuring that the verification logic does not deviate from the core principle of maintaining the order.

[0066] Based on the size relationship of the encoded values, a desired receiving order is generated. The desired order is consistent with the ascending order of the start time offset. Specifically, the generation method is as follows: when the system is initialized, the central node collects the start time offset codes of all channels, sorts them by the encoded values ​​from smallest to largest, and obtains the arrangement order of the channel numbers. This arrangement order is the desired receiving order and is stored in the central node's configuration file. For example, if the encoding value for channel 1 is 0 (corresponding to 0 seconds), the encoding value for channel 2 is 21 (corresponding to 21 seconds), and the encoding value for channel 3 is 42 (corresponding to 42 seconds), then the expected reception order is channel 1, channel 2, and channel 3. If the number of channels or the start time offset changes during system operation, the central node can reread the encoding values ​​of each channel and update the expected reception order according to the above rules to ensure the uniqueness and timeliness of the expected order. The smaller the encoding, the earlier the channel starts, and the earlier its expected position. The central node sorts the encoded data in ascending order to directly obtain the desired reception order. This desired order is an inherent order determined by a preset start offset, rather than a dynamic order calculated based on timestamps. The calculation logic is simple, with no accumulated errors, and the uniqueness of the desired order is guaranteed by pre-design, requiring no real-time adjustments. This step directly maps the order through the code size, with uniquely determined rules, requiring no complex algorithms. It adapts to the engineering needs of low-performance central nodes, ensuring a perfect match with the core requirement that the desired order aligns with the start order.

[0067] The actual receiving order is compared with the expected receiving order. If they match, the current order is confirmed to be valid; otherwise, the current data packet is discarded and the system waits for the next round of data packets.

[0068] Using the above method, the actual arrival order is compared with the expected order bit by bit. The specific comparison method is as follows: the central node extracts the channel number of each data packet after sorting in the same round according to the output order to form the actual receiving order sequence. The expected receiving sequence is read from the configuration file. The two sequences are compared one by one according to their positions. If the channel numbers at all positions are completely consistent, the current sorting is determined to be valid, and the data of this round is output to the subsequent modules. If the channel numbers at any position are inconsistent, the sorting is determined to be abnormal. The abnormal handling method is to discard all data packets of this round, start the re-reception mechanism, wait for the next round of data packet collection and transmission, and record the abnormal log for later investigation. If sorting anomalies occur for three consecutive rounds, an alarm signal is triggered, prompting engineers to check the link or channel status. This verification process is essentially a direct check of the unchanged order, rather than a check of time synchronization accuracy, which is one of the key features distinguishing this invention from traditional synchronization schemes. Traditional synchronization verification often achieves this by detecting time synchronization accuracy; this scheme only needs to compare the order to determine synchronization validity, eliminating the need for complex accuracy detection algorithms. Furthermore, the data stream order is stable after successful verification, directly meeting the input requirements of high-precision signal processing algorithms. This step can complete implicit synchronization verification without timestamps or calibration, improving system reliability in complex environments and adapting to the engineering needs of harsh operating conditions.

[0069] The system also includes the following steps to achieve lossy compressed transmission of multi-channel data using order invariance: The central node buffers data packets from all channels received in the same acquisition round, and concatenates the data from each channel into a super-long data frame in the order of reception, with the reception order being consistent with the start time order of each channel.

[0070] In this embodiment, since the core logic of maintaining the data order has been verified, the order can be directly used to replace the channel identifier, which is a compression logic that is difficult to achieve in traditional solutions. Conventional data compression often focuses on optimizing encoding algorithms, while this solution utilizes the inherent stability of the data order to remove redundant information, resulting in a significant difference in compression approach.

[0071] The specific splicing method is as follows: For data packets that have passed the same round of verification, the central node extracts the original data segments of each data packet in the expected receiving order (removing the round counter value, sending sequence number, start offset code and other identification fields in the header). The length of the raw data segment for each channel is fixed (32 bytes in this embodiment, which can be adjusted according to the sampling precision). The data segments are concatenated sequentially according to the channel order, i.e., raw data segment of channel 1 + raw data segment of channel 2 + ... + raw data segment of channel N, forming an ultra-long data frame. The header of the ultra-long data frame only retains the round identifier (1 byte), which is used to confirm the round during backend parsing. There are no other redundant fields. The concatenation process is implemented through memory copying, which does not require complex encoding operations and has high processing efficiency.

[0072] After completing the sorting and synchronization verification for the same round, the central node caches all valid data packets for that round. Based on the restored receiving order, it sequentially splices the acquired data segments from each channel end-to-end to form a single ultra-long data frame, reducing frame header overhead and transmission interaction counts, thus improving transmission efficiency. The splicing order is consistent with the physical startup order of the channels, allowing subsequent modules to directly restore channel ownership by location without additional identifier resolution, reducing backend processing complexity. Furthermore, the restored data stream order is completely consistent with the synchronization alignment result, without affecting the high-precision signal processing effect. This step can merge multiple fragmented data frames into a single continuous data frame, reducing frame header overhead and transmission interaction counts, and adapting to low-bandwidth, high-latency transmission scenarios.

[0073] The ultra-long data frame does not contain any channel identifiers or timestamp information; it only implicitly distinguishes the data from different channels by a fixed order.

[0074] It should be noted that this implicit differentiation method is an important condition for achieving high compression. Because the order can remain stable, channel identifiers and timestamps can be regarded as redundant information. Traditional multi-channel data usually relies on explicit identifiers for differentiation, while this solution achieves implicit differentiation by using a stable order, which can reduce the amount of redundant data.

[0075] It should be noted that the specific distinction rules are as follows: The backend parsing module pre-stores the number of channels and the length of the original data segment for each channel. During parsing, it extracts data segment by segment from the frame header according to the splicing order of the ultra-long data frames. The segment length is the preset length of the original data segment. The first segment is the data of channel 1, the second segment is the data of channel 2, and so on, until all channel data is extracted. For example, if there are 3 channels and each data segment is 32 bytes long, then bytes 0-31 after the frame header represent channel 1 data, bytes 32-63 represent channel 2 data, and bytes 64-95 represent channel 3 data. The ultra-long data frame only retains the original acquired sampled values ​​and does not carry additional information such as channel number, channel identifier, timestamp, or start offset encoding. Data from different channels occupies fixed-length, fixed-order segments within the ultra-long data frame. The channel number is implicitly and uniquely determined by the position index, making the parsing logic simple and requiring no additional computing power.

[0076] This step can reduce the amount of redundant data, achieve high-rate lossy compression transmission, adapt to engineering scenarios such as wireless long-distance transmission and storage constraints, reduce transmission and storage costs, and the order of the compressed data remains unchanged. After decompression, it can be directly used for high-precision signal processing.

[0077] The concatenated, ultra-long data frame is used as the output of this acquisition round for subsequent processing. Alternatively, if the backend supports implicit parsing, the ultra-long frame can be directly used as the output format to adapt to the interface requirements of different backend processing systems.

[0078] The specific output method is as follows: the central node sends the spliced ​​ultra-long data frame to the back-end storage or processing module through the transmission interface. During transmission, CRC32 verification is used, and the verification code is appended to the end of the ultra-long data frame (4 bytes) for the back-end to verify the data integrity. After receiving the data, the backend module first verifies the CRC32 checksum. If the verification passes, it parses the data according to the implicit differentiation rules mentioned above to restore the original data of each channel. If the verification fails, it requests the central node to resend the data for that round. The central node directly sends the spliced ​​ultra-long data frame to the storage, transmission, or backend algorithm module. The subsequent processing module, knowing the number of channels, the length of data per channel, and the fixed order in advance, extracts data from the corresponding position of the ultra-long data frame according to the preset structure, thus restoring the synchronized data of each channel. No additional parsing, verification, or mapping is required, reducing the backend processing pressure and latency. Moreover, the restored data stream is completely consistent with the synchronization alignment result and can be directly input into high-precision signal processing algorithms. This step achieves extremely simplified data stream output, reduces bandwidth, storage, and computing resource consumption, demonstrates the practical value of the invariant sequence design, and adapts to resource-constrained engineering scenarios.

[0079] After the central nodes are sorted in the restored order, the following steps are used to generate time coordinates that can be used by subsequent algorithms: The first arriving data packet is taken as the zero point of time. After verification, the time of the first arriving data packet in this round is recorded as the zero point of time. This zero point is only a relative starting point, determined by the arrival order, and is unrelated to the channel clock, start time, and drift. This virtual time base is established entirely based on the order and is a derivative application of the invariant order.

[0080] Traditional time coordinate generation relies on high-precision clocks to provide absolute timestamps. This solution, however, constructs a relative time base based on the order of data arrival, which can meet the algorithm interface requirements without relying on high-precision clocks.

[0081] Specifically, the central node sets up a time base register for each round. When the first data packet of that round is received, the central node reads the time value of its local system clock (with millisecond precision) and uses this time value as the zero point of that round, storing it in the time base register. The virtual time coordinates of all subsequent data packets in that round are calculated based on this zero point, independent of the local clocks of each channel. This virtual time base is independent of the actual clock drift of each channel, reducing dependence on hardware clock precision and avoiding time coordinate deviations caused by insufficient clock precision, thus adapting to engineering scenarios using conventional industrial-grade clocks. This step establishes a globally unified time base in the simplest way, avoiding deviations from multiple clock sources, providing a standardized timing interface for subsequent algorithms, and ensuring that the relative relationship of the time coordinates perfectly matches the data order, without affecting the algorithm's processing accuracy.

[0082] Based on the preset nominal sampling interval, virtual sampling time is allocated sequentially and incrementally to the data packets arranged in the same sampling round.

[0083] Furthermore, to ensure the output data conforms to a standard timing format, a uniform interval assignment method is adopted. The virtual timestamps here are only used to meet the interface requirements of subsequent algorithms; their relative relationships are guaranteed by the data order, not the actual acquisition times of each channel's data, which differs from the precise timestamps in traditional solutions. High-precision signal processing typically requires input data to have a standard timing relationship. This solution generates compliant virtual time coordinates through sequence and nominal intervals, no longer relying on absolute timestamps.

[0084] Specifically, the nominal sampling interval corresponds to the system sampling rate. In this embodiment, the sampling rate is 100Hz, and the nominal sampling interval is 10 milliseconds. Starting from the zero point of the current round, virtual time is allocated sequentially according to the order counter value of the data packets within the round. The virtual acquisition time is obtained by multiplying the zero point of the time by the order counter value and the nominal sampling interval. Data packets within the same channel are continuously allocated according to the order counter value, while data packets from different channels are interleaved according to the expected reception order, ensuring that the relative relationship of the virtual time is consistent with the timing relationship of data acquisition. The nominal sampling interval matches the system sampling rate; 100Hz corresponds to 10 milliseconds. Starting from zero, virtual time points are assigned sequentially in order, with each subsequent time point equal to the previous time point plus the nominal sampling interval. The generation logic is simple and has no accumulated error. This step can generate a standard, equally spaced time axis, which can be directly used by high-precision signal processing algorithms such as time difference direction finding and coherent accumulation, without the need for additional timing calibration, thus improving algorithm processing efficiency.

[0085] The virtual time is written to the header of the output data stream as a standardized timestamp for the data packet. Specifically, the virtual acquisition time is converted into a 64-bit timestamp (in milliseconds) and written to the 8-byte field reserved in the header of the data packet (located after the start offset encoding field). For ultra-long data frames transmitted in compressed format, virtual timestamps are stored in separate timestamp frames according to the order of data segments from each channel. Each timestamp frame corresponds one-to-one with the ultra-long data frame, and its format is round identifier (1 byte) + N 64-bit timestamps (N being the number of channels). During backend parsing, the timestamps are associated with the data segments sequentially. The virtual timestamps are written into preset fields in the data stream header, forming a complete output frame along with the channel identifier and the original data. Subsequent algorithms can directly read the tags to determine the relative time sequence without further synchronization calibration, reducing processing complexity and latency, and adapting to real-time processing requirements. This standardized time tag is an extension of the invariant time sequence; its core value lies in providing a unified interface rather than transmitting absolute time information. Furthermore, the tags correspond one-to-one with the data order, ensuring consistency in algorithm processing. This step enables the output data to inherently possess standard time coordinates, reducing the processing complexity of subsequent modules, ensuring the consistency of the algorithm's input data, and without altering the synchronization alignment of the data itself.

[0086] The virtual acquisition time allocation process does not read the actual local clock value of any acquisition channel. It should be understood that this design deliberately decouples the system from the channel's local clock to completely eliminate dependence on hardware clocks. This echoes the core idea of ​​this invention to make time synchronization issues irrelevant. The allocation of virtual times depends only on the order and nominal interval, and is independent of the actual clocks of each channel, avoiding timing chaos caused by inconsistencies in channel clocks, while further solidifying the design goal of eliminating the need for real-time compensation calculations.

[0087] The specific implementation guarantee is as follows: when allocating virtual time, the central node only reads the zero point of the local system clock (only once) and the sequence counter value of the data packet, and does not perform clock synchronization or clock reading interaction with any acquisition channel; the local clock of each acquisition channel is only used to control its own acquisition rhythm, and its accuracy deviation or drift is isolated by the safety boundary of the start offset, and does not affect the allocation accuracy of the virtual time.

[0088] The entire allocation process only uses the arrival order and nominal sampling interval, without reading, using, or correcting any channel local clock; the speed and drift of the channel clock do not affect the allocation result, only the correct order needs to be ensured, adapting to acquisition channels with different clock precision, and reducing the threshold for hardware selection.

[0089] This step reduces the impact of clock drift on synchronization from a mechanism perspective, achieving a synchronization method without calibration or correction, improving the compatibility and reliability of the solution, and the generated virtual time stamp fully meets the interface requirements of high-precision signal processing algorithms.

[0090] In the output stage, the central node outputs data streams or ultra-long data frames with standardized time tags to the storage and analysis terminal, reducing the traditional calibration process, reducing the computational load, shortening the data output delay, reducing the possibility of errors introduced by the calibration algorithm, adapting to the engineering requirements of real-time storage and real-time analysis, and the output data has been synchronized and aligned, and can be directly used for high-precision signal processing.

[0091] The system also includes the following steps to dynamically adjust the expected maximum continuous operating time: The central node records the local arrival time of each data packet (for internal monitoring only, not for data alignment or synchronization calibration) and calculates the deviation between the arrival time interval of two adjacent data packets and the nominal sampling interval. This arrival time interval is unaffected by the clock drift of each channel and only reflects the uniformity of the central node's local clock, thus not introducing additional synchronization errors. To achieve long-term stable operation, the system introduces a dynamic adaptive mechanism. It should be noted that the received timestamp here is only used to monitor drift trends and is not used for data alignment or time compensation, which differs from the purpose of timestamps in traditional schemes. This avoids adjustment deviations caused by timestamp errors, and the monitoring logic is simple, does not increase the additional computing power burden, and fully follows the core design of not requiring real-time compensation calculations.

[0092] Specifically: When the central node receives each data packet, it records the receiving timestamp of the local system clock (with millisecond precision) and stores it in the temporary buffer of that data packet; For data packets within the same channel and the same round, after sorting them by sequential counter value from smallest to largest, the receiving timestamps of two adjacent data packets are taken, and the difference between the latter and the former timestamps is calculated to obtain the actual receiving time interval. The nominal sampling interval is a preset 10 milliseconds. The difference between the actual receiving time interval and 10 milliseconds is calculated. This difference is the receiving time interval deviation between two adjacent data packets. A positive deviation indicates that the actual interval is greater than the nominal interval, and a negative deviation indicates that the actual interval is less than the nominal interval. For each adjacent data packet pair generated in each channel, the above calculation is performed to obtain a continuous deviation sequence. In this step, the central node records the reception timestamp of the verified valid data packets, calculates the actual reception interval for consecutive data packets in the same channel, and subtracts it from the nominal sampling interval to obtain the deviation value. This deviation only reflects the relative speed of the clock and does not depend on any absolute clock. The calculation logic is simple and requires little computing power. This step can obtain drift-related characteristics without affecting the synchronization logic, providing data support for duration adjustment, ensuring the stability of the sequence during long-term operation, and avoiding sequence disorder caused by drift accumulation exceeding expectations.

[0093] Based on the cumulative trend of the deviation, the relative drift rate of the local clock of each channel is estimated in real time.

[0094] Furthermore, the instantaneous deviation is transformed into a quantifiable trend indicator. This estimation process is only used to dynamically adjust the safe working window, rather than to perform drift compensation. It is an extension of the drift accumulation characteristic and can achieve trend estimation without the need for a complex drift model, without violating the core principle of not requiring real-time compensation calculation.

[0095] The central node performs sliding accumulation and trend fitting on multiple consecutive sets of deviations to obtain the current relative drift rate. Specifically, the central node is configured with a sliding window buffer for each channel, and the window size K is a fixed integer of 20 (which can be adjusted according to the monitoring accuracy, ranging from 10 to 50). The receive time interval deviations of the K most recent adjacent data packets of this channel are cached in chronological order. When a new deviation occurs, the earliest set of deviations is discarded to ensure that the cache always contains the latest K sets of deviations. Calculate the sliding cumulative bias: Add the absolute values ​​of the K groups of biases in the buffer to obtain the sliding cumulative bias value; Calculate the total time span: K groups of deviations correspond to K+1 data packets, and the total time span is the difference in the timestamp of the K+1th data packet and the first data packet (unit: seconds); Calculate the average deviation change per unit time: Average deviation change = sliding cumulative deviation value / total time span; Calculate the relative drift rate: Relative drift rate = average deviation change / nominal sampling interval, expressed in ppm. This rate directly reflects the degree to which the clock deviates from the nominal frequency. The calculation process only uses accumulation and division operations, and online trend estimation can be completed without establishing a clock drift model.

[0096] This step can obtain the current clock drift status in real time, providing a quantitative basis for dynamically adjusting the expected longest continuous working time, enabling the safety boundary to adapt to complex scenarios with changing drift rates, and ensuring that the constraints of unchanged order always hold.

[0097] When the estimated relative drift rate is lower than the preset threshold, the expected maximum continuous working time is extended; when the estimated relative drift rate is higher than the preset threshold, the expected maximum continuous working time is shortened and a reconfiguration warning is issued.

[0098] In one implementation, the preset threshold is calculated as: upper bound of maximum relative drift accumulation / initial expected maximum continuous working time. For example, in this embodiment, the threshold is approximately 10.368 seconds / 86400 seconds ≈ 120 ppm, and the initial expected maximum continuous working time is 24 hours. The specific adjustment logic is as follows: if the estimated relative drift rate of a channel is lower than the preset threshold for three consecutive times, the expected maximum continuous working time of that channel is extended by 20%, and the extended duration does not exceed twice the initial duration. If the relative drift rate of a channel is estimated to be higher than the preset threshold twice in a row, the expected maximum continuous working time of that channel will be shortened by 30%, and the shortened time will not be less than 50% of the initial time. If the relative drift rate of all channels is higher than the preset threshold, a reconfiguration warning will be issued, prompting engineers to remeasure the drift characteristics and adjust the start-up time offset.

[0099] If the drift is slow, the working time is extended to improve the system's continuous working capability, reduce the number of re-initializations, and adapt to long-term uninterrupted data acquisition scenarios. If the drift is fast, the working time is shortened to reduce the risks caused by sequence changes and to issue early warning prompts for timely intervention by engineers. This adjustment logic is still based on the core premise that the cumulative drift does not exceed the starting offset. It is a flexible application of the understanding that drift is bounded within a limited working period. It can adapt to drift dynamic changes without manual adjustment, and the adjustment process does not involve any real-time compensation calculations or clock calibration.

[0100] This step can improve the robustness of the system under complex scenarios such as temperature changes and long-term operation, ensure that the core objective of maintaining the same order under different operating conditions, and ensure that the synchronization alignment effect continuously meets the requirements of high-precision signal processing.

[0101] The adjustment of the expected longest continuous working time does not change the start time offset of each channel, but is only used to update the validity judgment of the sequential invariance in subsequent working periods.

[0102] To prevent the actual cumulative drift from exceeding the safety boundary due to lag in drift rate estimation or sudden temperature changes, this method introduces two safety margins in the dynamic adjustment: First, the estimated relative drift rate uses the average value of a 10-second sliding window, rather than the instantaneous value, to filter out short-term jitter; Second, when determining whether the cumulative drift amount within the shortened working time is less than the difference between the start time offset and the estimated drift rate, the estimated drift rate is multiplied by a safety factor of 1.5, and then the cumulative drift amount is calculated. If the drift amount is still less than the difference between the offset and the estimated drift amount after multiplying by 1.5, then continued operation is allowed; otherwise, a re-initialization is triggered immediately, without waiting for the next round of self-checks.

[0103] The above design ensures that even in the worst-case scenario of a sudden change in drift rate (e.g., the rate doubles within 1 second), the system still has sufficient response time and safety redundancy to maintain the order.

[0104] It should be noted that the duration adjustment only affects the validity judgment window and does not change the established synchronization structure. This also reflects the open-loop front-end design concept of this invention. There is no need to adjust the start offset to adapt to drift. Only the safe working window needs to be dynamically adjusted to avoid data corruption during the adjustment process. Moreover, the adjustment logic is simple and will not increase the complexity of the system.

[0105] The specific validity judgment method is as follows: the adjusted expected longest continuous working time is used to judge whether the safety boundary of the current start time offset is sufficient, that is, whether the start time offset is still greater than the product of the relative drift rate and the adjusted expected longest continuous working time; If the adjusted start time offset is found to be less than the product, a re-initialization process is triggered; otherwise, operation continues with the current start time offset. Duration adjustment only changes the effective judgment range for drift exceeding limits self-checks, without modifying core parameters such as start offset, splicing order, and time coordinate generation rules. This ensures uninterrupted system operation and data integrity, adapting to the engineering requirements of continuous data acquisition and further solidifying the stability of unchanged order.

[0106] The system also includes a step for self-checking drift exceeding limits using the receiving order: The central node records the acquisition channel identifier corresponding to each received data packet. After completing a round of receiving valid data packets, the central node records the channel identifiers in the actual arrival order, forming the channel identifier sequence for the current round. Specifically, the central node sets up an identifier sequence buffer for each round. Following the order in which data packets are received, the channel number of each data packet is written into the buffer, forming a one-dimensional array of channel identifier sequences. The length of this sequence is the same as the total number of data packets in that round, and each element is an 8-bit channel number. This sequence is stored in the central node's temporary memory. After the round processing is complete, the sequences of three rounds are retained for traceability, and the rest are automatically deleted. This record is only related to the physical arrival order and does not rely on clock, timestamp, or offset calculations. Its recording is based on sequence rather than time, making the operation simple and reliable with no additional sources of error. This is a direct monitoring of the unchanged sequence, rather than monitoring the drift itself, which aligns with the core idea of ​​this invention: prioritizing sequence over time.

[0107] This step obtains real-time sequence characteristics in the simplest way, providing objective data for exceeding limits, avoiding misjudgments caused by time parameters, and ensuring that the self-checking logic does not deviate from the core of maintaining the sequence.

[0108] The currently received channel identifier sequence is compared bit by bit with the preset startup sequence.

[0109] To detect inconsistent temperature trends, the central node performs the following additional checks: After each round of reception, it calculates the statistical value of the offset direction of the actual reception order of each channel relative to the expected order (e.g., a channel consistently leading or lagging behind its expected position). If the offset direction of the same channel is consistent and the offset increases for five consecutive rounds, it is determined that there is a local temperature anomaly at the location of that channel. At this time, the central node issues a "temperature trend inconsistency" alarm and suggests redeploying the channel or adding insulation measures. Before the alarm is cleared, the system automatically shortens the expected maximum continuous working time to 50% of the original value and triggers log recording, but does not stop data collection to balance security and availability.

[0110] The preset startup sequence is a fixed channel order corresponding to the channel startup time offset from small to large. Specifically, it is a cyclic sequence of the expected receiving order stored in the central node configuration file, and the sequence length is consistent with the channel identifier sequence length of the current round.

[0111] The specific comparison method is as follows: the corresponding elements of the current channel identifier sequence are compared with those of the preset start-up sequence, and each element is checked for consistency. The number of misalignments in the entire sequence is counted and recorded as the misalignment count, and the position index of the misalignment is also recorded. The preset start-up sequence is a fixed channel order corresponding to the channel start-up time offset from smallest to largest. The central node compares the actual channel identifier sequence with the preset sequence element by element and marks the misalignment positions. This comparison only performs sequence matching and does not involve numerical calculations such as time, frequency, or drift. Its essence is a direct verification of the unchanged order, rather than a test of time synchronization accuracy. The calculation logic is simple and the real-time performance is high.

[0112] This invention directly determines the system's operating status through sequential comparison, more closely aligning with the actual goal of synchronization, without requiring complex hardware and algorithms. This step can achieve drift state monitoring through pure logical comparison, with low computational load and high real-time performance, adapting to the engineering needs of low-computing-power central nodes and ensuring timely detection of sequence disorder risks.

[0113] Furthermore, when more than M sequence misalignments occur in N consecutive comparisons, the system is deemed to have exceeded the preset range of maximum relative drift accumulation. In this embodiment, N=5 and M=2 (i.e., misalignment occurs in more than 2 out of 5 consecutive rounds). In practical applications, this can be adjusted within the above range according to system reliability requirements. A larger N indicates a more reliable decision but a slower response, while a smaller M indicates greater sensitivity. To avoid misjudgments caused by transmission jitter, this invention employs a statistical decision method, where N is the number of consecutive detection rounds and M is the maximum allowed number of misalignments, the specific values ​​of which are determined based on the round period and the number of channels. The decision logic is as follows: the central node records the misalignment count for N consecutive rounds; if misalignment occurs in more than M rounds, it is determined that the system has exceeded the preset range of the maximum relative drift accumulation. If the number of misaligned rounds in N consecutive rounds does not exceed M, it is considered normal, and only a misalignment log is recorded. The core of this judgment logic is to determine whether the sequence is invalid, rather than focusing on the specific value of the drift amount, which is fundamentally different from the traditional scheme that triggers compensation based on the drift amount threshold. The judgment is based solely on the statistical characteristics of sequence misalignment, without relying on clock parameters or temperature measurements. It exhibits strong anti-interference capabilities, and the decision result is directly related to the core target whose sequence remains unchanged, ensuring that the synchronization alignment effect is not compromised. This step enhances self-test robustness through statistical judgment, effectively distinguishing between transmission jitter and actual drift exceeding limits, ensuring stable system operation in complex communication environments, and guaranteeing that data can always be directly used for high-precision signal processing.

[0114] The automatic reinitialization process is triggered to reallocate the start time offset of each channel. Upon exceeding the limit, the central node automatically issues a reset command. The specific process is as follows: the central node suspends data acquisition and processing for the current round and broadcasts the reinitialization command to all acquisition channels; upon receiving the command, each channel stops acquisition, resets the round counter and sequence counter to 0, and starts the timer to reload the initial start time offset; the central node remeasures the current extreme frequency drift rate of each channel, recalculates the upper limit of the maximum relative drift accumulation according to the aforementioned calculation steps, and then reallocates the start time offset of each channel by a factor of 2, writing it to the local storage module of each channel. After re-initialization, the central node sends a start command, and each channel begins acquisition according to the new start time offset, restoring normal operation. The entire process requires no manual intervention and takes no more than 5 minutes, ensuring rapid system recovery. Each acquisition channel resets its acquisition status. The system regenerates and allocates the startup time offset according to the aforementioned calculation method for the maximum relative drift accumulation, restoring sufficient sequential safety redundancy; Traditional synchronization failures are often resolved by recalibrating the clock or enhancing compensation. This solution, however, rebuilds a stable sequence simply by reallocating the startup offset. Its self-healing logic is simpler, and the data synchronization alignment is quickly restored after self-healing without affecting subsequent high-precision signal processing. This step enables self-healing after drift exceeds limits, ensuring the system continues to operate reliably under long-term, extreme temperature conditions, and guaranteeing the long-term effectiveness of the core objective of maintaining the unchanged sequence.

[0115] Because the difference in startup time offset is greater than the maximum relative drift accumulation, and communication out-of-order has been eliminated in advance, the final order of data packets in each channel is consistent with the startup order, without the need for any clock correction or synchronization calibration.

[0116] The core innovation of this invention lies in the fact that by using the core technical feature of preset start-up time offset and the difference between offsets being greater than the maximum relative drift accumulation, combined with communication disorder elimination, sequence self-checking and automatic reset, dynamic adjustment of the expected longest continuous working time, and ultra-long data frame compression transmission based on sequence invariance, the complex synchronization problem is simplified into a sequence guarantee problem. This achieves multi-channel synchronization without external synchronization signals, additional sensors, or real-time compensation calculations, and the aligned data can be directly used for high-precision signal processing algorithms.

[0117] In this embodiment, the maximum relative drift upper limit is 10.368 seconds, and the starting offset difference is set to twice, i.e., 21 seconds. Even under extreme drift conditions, the sequence can remain stable. When long-term operation leads to drift accumulation and continuous misalignment exceeds the limit, the system can automatically reinitialize and redistribute the offset. At the same time, by dynamically adjusting the duration and compressing the transmission, the system balances stability, efficiency, and bandwidth usage, ensuring synchronization reliability from multiple dimensions such as principle, anti-out-of-order, self-testing, self-healing, dynamic adaptation, and efficient transmission.

[0118] First, this invention uncovers and applies an inherent law that has long been overlooked by existing technologies: as long as the data packet reception order of each channel is consistent with the startup order, data alignment can be achieved through simple sorting; when the initial startup offset is greater than the maximum relative drift accumulation, the order can remain stable for a long time. Conventional approaches tend to actively combat drift, while this solution utilizes the characteristic of slow drift accumulation to isolate the drift impact with a safety boundary, thereby overcoming the constraints imposed by time synchronization.

[0119] Second, this invention simplifies the technical approach by transforming the traditional "multi-channel acquisition times aligned on the same time axis" into "multi-channel data packet order kept constant." After reconstruction, the system no longer relies on timestamps, external references, and real-time feedback, and can achieve stable open-loop operation with only one pre-configuration. The central node is simplified from complex time alignment to basic sorting operations, reducing hardware and computing power requirements while ensuring that the data order meets the input requirements of high-precision algorithms.

[0120] Third, this invention no longer treats clock drift as an error that must be eliminated, but rather utilizes it as a quantifiable and tolerable system characteristic. Drift has a definite cumulative upper limit; as long as the initial offset covers this upper limit, the order can be guaranteed not to be reversed. Stable synchronization can be achieved without complex compensation or additional hardware, improving adaptability in scenarios without external time synchronization and with limited resources.

[0121] Building upon this foundation, the solution also offers several extended advantages: relying on sequence invariance to achieve implicit differentiation and compressed transmission of multi-channel data, reducing bandwidth consumption; generating virtual time coordinates based on sequence provides a standard interface for backend algorithms, eliminating reliance on absolute timestamps; achieving drift self-checking and automatic reset through sequence comparison simplifies the system's self-healing logic; and combining dynamic working duration adjustment enables the system to adapt to temperature changes and long-term operating conditions, reducing manual intervention.

[0122] The aforementioned technical effects support each other, enabling the solution to be directly applied to practical scenarios such as portable electromagnetic monitoring equipment, underground engineering electromagnetic acquisition systems, and multi-channel monitoring networks in environments with strong interference. This effectively solves the problem that existing synchronization solutions cannot work stably in scenarios without external time synchronization and with limited resources.

[0123] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for multi-channel synchronous acquisition of electromagnetic environment data, characterized in that, include: Multiple acquisition channels are deployed in the same physical area, with each channel operating its local clock independently. A start time offset is pre-assigned to each channel so that each channel starts acquisition sequentially according to a preset time order, and the difference between the start time offsets of any two adjacent channels is greater than the maximum relative drift accumulation of the local clocks of all channels within the expected longest continuous working time. After each channel is started, it independently performs continuous data acquisition according to its local clock and sends the acquired data packets to the central node in real time. After receiving the data packets, the central node sorts them directly according to the order in which they were received, and outputs the sorted data as synchronized, multi-channel data. Since the difference in the start time offset is greater than the maximum relative drift accumulation, the receiving order of data packets in each channel is always consistent with the start time order, and no timestamp correction or synchronization calibration is required during the acquisition process.

2. The method for multi-channel synchronous acquisition of electromagnetic environment data according to claim 1, characterized in that, The maximum relative drift accumulation is determined by the following steps: The frequency drift characteristics of the local clock of each acquisition channel within the expected operating temperature range are obtained, and the frequency drift characteristics include the monotonic correspondence between the drift direction and temperature change; Based on the principle that all channels within the same physical region experience the same temperature change trend, it is determined that the relative drift of any two channels shall not exceed the sum of their absolute drifts, and the sum of the absolute values ​​shall be taken as the upper limit of safety. The upper bound of the maximum relative drift accumulation is obtained by multiplying the sum of the largest and second largest absolute drift in each channel, or by taking twice the maximum absolute drift when all channels are of the same type, by the expected longest continuous operating time.

3. The method for multi-channel synchronous acquisition of electromagnetic environment data according to claim 1, characterized in that, It also includes a step of performing a drift over-limit self-check using the receiving order: The central node records the acquisition channel identifier corresponding to each received data packet; Compare the currently received channel identifier sequence with the preset startup sequence bit by bit; When more than M sequence misalignments occur in N consecutive comparisons, the system is determined to have exceeded the preset range of maximum relative drift accumulation. Trigger an automatic reinitialization process to reallocate the startup time offsets for each channel.

4. The method for multi-channel synchronous acquisition of electromagnetic environment data according to claim 1, characterized in that, Each acquisition channel's data packet contains the channel's start time offset encoding. The central node performs implicit synchronization verification through the following steps: Extract the start time offset code of each channel from the received data packets; Based on the size relationship of the encoded values, a desired receiving order is generated; The actual receiving order is compared with the expected receiving order. If they match, the current order is confirmed to be valid; otherwise, the current data packet is discarded and the system waits for the next round of data packets.

5. The method for multi-channel synchronous acquisition of electromagnetic environment data according to claim 1, characterized in that, The difference in startup time offset is greater than the maximum relative drift accumulation setting, specifically including the following steps: Pre-measure the extreme frequency drift rate of the local clock of each acquisition channel at the extreme operating temperature; Multiply the extreme frequency drift rate by the expected longest continuous operating time, and then multiply by a temperature change rate correction factor to obtain the extreme cumulative drift of a single channel. Add the extreme cumulative drift values ​​of any two channels to obtain the upper bound of the maximum relative cumulative drift value; Set the difference in startup time offsets to an integer multiple of this upper bound, where the multiple is greater than or equal to 2.

6. The method for multi-channel synchronous acquisition of electromagnetic environment data according to claim 1, characterized in that, It also includes the following steps to eliminate the interference of out-of-order communication links on the receiving order: When each acquisition channel sends a data packet, it embeds the local acquisition round counter value and the sending sequence number within the same round into the data packet; After receiving the data packet, the central node first groups it according to the collection round counter value, and then sorts it according to the sending sequence number within the same round to restore the correct collection order. Data packets from different rounds are not sorted across rounds to avoid cross-round order errors caused by fluctuations in communication latency.

7. The method for multi-channel synchronous acquisition of electromagnetic environment data according to claim 1, characterized in that, After the central nodes are sorted according to the receiving order, time coordinates that can be used by subsequent algorithms are generated through the following steps: The first arriving data packet is taken as the zero point of time; Based on the preset nominal sampling interval, virtual sampling time is allocated sequentially and incrementally to the data packets arranged in the same sampling round. The virtual acquisition time is written into the header of the output data stream as a standardized timestamp for the data packet; The virtual acquisition time allocation process does not read the actual local clock value of any acquisition channel.

8. The method for multi-channel synchronous acquisition of electromagnetic environment data according to claim 1, characterized in that, The following steps are also included to handle situations where the number of channels exceeds the concurrent capacity of the communication link: Multiple acquisition channels are grouped according to the order of start time offset, and the number of channels in each group does not exceed the maximum number of conflict-free concurrency of the communication link; Each channel within the same group is started sequentially according to the rule that the difference in the start time offset is greater than the maximum relative drift accumulation. A protection interval greater than the maximum drift accumulation within a group is set between different groups to prevent the order of groups from being interleaved.

9. The method for multi-channel synchronous acquisition of electromagnetic environment data according to claim 1, characterized in that, The following steps are also included to dynamically adjust the expected maximum continuous operating time: The central node records the timestamp of the data packets received by each acquisition channel and calculates the deviation between the time interval of receiving two adjacent data packets and the nominal sampling interval. Based on the cumulative trend of the aforementioned deviation, the relative drift rate of the local clock of each channel is estimated in real time. When the estimated relative drift rate is lower than a preset threshold, the expected maximum continuous working time is extended; when the estimated relative drift rate is higher than the preset threshold, the expected maximum continuous working time is shortened and a reconfiguration warning is issued. The preset threshold is determined based on the ratio of the upper bound of the maximum relative drift accumulation to the initial expected longest continuous working time. The adjustment of the expected longest continuous working time does not change the start time offset of each channel, but is only used to update the validity judgment of the sequential invariance in subsequent working periods.

10. The method for multi-channel synchronous acquisition of electromagnetic environment data according to claim 1, characterized in that, The steps also include using order invariance to achieve lossy compressed transmission of multichannel data: The central node buffers data packets from all channels received in the same collection round, and concatenates the data from each channel into a super-long data frame in the order of reception, wherein the reception order is consistent with the start time order of each channel; The ultra-long data frame does not contain any channel identifier or timestamp information, and only implicitly distinguishes the data of different channels by a fixed order position; The spliced ​​ultra-long data frame is used as the output of this acquisition round for subsequent processing.

Images