Pulse description word data playback method and system based on structured data block

Abstract

The invention discloses a pulse description word data playback method and system based on a structured data block, which renovates the traditional streaming processing architecture of reading while computing into an intelligent architecture of separating off-line preprocessing and on-line lightweight playback, in the off-line preprocessing stage, the original PDW stream is intelligently analyzed and segmented, and in the on-line lightweight playback stage, the original PDW stream is intelligently analyzed and segmented; packaging into a structured data block rich in metadata such as time, frequency and radiation source identification, and synchronously establishing a multi-dimensional index; in the playback stage, the system directly loads the prefabricated index and the data block, the target data block is quickly positioned by querying the index, and multi-channel synchronization and accurate scheduling playback are realized by using metadata of the data block. According to the method, the technical problems of high random access delay, difficulty in multi-channel synchronization, incapability of intelligent retrieval and the like in a traditional method are fundamentally solved, millisecond retrieval, top-speed scene switching and high-fidelity playback of massive PDW data are realized, and the utilization rate of system resources is remarkably improved.

Description

Technical Field

[0001] This invention relates to signal processing technology, specifically to a method and system for playing back pulse descriptor data based on structured data blocks. Background Technology

[0002] A pulse descriptor word (PDW) describes the key characteristics of each pulse seen by a receiver, including frequency, amplitude, pulse repetition interval, pulse width, time of arrival (TOA), and angle of arrival. Typically, during testing or experimentation, PDW data is recorded for subsequent playback and to facilitate debugging and analysis of the signal sorting program. Existing PDW playback technologies (such as sequential streaming) suffer from the following drawbacks: 1) extremely high latency when randomly accessing massive amounts of PDW data or switching between scenes; 2) extremely difficult and unreliable precise time synchronization for multi-channel playback; 3) inability to perform intelligent retrieval and selective playback based on signal characteristics (such as frequency and radiation source); and 4) heavy computational burden and low system resource utilization during playback. Summary of the Invention

[0003] The purpose of this invention is to propose a method and system for playing back pulse descriptor data based on structured data blocks.

[0004] The technical solution for implementing this invention is: a method for playing back pulse descriptor data based on structured data blocks, characterized in that it includes:

[0005] Offline preprocessing stage: The raw pulse description word stream is analyzed and processed to generate structured data blocks containing metadata areas, and a retrieval index is built based on the metadata;

[0006] Online playback phase: When the system starts, it loads the retrieval index into memory; in response to a playback request, it queries the index to determine the target data block sequence, schedules and loads the target data blocks into the playback buffer, and controls the timing output of the pulse descriptor based on its metadata and the global synchronization clock.

[0007] The offline preprocessing stage and the online playback stage are independent of each other in terms of logic and runtime sequence. The signal sorting and feature extraction calculations in the offline preprocessing stage are not performed in the online playback stage.

[0008] Furthermore, the structured data block includes a data block header and a pulse descriptor data body, wherein:

[0009] The data block header includes a metadata area;

[0010] The metadata area stores summary information describing the overall characteristics of the pulse descriptors within the data block, including at least one of the following: start timestamp, end timestamp, minimum frequency, maximum frequency, average frequency, list of radiation source identifiers, total number of PDWs, and data volume verification code.

[0011] Furthermore, structured data blocks are generated, specifically through the following method:

[0012] Based on preset signal logic boundaries and protective thresholds, the continuous pulse description word stream is divided into data segments and encapsulated into independent data blocks;

[0013] Among them, the signal logic boundary includes at least one of the following: the radiation source identifier changes, the signal parameter jump amplitude exceeds the preset threshold of the historical moving average, and the arrival time interval of two consecutive pulse description words exceeds the silent period threshold.

[0014] The protective threshold includes at least one of the following: the number of pulse descriptors to be encapsulated exceeds a first quantity threshold, or the time span of the pulse descriptors to be encapsulated exceeds a first time threshold.

[0015] Furthermore, the retrieval index includes at least one of the following:

[0016] A B+ tree time index is built based on the start timestamp of the data block, and its leaf nodes store the data block identifier, file offset, and end timestamp.

[0017] A bitmap frequency index is constructed based on the frequency range of the data block, dividing the working frequency band into frequency slots, and establishing a bit vector indicating the relevant data block for each frequency slot;

[0018] The radiation source inverted index is built based on the radiation source identifier of the data block, which creates a list of data block identifiers containing the radiation source for each radiation source identifier.

[0019] Furthermore, in the online replay stage, the specific method is as follows:

[0020] Request parsing and playlist generation steps: Parse the time conditions and signal characteristic conditions in the playback request, query the corresponding indexes in parallel, perform a logical AND operation on the query results, and generate the target data block sequence and its storage location;

[0021] Intelligent scheduling and preloading steps: Based on the generated playlist, the target data block is loaded into the playback buffer using asynchronous I / O and pre-reading mechanisms, and I / O latency is hidden using double buffering or circular buffering.

[0022] Synchronization and real-time output steps: Using a high-precision hardware clock as the global time reference, control the output pulse descriptor words of each playback channel.

[0023] Further, the steps for synchronizing and real-time output are as follows:

[0024] Block-level coarse synchronization: controls different playback channels to start processing data blocks with the same start time base simultaneously according to the playlist;

[0025] Pulse-level precise synchronization: For each pulse descriptor in the playback buffer, calculate its output absolute time. Output absolute time = hardware clock time at the start of playback + (arrival time of the pulse descriptor - scene start time) * speed factor; each channel waits precisely until the hardware clock reaches the output absolute time before outputting the pulse descriptor.

[0026] A pulse descriptor data playback system based on structured data blocks, implementing the aforementioned pulse descriptor data playback method based on structured data blocks, includes:

[0027] The offline preprocessing module is used to analyze and process the raw pulse description word stream, generate structured data blocks containing metadata areas, and build a retrieval index based on the metadata.

[0028] The online playback engine loads the retrieval index into memory when the system starts; in response to a playback request, it queries the index to determine the sequence of target data blocks, schedules and loads the target data blocks into the playback buffer, and controls the timing output of the pulse descriptor based on its metadata and the global synchronization clock.

[0029] The offline preprocessing module and the online playback engine are independent of each other. The online playback engine does not include computational units for signal sorting and feature extraction.

[0030] Furthermore, the offline preprocessing module includes:

[0031] The data block generation unit is used to segment and encapsulate the pulse description word stream into structured data blocks according to the signal logic boundary and protective threshold, and to calculate its metadata;

[0032] The index building unit is used to extract metadata from each data block and build at least one of the following: B+ tree time index, bitmap frequency index, and radiation source inverted index.

[0033] Online playback engines include:

[0034] The index query unit is used to parse the playback request and query the index in memory in parallel to generate the playlist;

[0035] The scheduling and loading unit is used to asynchronously schedule I / O resources according to the playlist and load data blocks using a pre-read mechanism;

[0036] The synchronous output unit is used to perform block-level coarse synchronization and pulse-level fine synchronization based on a high-precision hardware clock, and to control multi-channel data output.

[0037] Furthermore, it also includes:

[0038] The first storage medium is used to store structured data blocks in the form of files;

[0039] The second storage medium is used to store the retrieval index in the form of independent files;

[0040] The online replay engine loads the entire retrieval index file into memory during initialization for index query unit operations.

[0041] Compared with existing technologies, the significant advantages of this invention are: by introducing a series of collaborative and innovative technologies such as data block structure, multi-dimensional indexing, and offline / online separation architecture, it systematically and fundamentally solves the defects of traditional PDW playback methods, and achieves orders of magnitude improvement in random access speed, synchronization accuracy, retrieval intelligence, and resource efficiency. Attached Figure Description

[0042] Figure 1 This is a flowchart of the pulse descriptor data playback system provided by the present invention.

[0043] Figure 2 This is a flowchart of the structured data block generation process of the present invention.

[0044] Figure 3 This is a preferred data block structure diagram of the present invention.

[0045] Figure 4 This is a flowchart of the online system of the present invention responding to a user's playback request. Detailed Implementation

[0046] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0047] 1. System Overall Architecture

[0048] Reference Figure 1 The pulse descriptor data playback method and system of the present invention adopts an intelligent architecture that separates offline preprocessing and online lightweight playback. The system mainly includes two subsystems:

[0049] Offline preprocessing system: This is a standalone data preparation platform, typically deployed on a data processing server. It receives raw, continuous PDW stream files and, through computationally intensive processing, outputs a structured data block file library and a corresponding multi-dimensional index file library. This process is executed once before the playback task begins.

[0050] Online Real-Time Playback System: This is an operator-oriented real-time interactive platform that can be deployed on a playback control terminal or server. Upon task startup, it loads pre-processed index and data block files, responds to user commands, and achieves low-latency, high-precision playback. This system does not include computationally intensive logic such as PDW signal sorting and feature extraction.

[0051] The two systems are coupled through a data block file library and an index file library, and are completely separated in terms of logic and operation.

[0052] 2. Detailed description of offline preprocessing system and methods

[0053] 2.1 Data Block Generation Module

[0054] Reference Figure 2 This module is responsible for transforming the raw PDW stream into structured data blocks, and its core is intelligent data segmentation decision-making. The steps are as follows:

[0055] Step 1: PDW Buffering and Parsing. The raw PDW stream is read in continuously. Each PDW contains at least the following parameters: arrival time, carrier frequency, pulse width, pulse amplitude, and angle of arrival. The PDWs are buffered in a first-in, first-out (FIFO) buffer.

[0056] Step 2: Intelligent Partitioning Decision. This is crucial for ensuring high cohesion and low coupling in data blocks. Partitioning decisions are dynamically triggered based on the following priority rules:

[0057] 1) External event trigger (highest priority): When a scenario marker (such as threat occurrence or phase change) sent by the simulation scenario is received, the process is immediately cut off.

[0058] 2) Signal logic boundary detection: Real-time analysis of the PDW sequence at the head of the buffer. If any of the following conditions are detected, it is determined to be a logic boundary and the signal is cut off:

[0059] The radiation source identifier (derived through a real-time sorting algorithm) has changed.

[0060] Key parameters (such as RF and PW) change abruptly, and the magnitude of the change exceeds the preset threshold (such as 15%) of the historical moving average.

[0061] A silent period occurs when the TOA interval of two consecutive PDWs exceeds the preset value (e.g., 100ms).

[0062] 3) Protective threshold trigger: To prevent data blocks from becoming too large, a protective upper limit is set. When the number of PDWs in the buffer exceeds the first threshold (e.g., 8192), or the time span exceeds the second threshold (e.g., 500ms), a forced split is executed.

[0063] Step 3: Metadata Calculation and Header Generation. Once the cut-off point is determined, the system performs statistical analysis on the PDW sequence to be encapsulated, generates metadata, and constructs the data block header. A preferred data block structure is as follows: Figure 3 As shown, it includes:

[0064] Synchronization word: a fixed value (e.g., 0xAA5555AA) used for fast positioning of the block start.

[0065] Block header: Contains metadata area, storing calculated summary information, such as: TOA of the first / last PDW in this block; minimum, maximum and average frequency of the PDWs in this block; list of radiation source identifiers identified in this block; total number of PDWs contained in this block; cyclic redundancy check code calculated for subsequent PDW data bodies.

[0066] Header checksum: A checksum calculated over the entire header (from the synchronization word to the cyclic redundancy check), used to verify the integrity of the header itself.

[0067] Step 4: Data Block Encapsulation and Writing. The calculated header and the corresponding PDW data body are sequentially written into the data block file to form a complete, self-describing data unit.

[0068] 2.2 Index Building Module

[0069] The preprocessing system traverses all generated data block files, but only reads their headers to build a multi-dimensional index in memory, which is then persisted as an index file.

[0070] Time Index: A B+ tree is constructed using the start time of each data block as the key. Leaf nodes store the data block ID, file offset, and end time. This index supports efficient log-time (O(logN)) location for any point in time.

[0071] Frequency Indexing: A bitmap index is used. The entire operating frequency band (e.g., 0.5-18GHz) is divided into frequency slots of equal width (e.g., each 100MHz). A bit vector is created for each slot, with a length equal to the total number of data blocks. If the frequency range [minimum frequency, maximum frequency] of the i-th data block intersects with a frequency slot, then the i-th bit of the bit vector for that slot is set to 1. This indexing supports constant-time (O(1)) initial screening for specific frequencies.

[0072] Radiation source index: An inverted index is used. A hash table is created where the key is the identifier of all occurrences of a radiation source, and the value is a list of block IDs (sorted by time) of all data blocks containing that radiation source. This index makes the time complexity of a query retrieved by radiation source O(1).

[0073] Example 1: Assume that the offline preprocessing system processes 10GB of raw PDW data, generating 16386 data blocks with a total size of 10.6GB (including header overhead). Subsequently, the index file is built, which is approximately 50MB in size and can be fully loaded into the server memory.

[0074] 3. Detailed description of online real-time playback system and methods

[0075] 3.1 System Initialization

[0076] After the online system starts, the index file (approximately 50MB) is first loaded into memory. The data block files are then opened in a mode that supports random access.

[0077] 3.2 Scheduling and Replay Process

[0078] Reference Figure 4 The process of the online system responding to user replay requests is as follows:

[0079] Step 1: Request Parsing and Playlist Generation. The user issues a command, such as to play back signals with frequencies in the 1.2-1.3 GHz range, starting from an absolute time T=100 seconds. The system parses the command and queries the index in memory in parallel:

[0080] Query the time index to find the data block containing T=100 seconds, and use it as the starting block.

[0081] The frequency index is queried, and the bitmaps of all frequency slots involved in 1.2-1.3GHz are ORed to obtain the bitmaps of all data blocks that may contain the target signal.

[0082] The query results based on the time and frequency conditions are logically ANDed to obtain a final list of target data block IDs (i.e., the playlist) and their file offsets. This process is completed within milliseconds.

[0083] Step 2: Intelligent Scheduling and Preloading. The I / O scheduler initiates asynchronous data block loading requests based on the playlist. A pre-read mechanism is employed: while playing back the current data block, several subsequent data blocks are pre-loaded into the playback buffer. Double buffering or circular buffering techniques are used to ensure that the output engine always has data available, while hiding disk I / O latency.

[0084] Step 3: Synchronization and Real-time Output. This is the core of multi-channel playback.

[0085] Block-level coarse synchronization: Each channel ensures that it starts processing data blocks with the same time reference simultaneously according to the playlist (achieved by comparing the start time of the data block header).

[0086] Pulse-level precise synchronization and output: The output engine uses a highly stable hardware clock (such as a PPS or PTP synchronization clock) as the global time base. For each PDW in the buffer, the following calculation is made: Output absolute time = Actual hardware clock time at playback start + (PDW.TOA - Scene start time) * Speed ​​factor. The engine waits precisely until the hardware clock reaches this absolute time before sending out the PDW. This ensures that the absolute and relative timing relationships of cross-channel signals are completely faithful to the original scene.

[0087] Step 4: Dynamic Process Management. The monitoring module runs continuously and supports:

[0088] Jump: When the user issues a new time point instruction, the process immediately returns to request parsing and playlist generation, generates a new playlist and schedules its loading, achieving instant switching.

[0089] Loop / Pause: Controls the time accumulator of the output engine.

[0090] Exception handling: If the data block header verification fails, the data will be discarded and an alarm will be triggered.

[0091] Example 2: During the online playback phase, a user requests to replay the signal of Radar A for a specific time period from a 10.5GB data block file. The system queries the inverted index of the radiation source and obtains 1200 data block IDs (approximately 750MB of data) related to Radar A within 1 millisecond. The scheduler loads only this 750MB of data, reducing I / O load by 93%. The playback start latency is less than 50 milliseconds, and the multi-channel synchronization error is less than 1 microsecond.

[0092] Accordingly, the core of this invention lies in updating the traditional read-on-demand streaming architecture to an architecture that separates offline preprocessing and online lightweight playback. In the offline preprocessing stage, the original Pulse Description Word (PDW) is segmented and encapsulated into data blocks with headers, metadata of the data blocks is extracted, and a multi-dimensional index is established. In the online playback stage, the index is queried according to the playback request to determine the sequence of data blocks that meet the conditions, the data blocks are scheduled and loaded, and multi-channel precision synchronous playback is achieved based on a global synchronization clock and metadata.

[0093] To address the high latency issues of random access and scene transitions: traditional methods employ streaming storage, requiring sequential scanning to locate the target time, with latency directly proportional to the data volume. This invention, in the offline preprocessing stage, divides the PDW stream into multiple data blocks based on time windows, pulse counts, or signal logic boundaries. Each data block's header contains its start and end times. Simultaneously, a time index (e.g., a B+ tree) is constructed based on the start times of all data blocks. During online playback, upon receiving a jump instruction to time T, the system directly locates the data block containing time T by querying the time index, obtains its storage address, loads only that data block, places it in the playback buffer, and immediately begins playback. This reduces the time complexity of random access from O(N) to O(logN) or even O(1), achieving millisecond-level scene transitions.

[0094] Addressing the challenge of precise time synchronization across multiple channels: Traditional methods involve each channel independently reading the Product Development Video (PDW) stream and relying on real-time software comparison of timestamps, resulting in high computational overhead and susceptibility to I / O latency. This invention provides globally consistent time metadata (start and end times) at the beginning of each data block. At the start or jump of playback, each channel simultaneously locates the data block with the same time reference via indexing, achieving coarse synchronization at the data block level. Subsequently, each channel performs microsecond or nanosecond-level pulse output timing adjustments based on the fine timestamps of the PDW within the data block and a shared high-precision hardware clock, achieving fine synchronization at the pulse level. This simplifies the complex and fragile end-to-end software synchronization into rapid block-level alignment combined with intra-block fine-tuning, significantly improving synchronization accuracy and stability.

[0095] To address the issue of inability to support intelligent retrieval and selective playback based on signal characteristics: Traditional methods lack high-level feature summaries in PDW streams, requiring a full scan to select specific signals. This invention extracts signal feature statistics from each data block during data block generation, such as frequency range, pulse width range, and a list of radiation source identifiers, storing these as metadata in the header. Specialized indexes, such as a frequency range bitmap index and a radiation source inverted index, are then constructed based on this feature metadata. When a user needs to play back signals matching specific characteristics (e.g., frequencies around 1.2 GHz, originating from a specific radiation source), the system quickly filters out matching data blocks by querying the corresponding feature index, creating a filtered playlist, and only playing back these data blocks. This achieves intelligent retrieval and selective playback based on signal characteristics, reducing invalid data I / O by over 90%.

[0096] To address the issues of heavy computational burden and low resource utilization during playback: traditional methods require real-time parsing of PDW formats and feature calculations during playback, consuming significant CPU resources. This invention employs an architecture that separates offline preprocessing from online playback. The offline preprocessing stage completes all computationally intensive tasks, including signal sorting, feature extraction, data block encapsulation, and index construction. The online playback stage only needs to load the index and schedule data blocks; the playback engine simply outputs the PDW data according to timestamps, without performing any signal processing calculations. This removes the computational burden from the real-time playback path, making the playback process extremely lightweight, significantly reducing CPU utilization, and allowing system resources to focus on I / O scheduling and multi-channel synchronization. This enables high-performance playback of massive amounts of PDW data on ordinary commercial hardware.

[0097] In summary, compared with traditional streaming playback methods, this invention has achieved the following significant improvements:

[0098] 1. Order-of-magnitude improvement in retrieval and location speed: Replacing full scans with index queries reduces time jumps from minutes to milliseconds.

[0099] 2. Fundamental optimization of system resource utilization: The offline / online separation architecture enables heavy computation tasks to be performed in advance, the online playback path is greatly simplified, and the CPU utilization rate is reduced by more than 70%, making it possible to process massive amounts of data on general-purpose hardware.

[0100] 3. Breakthrough in functional flexibility: Based on inverted indexes, bitmap indexes, etc., semantic-level intelligent retrieval and selective playback of PDW data based on radiation source and frequency band were realized for the first time.

[0101] 4. Significantly enhanced synchronization accuracy and reliability: Block-level coarse synchronization achieved by utilizing the global time base in the data block header, combined with fine synchronization by the hardware clock, solves the drift and reset problems of traditional software synchronization.

[0102] The above description is merely a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical methods and inventive concepts of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A method for playing back pulse descriptor data based on structured data blocks, characterized in that, include: Offline preprocessing stage: The raw pulse description word stream is analyzed and processed to generate structured data blocks containing metadata areas, and a retrieval index is built based on the metadata; Online playback phase: The retrieval index is loaded into memory when the system starts; In response to a playback request, the index is queried to determine the sequence of target data blocks, the target data blocks are scheduled and loaded into the playback buffer, and the timing output of the pulse descriptor is controlled based on its metadata and the global synchronization clock. The offline preprocessing stage and the online playback stage are independent of each other in terms of logic and runtime sequence. The signal sorting and feature extraction calculations in the offline preprocessing stage are not performed in the online playback stage.

2. The pulse descriptor data playback method based on structured data blocks according to claim 1, characterized in that, The structured data block includes a data block header and a pulse descriptor data body, wherein: The data block header includes a metadata area; The metadata area stores summary information describing the overall characteristics of the pulse descriptors within the data block, including at least one of the following: start timestamp, end timestamp, minimum frequency, maximum frequency, average frequency, list of radiation source identifiers, total number of PDWs, and data volume verification code.

3. The pulse descriptor data playback method based on structured data blocks according to claim 1, characterized in that, The specific method for generating structured data blocks is as follows: Based on preset signal logic boundaries and protective thresholds, the continuous pulse description word stream is divided into data segments and encapsulated into independent data blocks; Among them, the signal logic boundary includes at least one of the following: the radiation source identifier changes, the signal parameter jump amplitude exceeds the preset threshold of the historical moving average, and the arrival time interval of two consecutive pulse description words exceeds the silent period threshold. The protective threshold includes at least one of the following: the number of pulse descriptors to be encapsulated exceeds a first quantity threshold, or the time span of the pulse descriptors to be encapsulated exceeds a first time threshold.

4. The pulse descriptor data playback method based on structured data blocks according to claim 1, characterized in that, The retrieval index includes at least one of the following: A B+ tree time index is built based on the start timestamp of the data block, and its leaf nodes store the data block identifier, file offset, and end timestamp. A bitmap frequency index is constructed based on the frequency range of the data block, dividing the working frequency band into frequency slots, and establishing a bit vector indicating the relevant data block for each frequency slot; The radiation source inverted index is built based on the radiation source identifier of the data block, which creates a list of data block identifiers containing the radiation source for each radiation source identifier.

5. The pulse descriptor data playback method based on structured data blocks according to claim 1, characterized in that, During the online replay phase, the specific method is as follows: Request parsing and playlist generation steps: Parse the time conditions and signal characteristic conditions in the playback request, query the corresponding indexes in parallel, perform a logical AND operation on the query results, and generate the target data block sequence and its storage location; Intelligent scheduling and preloading steps: Based on the generated playlist, the target data blocks are loaded into the playback buffer using asynchronous I / O and pre-read mechanisms, and I / O latency is hidden using double buffering or circular buffering. Synchronization and real-time output steps: Using a high-precision hardware clock as the global time reference, control the output pulse descriptor words of each playback channel.

6. The pulse descriptor data playback method based on structured data blocks according to claim 5, characterized in that, The steps for synchronous and real-time output are as follows: Block-level coarse synchronization: controls different playback channels to start processing data blocks with the same start time base simultaneously according to the playlist; Pulse-level precise synchronization: For each pulse descriptor in the playback buffer, calculate its output absolute time. Output absolute time = hardware clock time at the start of playback + (arrival time of the pulse descriptor - scene start time) * speed factor; each channel waits precisely until the hardware clock reaches the output absolute time before outputting the pulse descriptor.

7. A pulse descriptor data playback system based on structured data blocks, characterized in that, The method for playing back pulse descriptor data based on structured data blocks according to any one of claims 1-6 includes: The offline preprocessing module is used to analyze and process the raw pulse description word stream, generate structured data blocks containing metadata areas, and build a retrieval index based on the metadata; The online playback engine loads the retrieval index into memory when the system starts; in response to a playback request, it queries the index to determine the sequence of target data blocks, schedules and loads the target data blocks into the playback buffer, and controls the timing output of the pulse descriptor based on its metadata and the global synchronization clock. The offline preprocessing module and the online playback engine are independent of each other. The online playback engine does not include computational units for signal sorting and feature extraction.

8. The pulse descriptor data playback system based on structured data blocks according to claim 7, characterized in that, The offline preprocessing module includes: The data block generation unit is used to segment and encapsulate the pulse description word stream into structured data blocks according to the signal logic boundary and protective threshold, and to calculate its metadata; The index building unit is used to extract metadata from each data block and build at least one of the following: B+ tree time index, bitmap frequency index, and radiation source inverted index. Online replay engines include: The index query unit is used to parse the playback request and query the index in memory in parallel to generate the playlist; The scheduling and loading unit is used to asynchronously schedule I / O resources according to the playlist and load data blocks using a pre-read mechanism; The synchronous output unit is used to perform block-level coarse synchronization and pulse-level fine synchronization based on a high-precision hardware clock, and to control multi-channel data output.

9. The pulse descriptor data playback system based on structured data blocks according to claim 7, characterized in that, Also includes: The first storage medium is used to store structured data blocks in the form of files; The second storage medium is used to store the retrieval index in the form of independent files; The online replay engine loads the entire retrieval index file into memory during initialization for index query unit operations.

Images