A spaceflight data pipeline parsing method based on state machine driving and semantic segmentation

Abstract

This application relates to a pipelined parsing method for aerospace data based on state machine-driven and semantic segmentation, belonging to the field of satellite remote sensing. The method includes the following steps: receiving raw data transmission and dividing it into independent channel data streams according to virtual channels; synchronously performing header identification and packet reassembly on each channel data stream to output a complete data packet set; further decomposing the complete data packet set into several logically independent small data block units with consistent update rates based on semantic configuration; allocating an independent parsing thread to each small data block unit for parallel processing, and finally parsing to obtain a structured aerospace data table. The method uses a built-in pipelined state machine to automatically drive the above four steps sequentially based on the execution results of each stage, and enters an error state and records the migration trajectory in case of an anomaly. This application solves the problem of difficult parsing of complex data protocols in novel aerospace payloads.

Description

Technical Field

[0001] This application relates to the field of satellite telemetry technology, and in particular to a method for parsing aerospace data pipelines based on state machine-driven and semantic segmentation. Background Technology

[0002] As the complexity of space missions increases, new payloads such as laser interferometers typically contain multiple functionally independent subsystems. To facilitate onboard debugging of these new payloads and conserve valuable onboard transmission bandwidth and storage resources, their downlink data protocols are designed to be extremely complex, typically employing techniques such as multi-virtual channel interleaving, multi-modal protocol structures, data multiplexing, and multi-rate data interleaving.

[0003] Space data mainly includes two categories: telemetry data and data transmission data. Telemetry data is primarily used for monitoring the health status of satellite platforms and is handled by the satellite system integrator or commercial satellite companies, with relatively standardized processing procedures. Data transmission data, on the other hand, mainly consists of scientific exploration data from the payloads and belongs to individual user units (such as research institutes and universities), each responsible for its own analysis. Due to the differences in the detection targets, operating modes, and protocol designs of various payloads, as well as the varying technical capabilities of different user units and the lack of standardized data transmission data processing procedures, each unit faces significant difficulties in analyzing data transmission data, resulting in substantial duplication of development work.

[0004] Current telemetry processing systems aim to achieve unified processing of telemetry data from multiple models by constructing modules such as basic data management, configuration information management, index tables, and preprocessing. However, this technology still exhibits the following major technical bottlenecks when addressing the complex requirements of new payloads: While existing technologies manage different frame or packet formats through configuration files and a basic database, their core parsing logic and data models are relatively fixed, making it difficult to dynamically adapt to changes in parameter definitions for the same physical byte segment under different operating modes. This necessitates predefining numerous parsing views or modifying underlying code for different modes. Furthermore, the parameter processing algorithms are rigid and lack flexible extensibility, failing to support the flexible definition of complex physical quantities (such as temperature parameters requiring specific compensation algorithms), thus limiting the system's rapid response and accurate parsing capabilities during the debugging of new payloads. It cannot handle multi-rate data interleaving, employing a flat parsing architecture based on full data packets. Its data processing flow is based on a unified timing framework, lacking effective identification, separation, and timing reconstruction mechanisms, resulting in the inability to restore the original timing characteristics and correlations of each parameter, hindering data display and analysis. Finally, the system architecture is closed, data processing efficiency needs improvement, and it cannot achieve on-demand parsing, data isolation, or targeted data distribution according to actual needs. Summary of the Invention

[0005] This application provides a method for parsing aerospace data pipelines based on state machine-driven and semantic segmentation, which can solve the problem of difficult parsing of complex data protocols of new aerospace payloads.

[0006] To achieve the above objectives, this application provides a method for parsing aerospace data pipelines based on state machine-driven and semantic segmentation, comprising the following steps: S0. Initialize the pipeline state machine and define the state sets of channel segmentation state, packet reassembly state, semantic segmentation state, and parameter calculation state. S1. When the pipeline state machine is in the channel segmentation state, it receives the original data transmission data, segments it into independent channel data streams according to the virtual channel in parallel, and automatically migrates to the packet reassembly state after completion. S2. When the pipeline state machine is in the packet reassembly state, it synchronously performs packet header identification and packet reassembly on each channel data stream, outputs a complete set of data packets, and automatically transitions to the semantic segmentation state after completion. S3. When the pipeline state machine is in the semantic segmentation state, the complete data packet set is further decomposed into several logically independent small data block units with consistent update rates based on semantic configuration. After completion, it automatically migrates to the parameter calculation state. S4. When the pipeline state machine is in the parameter calculation state, an independent parsing thread is assigned to each small data block unit for parallel calculation, and finally the structured aerospace data is obtained.

[0007] In this application, a pipelined state machine drives the automatic flow of four stages: channel segmentation, packet reassembly, semantic segmentation, and parameter calculation. This achieves automation and controllability of the parsing process, eliminating the need for manual triggering of each stage and significantly reducing manual operation steps, thus improving parsing efficiency and ease of use. The fully parallelized architecture significantly improves processing speed and fundamentally enhances overall throughput. Through the synergistic effect of two-level decoupled configuration of semantic segmentation and parsing rule files, the parsing challenges of data byte segment reuse, dynamic switching of working modes, and multi-rate data interleaving are fundamentally solved. The entire processing flow is entirely driven by the first configuration file and the second configuration file. The system is driven by configuration files, third-party configuration files, and parsing rule files, decoupling the parsing logic from the protocol format. When faced with new satellite models or payload protocols, only the corresponding configuration files need to be updated or added, without modifying the core code of the parsing software. This greatly enhances the system's versatility and scalability, and shortens the adaptation cycle. It ensures that data is supplied on demand, eliminating the possibility of unauthorized access to the full original data from the process perspective, conforming to the principle of minimization in data security, and effectively reducing the risk of sensitive information leakage. It reconstructs the true time series of each parameter in its original subsystem, providing an accurate and reliable time series foundation for subsequent data analysis, diagnosis, and tracing. Attached Figure Description

[0008] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments of this application will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0009] Figure 1 This is a state diagram of the aerospace data pipeline parsing method based on state machine driving and semantic segmentation provided in the embodiments of this application; Figure 2 The diagram shows the structure of each module in the aerospace data pipeline parsing method based on state machine driving and semantic segmentation provided in the embodiments of this application. Detailed Implementation

[0010] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0011] This application's embodiments are applied to a ground-based testing software, achieving versatility in telemetry and data transmission processing, and providing support for ground-based telemetry / data transmission data analysis. Based on the proposed four-level progressive processing architecture, the specific implementation process is described in detail below, taking the downlink data analysis of a satellite laser interferometer payload as an example.

[0012] like Figure 1 and Figure 2 As shown in the figure, this application provides a method for parsing aerospace data pipelines based on state machine-driven and semantic segmentation, including the following steps: S0. Initialize the pipeline state machine and define the state sets of channel segmentation state, packet reassembly state, semantic segmentation state, and parameter calculation state. S1. When the pipeline state machine is in the channel segmentation state, it receives the original data transmission data, segments it into independent channel data streams according to the virtual channel in parallel, and automatically migrates to the packet reassembly state after completion. S2. When the pipeline state machine is in the packet reassembly state, it synchronously performs packet header identification and packet reassembly on each channel data stream, outputs a complete set of data packets, and automatically transitions to the semantic segmentation state after completion. S3. When the pipeline state machine is in the semantic segmentation state, the complete data packet set is further decomposed into several logically independent small data block units with consistent update rates based on semantic configuration. After completion, it automatically migrates to the parameter calculation state. S4. When the pipeline state machine is in the parameter calculation state, an independent parsing thread is assigned to each small data block unit for parallel calculation, and finally the structured aerospace data is obtained.

[0013] Furthermore, the pipeline state machine maintains the current state pointer and automatically drives the state transition based on the execution results of each stage, the configuration file loading status, and the data validity. The states defined by the pipeline state machine include initialization state, channel segmentation state, packet reassembly state, semantic segmentation state, parameter calculation state, completion state, and error state. When the pipeline state machine enters an error state, perform the following operations: output the error status code and error information to the log and user interface; mark the status of the corresponding file as failed; prevent the file from automatically entering the next parsing stage; and record the complete state transition trajectory; the transition trajectory includes the state before the transition, the state after the transition, the triggering reason, and the timestamp.

[0014] Specifically, S1 further includes the following steps: Receive raw data transmission data that is mixed and interleaved with several physical channels; the raw data transmission data is obtained by mixing and transmitting data from several physical channels on the satellite and receiving it through ground receiving equipment; the physical channels include CAN bus, RS-422 bus or IIC bus; Based on the data frame structure, a first configuration file is created and loaded. The first configuration file adopts JSON format and is used to describe the overall semantic structure and data segmentation rules of the data frame. The semantic structure includes the field arrangement of frame header, frame tail, data field, spacecraft identifier and virtual channel information, and defines the bit width occupied by each field, spacecraft identifier and virtual channel classification information. The parsing engine runs several channel identification coroutines in parallel, extracts all data frames of different virtual channels in the original data transmission data according to the first configuration file, and automatically identifies and segments the data frames in parallel to obtain their own independent channel data streams; the format of the channel data stream is a binary data stream or a binary file; When S1 completes the parallel splitting of all virtual channels and at least one channel data stream is not empty, the state machine automatically transitions from the channel splitting state to the packet reassembly state and enters S2; if the configuration file fails to load or there is no valid data frame during the splitting process, the state machine enters the error state and records the migration trajectory.

[0015] In one specific embodiment, raw data downlinked from the X-band satellite is received, which simultaneously includes CAN and four 422-channel data. The first configuration file stage1_config.json is loaded, defining the frame header as 0x1ACFFC1D, and the virtual channel identifier VCID occupies 6 bits, where VCID=010100 corresponds to the CAN bus, and VCID=010101 corresponds to the laser interferometer scientific data. Multiple channel identification coroutines are run in parallel, and data frames with different VCIDs are written to science_vcid_{VCID}.bin respectively. 500MB of raw data is sorted within 5 seconds. If only laser interferometer data needs to be processed, only the corresponding VCID item is retained in the configuration file, and the other channels are automatically ignored.

[0016] Specifically, S2 further includes the following steps: According to the data packet protocol agreed upon between the satellite and the payload, a second configuration file is created and loaded; the second configuration file adopts JSON format and defines the header structure, packet length information and output file name format of various data packets in a declarative syntax; The header information of the channel data stream is processed in parallel. The packet length of the channel data stream is verified according to the second configuration file. Incomplete packets are discarded and packet loss events are recorded. The target packet type to be extracted is set, and channel data streams that do not match the target packet type are filtered. Finally, a complete set of data packets is obtained. The header of the complete set of data packets is complete and the structure is clear. The incomplete packets are channel data streams whose packet length is not equal to that of the second configuration file. When S2 completes packet reassembly of all channel data streams and outputs a valid complete set of data packets, the state machine automatically transitions from the packet reassembly state to the semantic segmentation state and enters S3; if continuous packet loss exceeds a preset threshold or a fatal error occurs during packet reassembly, the state machine enters an error state.

[0017] In one specific embodiment, the laser interferometer data stream includes a large data mode packet Type0x10 (packet header 0xAA55, packet length 14472 bytes) and a normal working mode packet Type0x11 (packet header 0xBB66, packet length 2686 bytes). A second configuration file is loaded, and science_vcid_010101.bin is scanned concurrently. Upon identifying 0xBB66, 2686 bytes are read to form a complete packet; otherwise, the packet is discarded and the packet loss event is recorded. Only Type0x11 packets are retained by configuration, automatically filtering out irrelevant packets.

[0018] Specifically, S3 further includes the following steps: Based on the data packet protocol corresponding to each payload, a third configuration file is created and loaded; the third configuration file defines the global semantic structure and segmentation rules of the data packets in a declarative syntax; and defines the specific location of the mode identifier byte or mode identifier bit, and its mapping relationship with each working sub-mode; The macroscopic semantic structure of the complete data packet set is parsed according to the third configuration file. The mode feature bits are read, and the current working mode is queried and determined. The semantic structure includes: the specific position of the mode identifier byte or mode identifier bit and its mapping relationship with each working sub-mode, as well as the byte range, time and original data update rate information of one or more multiplexed segments in the data packet. According to the semantic definition in the third configuration file, the complete data packet set is semantically segmented in parallel to obtain several small data block units. The segmentation rules include: common segmentation rules (common_split_rules) applicable to all sub-modes; segmentation rules (split_rules) specific to each sub-mode, where common segmentation rules can be reused through reference fields (ref); and dedicated segmentation rules (CAN_split_rules) for CAN bus data frames. The internal update rate of each small data block unit is consistent and semantically singular, and each small data block unit uniquely corresponds to one sub-mode. Each small data block unit consists of non-multiplexed segment data and multiplexed segment data. The non-multiplexed segment data comes from byte segments in the complete data packet set whose meaning does not change with the working mode (e.g., timecode, satellite status, which exist in any working mode and have the same meaning). The multiplexed segment data comes from byte segments in the complete data packet set whose meaning changes with the working mode (e.g., bytes 100-300, corresponding to laser phase parameters in frequency stabilization mode and scan control status parameters in phase-locked mode). The multiplexed segment data extracted in different working modes are stored in their respective independent files. When S3 completes all semantic segmentation and outputs at least one small data block unit, the state machine automatically transitions from the semantic segmentation state to the parameter calculation state and enters S4; if there is no valid output or pattern recognition fails, the state machine enters the error state.

[0019] In a specific embodiment, the Type0x11 packet contains three operating modes: frequency stabilization mode, phase-locked loop mode, and optical amplification mode. Bytes 100-300 correspond to different parameter semantics in different modes. The update rate of non-multiplexed segments (such as timecode and laser status) is 1Hz, and the update rate of multiplexed segments is 10Hz. A third configuration file is loaded, defining mode feature bits (bits 3 and 2 of byte 98 and bit 2 of byte 99) through flag_bytes_config, and mapping them to the three modes through mode_mapping. In common_split... The `_rules` section defines common segmentation rules (e.g., bytes 0-4 are segmented into subsystem status data, and bytes 5-84 into high-precision time). In each sub-mode configuration, specific scientific data segments are defined using `split_rules`, and common rules can be referenced. After the parser reads the mode feature bits, it writes 1Hz data into a common data block (e.g., laser data.bin) and 10Hz data into a mode-specific data block (e.g., writing stable frequency scientific data.bin in stable frequency mode). The original 2686-byte packet is decoupled into multiple small data blocks with consistent update rates and semantically simple structures.

[0020] Specifically, S4 further includes the following steps: A parsing rule file is created for each small data block unit. If the current data block originates from a bus channel such as CAN, the predefined packet structure is preferentially read from the telemetry XML configuration file as the parsing rule. If the corresponding definition does not exist in the telemetry XML or the current data block originates from a non-CAN channel, a user-defined Excel parsing rule file is loaded. The parsing rule file adopts different formats depending on the data source: if the data block originates from a bus channel such as CAN, the parsing rule file is a telemetry XML configuration file. This is because the aerospace field commonly uses telemetry XML to define the packet structure of telemetry data transmitted by satellite personnel, and the format of this type of data remains consistent regardless of whether it is transmitted through a telemetry channel or a data transmission channel. Therefore, the existing telemetry XML configuration can be reused for parsing. Otherwise, the parsing rule file is an Excel parsing rule file, which adopts a hierarchical structure with multiple nested worksheets. The parsing rule file fully defines the specific parsing rules for all parameters within the corresponding data block. The definition content includes at least the parameter name, data type, data path sequence, and the calculation formula for the parameter value. A dynamic execution mechanism is adopted to encapsulate custom expressions into executable functions based on computational requirements; Small data block units and parsing rule files are distributed to a multi-threaded parsing pipeline; each thread performs parameter calculations on the small data block units according to the parsing rule file and performs parallel calculations on the small data block units through executable functions; after all threads are completed, the structured aerospace data of each small data block unit is obtained. When S4 completes parameter calculation for all parsing threads and generates a structured aerospace data table, the state machine automatically transitions from the parameter calculation state to the completion state. If any parsing thread encounters an exception, the state machine enters an error state, outputs an error message, and prevents the file from automatically entering the subsequent process.

[0021] In a specific implementation, taking FreqStable.bin (frequency-stabilized mode data block) as an example, physical quantities such as laser phase and optical path difference need to be calculated. The parsing rule file parse.xlsx is loaded, where the FreqStable worksheet defines the path order, data type, and processing formula for each parameter. For the key parameter "fine optical path difference," its calculation involves polynomial fitting, temperature compensation, and system delay correction. The Python function definition is directly written into the corresponding cell in Excel. The parsing engine creates an independent thread for this data block, dynamically loads and executes the function, and passes in the original byte decoding result. Other threads process data blocks such as phase-locked loop mode data block PLL.bin and optical amplification mode data block OPAmp.bin in parallel. After all threads are completed, based on the update rate of each data block (10Hz and 1Hz) and a unified time base, a multi-sheet structured aerospace data table with microsecond-level timestamps and containing all interferometer parameters is generated, which is directly used by scientific analysis software.

[0022] The technical solution of this application embodiment has the following advantages or beneficial effects: State machine-driven automatic flow: Through the built-in pipelined state machine, the four stages of channel segmentation, packet reassembly, semantic segmentation and parameter calculation are automatically connected without manual intervention, which significantly improves processing efficiency and system controllability. The full-process parallel architecture: the multi-level parallel processing mechanism from channel, data packet to data block enables the system to meet the low latency requirements of real-time frame-by-frame parsing, as well as to efficiently complete the batch processing of large files afterward, and the overall throughput is fundamentally improved. Configuration decoupling: Through a configuration architecture that separates semantic segmentation from parameter parsing rules, the problem of parsing data byte segment reuse, dynamic switching of working modes, and multi-rate data interleaving is fundamentally solved. Configuration-driven versatility: The entire processing flow is driven by configuration files, decoupling the parsing logic from the protocol format implementation. When faced with new satellite models or payload protocols, only the corresponding configuration files need to be updated or added, without modifying the core code; Data security and on-demand supply: The process eliminates the possibility of unauthorized parties accessing the full amount of raw data, which is in line with the principle of minimization of data security. Precise time series reconstruction: The true time series of each parameter in its original subsystem is reconstructed, providing an accurate and reliable time series basis for subsequent data analysis, diagnosis and tracing.

[0023] The above example fully demonstrates a typical parsing process: starting from the raw mixed data transmitted from the satellite's X-band, the data is segmented to obtain the laser interferometer scientific data stream. Packet reassembly extracts the complete Type 0x11 data packets. Semantic segmentation decouples the large packets into a frequency-stabilized mode-specific data block, FreqStable.bin, and a common data block. Finally, parameter calculation outputs a structured aerospace data table with microsecond-level timestamps. Throughout the process, the pipeline state machine automatically drives the four stages sequentially without manual intervention. If any stage experiences a configuration file loading failure, continuous packet loss exceeding the threshold, or a parsing thread anomaly, the state machine enters an error state, outputs an error message, and prevents further processing, while simultaneously recording the complete state transition trajectory.

[0024] In the several embodiments provided in this application, it should be understood that the disclosed methods and apparatus can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces, or indirect coupling or communication connection between devices or units, and may be electrical, mechanical, or other forms.

[0025] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can be physically included separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in the form of hardware plus software functional units.

[0026] The above description is the preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principles described in this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A space data pipeline parsing method based on state machine driving and semantic segmentation, characterized in that, It includes the following steps: S0. Initialize the pipeline state machine and define the state sets of channel segmentation state, packet reassembly state, semantic segmentation state, and parameter calculation state. S1. When the pipeline state machine is in the channel segmentation state, it receives the original data transmission data, segments it into independent channel data streams according to the virtual channel in parallel, and automatically migrates to the packet reassembly state after completion. S2. When the pipeline state machine is in the packet reassembly state, it synchronously performs packet header identification and packet reassembly on each channel data stream, outputs a complete set of data packets, and automatically transitions to the semantic segmentation state after completion. S3. When the pipeline state machine is in the semantic segmentation state, the complete data packet set is further decomposed into several logically independent small data block units with consistent update rates based on semantic configuration. After completion, it automatically migrates to the parameter calculation state. S4. When the pipeline state machine is in the parameter calculation state, an independent parsing thread is assigned to each small data block unit for parallel calculation, and finally the structured aerospace data is obtained.

2. The state machine driven and semantic segmentation based space data pipeline parsing method according to claim 1, characterized in that, The pipeline state machine maintains the current state pointer and automatically drives the state transition based on the execution results of each stage, the configuration file loading status, and the data validity. The states defined by the pipeline state machine include initialization state, channel segmentation state, packet reassembly state, semantic segmentation state, parameter calculation state, completion state, and error state. When the pipeline state machine enters an error state, perform the following operations: output the error status code and error information to the log and user interface; mark the status of the corresponding file as failed; prevent the file from automatically entering the next parsing stage; and record the complete state transition trajectory; the transition trajectory includes the state before the transition, the state after the transition, the triggering reason, and the timestamp.

3. The state machine driven and semantic segmentation based space data pipeline parsing method according to claim 2, characterized in that, S1 further includes the following steps: Receive raw data transmission data that is mixed and interwoven with several physical channels; Based on the data frame structure, create and load the first configuration file; The parsing engine runs several channel identification coroutines in parallel, extracts all data frames of different virtual channels in the original data transmission data according to the first configuration file, and automatically identifies and segments the data frames in parallel to obtain their respective independent channel data streams.

4. The state machine driven and semantic segmentation based space data pipeline parsing method according to claim 3, characterized in that, The original data transmission data is obtained by mixing and transmitting data from several physical channels on the satellite and receiving it through ground receiving equipment; The physical channels include CAN bus, RS-422 bus or IIC bus; The first configuration file adopts JSON format and is used to describe the overall semantic structure of the data frame and the data segmentation rules. The semantic structure includes the field arrangement of frame header, frame tail, data field, spacecraft identifier and virtual channel information, and defines the bit width occupied by each field, spacecraft identifier and virtual channel classification information. The channel data stream is in the format of a binary data stream or a binary file; When S1 completes the parallel splitting of all virtual channels and at least one channel data stream is not empty, the state machine automatically transitions from the channel splitting state to the packet reassembly state and enters S2; if the configuration file fails to load or there is no valid data frame during the splitting process, the state machine enters the error state and records the migration trajectory.

5. The aerospace data pipeline parsing method based on state machine driving and semantic segmentation according to claim 1, characterized in that, S2 also includes the following steps: Create and load a second configuration file according to the data packet protocol agreed upon between the satellite and the payload; The packet header information of the channel data stream is processed in parallel. The packet length of the channel data stream is verified according to the second configuration file. Incomplete packets are discarded and packet loss events are recorded. Set the target packet type to be extracted, filter channel data streams that do not match the target packet type; finally obtain a complete set of data packets, the header of which is complete and the structure is clear.

6. The aerospace data pipeline parsing method based on state machine driving and semantic segmentation according to claim 5, characterized in that, The second configuration file uses JSON format and defines the header structure, packet length information, and output file name format of various data packets using declarative syntax; The incomplete packet is a channel data stream whose length is not equal to that of the second configuration file packet; When S2 completes packet reassembly of all channel data streams and outputs a valid set of complete data packets, the state machine automatically transitions from the packet reassembly state to the semantic segmentation state and enters S3. If continuous packet loss exceeds a preset threshold or a fatal error occurs during packet reassembly, the state machine enters an error state.

7. The aerospace data pipeline parsing method based on state machine driving and semantic segmentation according to claim 1, characterized in that, S3 also includes the following steps: Based on the data packet protocol corresponding to each payload, a third configuration file is created and loaded; Based on the third configuration file, the macroscopic semantic structure of the complete data packet set is parsed, the pattern feature bits are read, and the current working mode is queried and determined. Based on the semantic definition of the third configuration file, the complete data packet set is semantically segmented in parallel to obtain several small data block units.

8. The aerospace data pipeline parsing method based on state machine driving and semantic segmentation according to claim 7, characterized in that, The third configuration file defines the global semantic structure and segmentation rules of the data packets using a declarative syntax; it defines the specific location of the mode identifier byte or mode identifier bit and its mapping relationship with each working sub-mode; The semantic structure includes: the specific location of the pattern identifier byte or pattern identifier bit and its mapping relationship with each working sub-mode, as well as the byte range, time and original data update rate information of one or more multiplexed segments in the data packet; The segmentation rules include: common segmentation rules applicable to all sub-modes, segmentation rules specific to each sub-mode, and special segmentation rules for CAN bus data frames; the segmentation rules specific to each sub-mode reuse the common segmentation rules by referencing fields. The small data block unit consists of non-multiplexed segment data and multiplexed segment data. The non-multiplexed segment data comes from byte segments in the complete data packet set whose meaning does not change with the working mode, while the multiplexed segment data comes from byte segments in the complete data packet set whose meaning changes with the working mode. The multiplexed segment data extracted under different working modes are stored in their respective independent files. When S3 completes all semantic segmentation and outputs at least one small data block unit, the state machine automatically transitions from the semantic segmentation state to the parameter calculation state and enters S4; if there is no valid output or pattern recognition fails, the state machine enters the error state.

9. The aerospace data pipeline parsing method based on state machine driving and semantic segmentation according to claim 1, characterized in that, S4 also includes the following steps: A parsing rule file is created for each small data block unit. If the current data block comes from a bus channel such as CAN, the predefined packet structure is read from the telemetry XML configuration file as the parsing rule. If there is no corresponding definition in the telemetry XML or the current data block comes from a non-CAN channel, the user-defined Excel parsing rule file is loaded. A dynamic execution mechanism is adopted to encapsulate custom expressions into executable functions based on computational requirements; Small data block units and parsing rule files are distributed to a multi-threaded parsing pipeline; each thread performs parameter calculations on the small data block units according to the parsing rule file and performs parallel calculations on the small data block units through executable functions; after all threads have completed, the structured aerospace data of each small data block unit is obtained.

10. The aerospace data pipeline parsing method based on state machine driving and semantic segmentation according to claim 9, characterized in that, The parsing rule file adopts different formats depending on the data source: if the data block comes from the CAN bus channel, the parsing rule file is a telemetry XML configuration file; Otherwise, the parsing rule file is an Excel parsing rule file, which adopts a hierarchical structure with multiple nested worksheets; The parsing rule file fully defines the specific parsing rules for all parameters within the corresponding data block. The definition includes parameter name, data type, data path order, and calculation formula for parameter value. When S4 completes parameter calculation for all parsing threads and generates a structured aerospace data table, the state machine automatically transitions from the parameter calculation state to the completion state. If any parsing thread encounters an exception, the state machine enters an error state, outputs an error message, and prevents the file from automatically entering the subsequent process.

Images