Real-time data transmission and processing method of intelligent inspection system based on Beidou positioning

Abstract

The invention relates to the technical field of intelligent inspection and industrial internet communication control, and discloses an intelligent inspection system real-time data transmission and processing method based on Beidou positioning, and the method comprises the following steps: obtaining a three-dimensional space coordinate and a speed vector, and collecting operation state data; calculating a network quality index and constructing a spatial network signal quality topology library; performing anomaly detection on the operation state data, outputting scores, and adjusting a positioning sampling frequency and a protocol heartbeat sending interval; calculating a predicted space coordinate, extracting a predicted network quality index, triggering burst transmission before entering a blind area, hanging a radio frequency front end after entering the blind area, and executing lossless compression caching; if serious abnormity occurs in the blind area, abnormal coordinates are broadcasted through short-distance communication, the candidate terminal calculates the relay fitness weight based on the network quality and the spatial distance, and a main relay node is selected to establish a transparent transmission tunnel for message forwarding. According to the invention, high-reliability transmission and real-time interaction of inspection data in a complex shielding environment are guaranteed.

Description

Technical Field

[0001] This invention relates to the field of intelligent inspection and industrial internet communication control technology, specifically to a real-time data transmission and processing method for an intelligent inspection system based on BeiDou positioning. Background Technology

[0002] With the development of industrial digitalization, portable intelligent inspection terminals with satellite positioning modules are widely used in large and complex production scenarios such as petrochemicals and power pipeline corridors to continuously collect equipment operating status parameters and spatial trajectory data of inspection personnel. However, these industrial sites are usually densely packed with large metal storage tanks, dense pipelines, and semi-enclosed spaces, which causes severe physical obstruction and multipath attenuation of wireless communication signals, easily creating network communication blind spots in local areas.

[0003] Existing inspection data transmission systems typically employ fixed timed reporting mechanisms and uniform hardware sampling frequencies. When inspection equipment enters a blind zone without network signal coverage, traditional solutions often can only passively accept communication interruptions and perform basic local buffering. Due to the lack of prior perception of the quality of spatial network signal distribution and the lack of forward-looking prediction of terminal movement trajectories, the equipment cannot transmit high-priority abnormal status data in advance before a complete network outage, resulting in significant time lags in the reporting of critical operating conditions or alarm information.

[0004] Meanwhile, the fixed sampling frequency of the positioning module and the network heartbeat keep-alive interval cannot adapt to dynamic changes in operating conditions. When the equipment has detected abnormalities in on-site environmental parameters or machine operating status, the cloud control center struggles to obtain high-resolution data on on-site spatial trajectories and details of sudden state changes due to the system's preset low-frequency acquisition strategy. Furthermore, when the inspection terminal triggers a serious anomaly alarm within a network blind spot, the isolated terminal device, having lost its connection to the wide area network backbone, cannot autonomously coordinate with other nodes in the surrounding normal communication environment to relay data links, thus creating security risks. This rigid data transmission and hardware scheduling mechanism limits the reliability of data interaction and emergency response efficiency of the inspection system under complex and harsh operating conditions. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a real-time data transmission and processing method for an intelligent inspection system based on BeiDou positioning. This method solves the problems of abnormal data loss, lag, and rigid transmission parameter scheduling caused by sudden changes in network status in complex physical shielding environments.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a real-time data transmission and processing method for an intelligent inspection system based on BeiDou positioning, comprising the following steps: The system acquires the three-dimensional spatial coordinates and instantaneous velocity vector output by the BeiDou positioning module, and simultaneously acquires the operational status data collected by the sensor module. It calculates the network quality index based on the network transmission parameters of the current communication link, performs spatial mapping between the network quality index and the corresponding three-dimensional spatial coordinates, and establishes and dynamically updates a spatial network signal quality topology library in local memory. It performs anomaly detection on the operational status data to output an anomaly severity score, and combines this score to cascade control the sampling frequency of the BeiDou positioning module and the protocol heartbeat transmission interval of the wireless communication module at the hardware and protocol stack levels. It predicts spatial coordinates based on the three-dimensional spatial coordinates and instantaneous velocity vector, and obtains the predicted network quality index by addressing in the spatial network signal quality topology library. When the predicted network quality index is lower than a set blind zone threshold, it triggers queue scheduling to perform burst concurrent transmission before the device enters the network blind zone, and suspends the radio frequency front-end of the wireless communication module after entering the blind zone. It performs lossless compression and local caching on the continuously acquired operational status data and three-dimensional spatial coordinates.

[0007] In the process of constructing and updating the spatial network signal quality topology library, the edge computing unit divides the geographic space into discrete grids and extracts the grid index of the current three-dimensional spatial coordinate mapping. For grid indices with repeated paths, an exponentially weighted moving average algorithm is used, which integrates a preset forgetting factor coefficient, the fresh network quality index calculated at the current sampling time, and the historical network quality feature values ​​stored in the grid index, to perform smooth iteration on the network quality feature values ​​within the grid, thereby filtering transient network jitter data.

[0008] In the anomaly detection and hardware cascade control phase, the temporal characteristics of the operational status data are input into the locally deployed isolated forest detection model to calculate and output an anomaly severity score normalized to the zero-to-one interval. When the anomaly severity score exceeds the preset anomaly judgment threshold, the system writes a control frame to the phase-locked loop frequency divider register of the BeiDou positioning module to positively increase the sampling frequency. Simultaneously, it modifies the long-connection keep-alive timer variable of the network protocol stack to exponentially reduce the protocol heartbeat transmission interval, thereby increasing the trajectory sampling density and connection liveness detection frequency in the anomaly area.

[0009] In the blind zone adaptive transmission scheduling process, when the predicted network quality index is lower than the blind zone threshold and the current real-time network quality index is within the normal range, the device is determined to be in a blind zone critical state. The underlying communication driver clears the regular operation log in the transmission buffer, assigns the highest priority to the operation status data carrying the abnormal marker and historical unacknowledged data frames, and uses all available bandwidth for concurrent burst transmission within the remaining time window before entering the blind zone.

[0010] Furthermore, multi-hop relay cooperative transmission based on spatial clustering is performed within the network blind zone. When the severity score of an anomaly is continuously detected to be greater than the set critical anomaly threshold, the current anomaly coordinates are extracted, serialized with device identifiers and operating status data to generate a relay broadcast data frame, which is then radiated outward through a short-range wireless communication module. Surrounding candidate inspection terminals receive this broadcast and calculate the spatial Euclidean distance between themselves and the anomaly coordinates. Each candidate inspection terminal calculates its relay fitness weight based on the positive gain component of its real-time network quality index and the logarithmic penalty component of the spatial Euclidean distance. Each candidate inspection terminal broadcasts within a random backoff contention time window and locally collects the relay fitness weights within the network. The node with the highest weight value automatically switches to become the primary relay node. The primary relay node establishes a transparent tunnel, receives short-range wireless communication data packets sent by terminals within the blind zone, extracts the underlying protocol messages, and performs WAN transparent forwarding to the cloud server.

[0011] During the phase of exiting the network blind zone, interrupted transmission recovery is performed. The device continuously calculates the predicted spatial coordinates at the base sampling frequency. When the predicted network quality index recovers to equal or higher than the blind zone threshold, a hardware wake-up signal is issued to restore power to the RF front-end and re-establish the network connection. The system prioritizes extracting lossless compressed data blocks carrying anomaly markers from the local cache for encryption and uploading. Subsequently, regular cached data blocks are sent in batches strictly according to the sampling timestamp and monotonically increasing sequence number.

[0012] For cloud-based streaming processing and integration of multi-source heterogeneous data, the cloud server receives the 3D spatial coordinate trajectory stream uploaded by the device, calls a weighted moving average model combined with continuous sampling points from historical backtracking to perform spatial smoothing and correction, and outputs calibrated standard trajectory points. The distributed streaming computing unit writes structured data to the time-series database shard nodes of the storage cluster and defines a heterogeneous system data transformation matrix through a business logic adaptation layer. The system multiplies the original structured data vector with this matrix, performs field name mapping, data type casting, and dimensional unification calculations, and generates target data packets conforming to the interface specifications of external third-party systems for targeted push, achieving closed-loop collaboration between the device's underlying operational data and the enterprise's upper-level management system.

[0013] This invention provides a real-time data transmission and processing method for an intelligent inspection system based on BeiDou positioning. It has the following beneficial effects: 1. This invention constructs a spatial network signal quality topology library and combines it with BeiDou positioning vectors for trajectory prediction, realizing pre-emptive response and adaptive scheduling for communication blind spots. The system calculates the predicted position based on instantaneous velocity and spatial coordinates. When the predicted network quality index is determined to be lower than the blind spot threshold, queue scheduling is triggered and cached data with anomaly markers is sent before the device actually enters the blind spot. After the device enters the blind spot, the radio frequency module is actively suspended and local lossless compression caching is performed, avoiding sudden network outages and data loss caused by physical shielding, and ensuring the integrity and timely reporting of critical operational data.

[0014] 2. This invention couples the data anomaly detection results from the edge side with the underlying hardware configuration logic of the device. By calculating the severity score of the anomaly, it dynamically adjusts the data acquisition and transmission parameters. When the system determines that the device's operating status data is abnormal, it directly modifies the hardware registers and protocol stack variables to positively increase the sampling frequency of the Beidou positioning module and exponentially shorten the network heartbeat transmission interval. This allows the inspection terminal to maintain low power consumption and low bandwidth usage under normal operating conditions, while automatically increasing the trajectory sampling density and connection survival detection frequency under abnormal operating conditions, thereby improving the acquisition accuracy of spatiotemporal data at the fault site.

[0015] 3. This invention addresses the situation where equipment experiences severe anomalies within network blind spots by designing a multi-hop relay collaborative transmission mechanism based on spatial location and network quality assessment. The system utilizes a short-range wireless communication module to broadcast the coordinates of the anomaly. Surrounding candidate terminals calculate relay fitness weights by combining the positive gain of their own backbone network quality index with the attenuation penalty of relative spatial distance, and select the optimal primary relay node through a competition protocol. When the WAN link of the inspection terminal is interrupted, a transparent tunnel is established through a local self-organizing network to forward the underlying protocol messages, ensuring that high-priority alarm events in shielded environments can be reliably delivered to the cloud system. Attached Figure Description

[0016] Figure 1 This is a diagram of the system physical architecture of the present invention; Figure 2 This is the main flowchart of the method of the present invention; Figure 3 This is a logical functional architecture diagram of the present invention; Figure 4 This is a two-dimensional planar diagram showing the mapping between the inspection trajectory and the network quality topology in this invention. Figure 5 This is a graph showing the change in the severity score of the anomaly over time. Figure 6 This is a dynamic response curve of the sampling frequency of the Beidou positioning module of the present invention; Figure 7 This is a dynamic response curve of the MQTT protocol heartbeat sending interval according to the present invention.

[0017] The system comprises: 10. Inspection terminal; 11. Beidou positioning module; 12. Sensor module; 13. Edge computing processor; 14. Wireless communication module; 15. Short-range wireless communication module; 20. Cloud server; 21. MQTT access gateway; 22. Streaming computing unit; 23. Distributed storage cluster; 24. Business logic adaptation layer; 30. Candidate inspection terminal; 40. Real-time data transmission and processing system based on Beidou positioning; 41. Multi-dimensional data acquisition and topology construction module; 42. Anomaly assessment and cascading control module; 43. Blind zone adaptive transmission scheduling module; 44. Multi-hop relay routing control module; and 45. Cloud-based streaming integrated processing module. Detailed Implementation

[0018] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. 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.

[0019] See attached document Figure 1 This invention provides a real-time data transmission and processing method for an intelligent inspection system based on BeiDou positioning, applicable to the intelligent inspection system. The intelligent inspection system includes an inspection terminal 10 and a cloud server 20.

[0020] The inspection terminal 10 includes a BeiDou positioning module 11, a sensor module 12, an edge computing processor 13, and a wireless communication module 14. The edge computing processor 13 is electrically connected to the BeiDou positioning module 11, the sensor module 12, and the wireless communication module 14 via a system bus. The wireless communication module 14 establishes a communication connection with the cloud server 20 via a wireless network. The wireless communication module 14 uses the MQTT protocol for data transmission.

[0021] During the inspection task initiation phase, the edge computing processor 13 sends initialization control commands to the BeiDou positioning module 11, sensor module 12, and wireless communication module 14. The BeiDou positioning module 11 receives the initialization control commands and configures the basic sampling frequency. The wireless communication module 14 receives the initialization control commands and configures the standard heartbeat transmission interval, while simultaneously initiating an MQTT connection request with the cloud server 20.

[0022] Sensor module 12 collects operational status data of the inspection environment according to a preset time period. The operational status data includes environmental eccentricity parameters, equipment pressure parameters, and gas concentration parameters. Beidou positioning module 11 synchronously acquires the three-dimensional spatial coordinates and instantaneous velocity vector of the installed inspection terminal 10 according to the basic sampling frequency.

[0023] Operational status data, three-dimensional spatial coordinates, and instantaneous velocity vectors are transmitted to the edge computing processor 13 via the system bus. The three-dimensional spatial coordinates are represented as follows: ,in, At the current sampling time, The longitude coordinates corresponding to the current sampling time. The latitude coordinates corresponding to the current sampling time. Here are the elevation coordinates corresponding to the current sampling moment. The instantaneous velocity vector is represented as... .

[0024] During data transmission, the wireless communication module 14 records the transmission delay and packet loss rate of the current communication link. The edge computing processor 13 obtains the transmission delay and packet loss rate from the wireless communication module 14 and calculates the network quality index according to the set logic. The formula for calculating the network quality index is as follows: ; in, Indicates the three-dimensional coordinates of the inspection terminal 10 Network quality index at the location; Indicates the current sampling time The corresponding transmission delay; Indicates the current sampling time The corresponding packet loss rate; A preset weighting coefficient representing transmission delay; The preset weighting coefficients represent the packet loss rate, and the weighting coefficients satisfy the following conditions: Constraints, This indicates the preset maximum tolerable transmission delay threshold for the current communication link. At this time... It is transformed into a dimensionless penalty factor.

[0025] The edge computing processor 13 will calculate the network quality index. Three-dimensional spatial coordinates acquired at the same time The edge computing processor 13 performs association mapping processing. Based on the results of the association mapping processing, it generates a spatial network signal quality topology library in the local storage unit of the inspection terminal 10.

[0026] The space network signal quality topology library consists of multiple data nodes, each containing a pair of paired coordinate data and network quality index data. As the inspection task continues, the edge computing processor 13 appends newly acquired correlation mapping results to the space network signal quality topology library according to the sampling period, performing dynamic updates to the topology library. The cloud server 20 synchronously receives the updated data stream of the space network signal quality topology library via the wireless communication module 14 and stores it in a mirrored format in the server database.

[0027] See attached document Figure 1 and attached Figure 2 This invention provides a real-time data transmission and processing method for an intelligent inspection system based on BeiDou positioning. The method specifically includes the following operational logic during the multi-dimensional spatiotemporal data acquisition stage: The BeiDou positioning module 11 receives ephemeris data, pseudorange observations, and carrier phase observations broadcast by the BeiDou satellite navigation system via a radio frequency antenna array. The BeiDou positioning module 11 has a built-in baseband signal processing chip, which performs real-time dynamic differential (RTK) calculations on the pseudorange and carrier phase observations, outputting a positioning data stream in NMEA0183 standard format.

[0028] The edge computing processor 13 reads the positioning data stream at a set baud rate through a universal asynchronous transceiver interface and extracts three-dimensional spatial coordinates containing high-precision timestamps. and instantaneous velocity vector The BeiDou positioning module 11 outputs a pulse-per-second (PPS) signal every second. The edge computing processor 13 receives the PPS signal and uses it as the absolute synchronization reference for the system's global clock.

[0029] Edge computing processor 13 generates a hardware-triggered interrupt instruction based on an absolute synchronization reference and sends the instruction to sensor module 12. In response to the interrupt instruction, sensor module 12 immediately performs analog-to-digital conversion and latching operations on the operating status data. This hardware-level triggering mechanism ensures that the operating status data output by sensor module 12 and the three-dimensional spatial coordinates output by BeiDou positioning module 11 achieve millisecond-level precise alignment in the time dimension.

[0030] The wireless communication module 14 internally runs a TCP / IP protocol stack and an MQTT client daemon. Based on the standard heartbeat transmission interval, the wireless communication module 14 sends protocol control messages (PINGREQ) to the cloud server 20. At the instant the protocol control message is sent, the edge computing processor 13 starts a high-precision local hardware timer.

[0031] When the wireless communication module 14 receives a response message from the cloud server 20 and triggers a receive interrupt, the edge computing processor 13 reads the count value of the high-precision local hardware timer and uses this count value as the transmission delay at the current sampling moment. Meanwhile, the wireless communication module 14 operates within a preset sliding time window. Within, the total number of application data packets sent to cloud server 20 is counted. and the total number of acknowledgment messages successfully received. Edge computing processor 13 computing Packet loss rate within the sliding time window The network quality index was calculated. Subsequently, the edge computing processor 13 allocates contiguous memory space in local non-volatile memory to construct a rasterized spatial network signal quality topology library. The edge computing processor 13 then configures the raster according to a preset raster resolution. This divides continuous geographic space into discrete two-dimensional or three-dimensional grids.

[0032] Edge computing processor 13 will acquire the current three-dimensional spatial coordinates Mapped to the corresponding raster index In China, the network quality index will be included. This is written as an attribute feature value of the raster index. When the inspection terminal 10 repeatedly passes through the same raster index during operation... When generating a new network quality index, the edge computing processor 13 uses the Exponentially Weighted Moving Average (EWMA) algorithm to iteratively update the attribute feature values ​​of the grid.

[0033] The iterative update formula for attribute feature values ​​is as follows: ; in, This indicates the raster index after smooth update. The current network quality characteristic value; This indicates that the raster index was stored in local non-volatile memory before the update. Historical network quality characteristics; Indicates the current sampling time in the raster index. The fresh network quality index is calculated in real time within the specified range. This represents the preset forgetting factor coefficient, whose value range is set to... .

[0034] By iteratively updating the formula, the edge computing processor 13 dynamically maintains the spatial network signal quality topology library, enabling the data nodes in the topology library to continuously reflect the long-term steady-state characteristics of the wireless network quality in the corresponding geographic spatial area, while filtering out transient network jitter data caused by multipath effects or instantaneous electromagnetic interference.

[0035] See attached document Figure 2 The present invention provides a real-time data transmission and processing method for an intelligent inspection system based on BeiDou positioning, which performs abnormal state assessment and underlying hardware parameter linkage control at the edge side of the inspection terminal 10.

[0036] Edge computing processor 13 receives operating status data output by sensor module 12. Edge computing processor 13 uses a built-in moving average filtering algorithm to perform time-domain smoothing and denoising on the ambient temperature, equipment pressure, and gas concentration parameters in the operating status data. After denoising, edge computing processor 13 calculates the mean square error and first-order difference of each parameter within a preset time window, and concatenates the mean square error and first-order difference into a multidimensional anomaly feature vector.

[0037] Edge computing processor 13 inputs multi-dimensional anomaly feature vectors into an isolated forest anomaly detection model stored in local memory. This model calculates the path length of the segmentation within the data space containing the input feature vectors and outputs an anomaly severity score for the current inspection status. Severity score of abnormality The value range is normalized to the interval [0,1]. The edge computing processor 13 will calculate the value... Stored in an internal register and read the preset exception detection threshold. Perform a numerical comparison operation.

[0038] When the numerical comparison result is At this time, the edge computing processor 13 triggers a system-level interrupt to execute a cascaded adjustment process of hardware and network parameters. In the first stage of this adjustment process, the edge computing processor 13 sends a Universal Asynchronous Receiver / Transmitter (UART) control frame to the BeiDou positioning module 11 through the hardware abstraction layer (HAL) of the operating system. This control frame modifies the configuration of the phase-locked loop frequency divider register inside the BeiDou positioning module 11, thereby directly increasing the sampling frequency of the spatial coordinates output by the BeiDou positioning module 11.

[0039] Dynamically adjusted BeiDou sampling frequency The calculation formula is as follows: ; in, This indicates the adjusted current BeiDou sampling frequency, expressed in Hertz (Hz). This indicates the maximum sampling frequency limited by the hardware firmware of the Beidou positioning module 11; This indicates the base sampling frequency under normal conditions; This represents the preset frequency gain adjustment constant; This indicates the severity score of the current output anomaly. After receiving the control frame, the BeiDou positioning module 11 switches to... The frequency output coordinate data increases the density of trajectory sampling points around the point where the anomaly occurs.

[0040] In the second phase of the control process, the edge computing processor 13 sends a configuration update signal to the MQTT protocol daemon controlling the wireless communication module 14 via inter-process communication (IPC). After receiving the signal, the wireless communication module 14 modifies the long-connection keep-alive timer variable of the upper-layer application of the TCP / IP protocol stack, thereby shortening the time interval for sending heartbeat messages to the cloud server 20.

[0041] Dynamically adjusted heartbeat sending interval The calculation formula is as follows: ; in, This indicates the adjusted current heartbeat transmission interval, in seconds (s). This indicates the standard heartbeat transmission interval under normal conditions; This represents the minimum small-beat transmission interval set to avoid wireless channel congestion; This represents the heart rate attenuation coefficient; It is a natural constant.

[0042] Wireless communication module 14 applies a new heartbeat transmission interval Then, the frequency of interaction and connection confirmation messages with the cloud server 20 is increased. After the sensor module 12 continuously collects data and inputs it into the model, the calculated... Descending to When the above conditions are met, the edge computing processor 13 generates a recovery instruction, which is written to the register of the Beidou positioning module 11 and the MQTT protocol daemon, respectively, restoring the above two parameters to their initial values. and .

[0043] The real-time data transmission and processing method of the intelligent inspection system based on Beidou positioning provided by the present invention executes the network transmission state machine switching logic based on location look-ahead by combining the edge computing processor 13 with the spatial network signal quality topology library.

[0044] The edge computing processor 13 receives the three-dimensional spatial coordinates output by the Beidou positioning module 11. and instantaneous velocity vector Then, read the preset prediction time window variable. The edge computing processor 13 calculates the spatial coordinates of the inspection terminal 10 at a future predicted time based on a linear kinematics model.

[0045] The formula for calculating predicted spatial coordinates is as follows: ; in, Indicates the time window after prediction Predicted spatial coordinates of the rear inspection terminal 10; Indicates the current sampling time Three-dimensional spatial coordinates; Indicates the current sampling time The instantaneous velocity vector in three-dimensional space; This represents the preset prediction time window constant, in seconds (s). The edge computing processor 13 calculates the predicted spatial coordinates. The data is converted into a corresponding raster index and then located in the spatial network signal quality topology database stored in local memory. The edge computing processor 13 extracts the network quality feature values ​​stored in the raster index and uses them as the predicted network quality index. The edge computing processor 13 will... Compared with the blind zone determination threshold pre-stored in the system register Perform numerical comparisons.

[0046] When the numerical comparison result satisfies And the current real-time network quality index At this time, the edge computing processor 13 determines that the inspection terminal 10 is in a critical state before entering the network blind zone and triggers the pre-compensation transmission mechanism. The edge computing processor 13 sends a queue scheduling instruction to the underlying driver controlling the wireless communication module 14. In response to the queue scheduling instruction, the wireless communication module 14 clears the regular operation log in the transmission buffer and assigns the highest transmission priority to the operation status data containing abnormal markers and historical unacknowledged data frames. Within the remaining prediction time window... Inside, the wireless communication module 14 uses all available bandwidth to perform burst concurrent transmissions.

[0047] As the inspection terminal 10 continues to move, the edge computing processor 13 monitors that the current real-time network quality index meets the requirements. During this period, the edge computing processor 13 sends a connection suspension control primitive to the TCP / IP protocol stack of the wireless communication module 14. The wireless communication module 14 shuts down the transmit channel of the radio frequency front-end, stops sending MQTT heartbeat messages, and freezes the disconnection reconnection timer of the MQTT client. During this period, the edge computing processor 13 encapsulates the operating status data collected by the sensor module 12 and the positioning data collected by the Beidou positioning module 11 into a data structure, executes a lossless data compression algorithm, and writes the compressed data block along with a monotonically increasing sequence number and timestamp into the ring buffer configured in the local non-volatile memory of the inspection terminal 10.

[0048] While the wireless communication module 14 is in a suspended state, the edge computing processor 13 continuously performs spatial coordinate prediction at the base sampling frequency. The calculation and topology library addressing. When the predicted network quality index obtained by addressing satisfies At that time, the edge computing processor 13 determines that the inspection terminal 10 is about to leave the network blind spot.

[0049] Edge computing processor 13 sends a hardware wake-up signal to wireless communication module 14. Wireless communication module 14 restores power to the RF front-end and sends a lightweight network probe message to the base station. After the underlying link layer establishes a handshake, wireless communication module 14 initiates an MQTT connection request with the CleanSession flag set to False. After the connection is established, edge computing processor 13 reads compressed data blocks from the circular buffer, and based on the QoS level mechanism of the MQTT protocol, prioritizes extracting abnormally marked data blocks and uploads them to cloud server 20 through an encrypted channel. Subsequently, it sends cached blocks of normal operating status data in batches, strictly according to the order of timestamps and sequence numbers.

[0050] The real-time data transmission and processing method of the intelligent inspection system based on Beidou positioning provided by this invention executes a hardware-level self-organizing network multi-hop routing establishment program when the inspection terminal 10 is in a network blind zone and an abnormal event occurs.

[0051] The inspection terminal 10 further includes a short-range wireless communication module 15. The short-range wireless communication module 15 is electrically connected to the edge computing processor 13 via a system bus. The edge computing processor 13 reads the current anomaly severity score stored in its internal registers. and the current real-time network quality index The edge computing processor 13 performs logical AND judgment operations, and when the judgment result simultaneously satisfies... and At that time, a relay request hardware interrupt signal is generated. Among them, This is the preset critical anomaly threshold.

[0052] In response to the relay request hardware interrupt signal, the edge computing processor 13 extracts the current three-dimensional spatial coordinates output by the Beidou positioning module 11 and marks them as the coordinates where the anomaly occurred. The edge computing processor 13 will determine the coordinates of the anomaly. The device's unique identifier and abnormal operating status data collected by sensors are serialized according to a preset media access control layer protocol format to generate relay broadcast data frames. The short-range wireless communication module 15 receives the relay broadcast data frames and radiates radio frequency broadcast signals outward at a preset maximum transmit power.

[0053] The intelligent inspection system also includes multiple candidate inspection terminals 30 distributed around the inspection terminal 10. The candidate inspection terminals 30 have the same physical hardware architecture as the inspection terminal 10, including a candidate short-range wireless communication module, a candidate Beidou positioning module, a candidate wireless communication module, and a candidate edge computing processor.

[0054] The candidate short-range wireless communication module of the candidate inspection terminal 30 receives the radio frequency broadcast signal and demodulates and outputs the relay broadcast data frame to its internal candidate edge computing processor. The candidate edge computing processor then parses and extracts the coordinates of the anomaly from the relay broadcast data frame. Simultaneously, the candidate edge computing processor reads its current real-time three-dimensional spatial coordinates from its own candidate BeiDou positioning module. The candidate edge computing processor executes the Euclidean distance calculation algorithm and outputs the spatial Euclidean distance between the candidate inspection terminal 30 and the inspection terminal 10. .

[0055] The candidate edge computing processor obtains the network quality index of the communication link to the cloud server 20 at the current moment through its own candidate wireless communication module. The candidate edge computing processor calls its internal floating-point unit to calculate its own relay fitness weight as a data forwarding node based on a preset mathematical model.

[0056] The formula for calculating the relay fitness weight is as follows: ; in, Indicates the first The relay fitness weight calculation value of 30 candidate inspection terminals; This represents the preset positive gain coefficient for main network quality; Indicates the first Each candidate inspection terminal 30 has its own real-time three-dimensional spatial coordinates The network quality index of the communication link at the location; This represents the preset spatial distance penalty coefficient; Coordinates indicating where the anomaly occurred With real-time three-dimensional spatial coordinates The spatial Euclidean distance between them; Represented by natural constant A logarithmic function with base 0.

[0057] Each candidate inspection terminal 30 calculates its own relay fitness weight. Then, a contention time window based on a random backoff algorithm is initiated. Within the contention time window, each candidate short-range wireless communication module broadcasts its respective relay fitness weight. Each candidate edge computing processor collects all weight values ​​broadcast within the network and stores them in a local memory array.

[0058] After the competition window ends, the candidate edge computing processors perform a descending sorting algorithm on their local memory arrays. When a candidate inspection terminal 30 determines its own computation... When the value is the maximum value in the array, the candidate inspection terminal 30 automatically switches its working mode to the main relay node.

[0059] The primary relay node sends a unicast acknowledgment handshake frame to the inspection terminal 10 through its candidate short-range wireless communication module. After receiving the unicast acknowledgment handshake frame, the edge computing processor 13 of the inspection terminal 10 encapsulates the entire MQTT protocol message to be sent as the payload into the data segment of the short-range wireless communication protocol to generate a transparent data packet.

[0060] The short-range wireless communication module 15 continuously sends transparent data packets to the main relay node. The candidate edge computing processor of the main relay node unpacks the transparent data packets and extracts the underlying MQTT protocol messages from the payload. The main relay node controls its internal candidate wireless communication modules to upload and forward the extracted underlying MQTT protocol messages to the cloud server 20 through its established TCP / IP backbone network link.

[0061] See attached document Figure 1 and attached Figure 2 This invention provides a real-time data transmission and processing method for an intelligent inspection system based on BeiDou positioning. The cloud server 20 performs distributed streaming processing, structured storage, and data integration mapping of external systems on the received multi-source inspection data.

[0062] The cloud server 20 is equipped with an MQTT access gateway 21, a streaming computing unit 22, a distributed storage cluster 23, and a business logic adaptation layer 24. The MQTT access gateway 21 receives MQTT messages from the inspection terminal 10 or the master relay node through a Transport Layer Security (TLS) encrypted channel. The MQTT access gateway 21 performs integrity checks on the messages and forwards the unstructured payload data that passes the checks to the streaming computing unit 22 in real time.

[0063] The streaming computing unit 22 is configured with a high-concurrency processing thread pool. The streaming computing unit 22 first performs protocol decoding on the payload data, extracting information including sequence number, timestamp, and anomaly severity score. and three-dimensional spatial coordinates The original data stream. For data packets transmitted through the main relay node, the streaming computing unit 22 restores the unique physical address of the original inspection terminal by parsing the routing mark in the packet header, ensuring the consistency of data source tracing.

[0064] The streaming computing unit 22 integrates a BeiDou coordinate correction module. Due to ionospheric delay or multipath effects, the original coordinate data fluctuates. The BeiDou coordinate correction module uses a weighted moving average model to smooth the trajectory. (The calibrated spatial coordinates are shown in the image.) The calculation formula is as follows: ; in, Indicates the current time The three-dimensional spatial coordinate vector after correction and calibration; Indicates the preset sliding window length; This indicates that we are going back to the previous time. The original three-dimensional spatial coordinate vector of each sampling point; The corresponding weighting coefficients are given, and they satisfy the following conditions: The coordinate data after correction is marked as standard trajectory points for subsequent inspection path reproduction.

[0065] Streaming computing unit 22 scores based on anomaly severity Branching is performed on the data stream. When a branch is detected in the data stream... At that time, the streaming computing unit 22 immediately extracts the corresponding three-dimensional spatial coordinates and running status data, generates a high-priority alarm event frame, and pushes it to the real-time monitoring queue of the business logic adaptation layer 24.

[0066] The distributed storage cluster 23 employs a hot-and-cold separation storage strategy. The streaming computing unit 22 writes standard trajectory points and real-time acquired equipment pressure parameters, ambient temperature parameters, and other time-series characteristic data into the time-series database within the distributed storage cluster 23. The time-series database, through a sharded indexing mechanism, supports sub-second queries based on time range and equipment identifier. Simultaneously, the streaming computing unit 22 stores basic equipment information, system configuration parameters, and alarm event logs in the relational database within the distributed storage cluster 23, achieving structured and persistent data storage.

[0067] The business logic adaptation layer 24 integrates with the enterprise's existing management system through a pre-defined interface conversion protocol. The business logic adaptation layer 24 includes an API mapping module, which defines a data mapping matrix. This process converts the standard data fields defined in this system into the format required by third-party systems (such as ERP and MES). The mapping logic for data field conversion is represented as follows: ; in, This represents the original structured data vector stored in this system; This represents a preset heterogeneous system data transformation matrix, used to perform field name renaming, data type casting, and unit unification processing; This indicates the target data packet that conforms to the interface specifications of the third-party system after conversion.

[0068] Business logic adaptation layer 24 uses defined RESTful API interfaces or message middleware to transmit target data packets. The data is pushed to designated external enterprise systems. This integration mechanism enables abnormal inspection events to directly trigger the work order generation process in the enterprise's production planning management system, and automatically associates inspection trajectory data with the maintenance records of the equipment lifecycle management system, completing a closed-loop collaboration between inspection data and production management operations.

[0069] See attached document Figure 3 This invention provides a real-time data transmission and processing system 40 based on BeiDou positioning. The system 40 operates in a hardware environment comprised of an inspection terminal 10 and a cloud server 20. The BeiDou-based real-time data transmission and processing system 40 may include: a multi-dimensional data acquisition and topology construction module 41, an anomaly assessment and cascade control module 42, a blind zone adaptive transmission scheduling module 43, a multi-hop relay routing control module 44, and a cloud-based streaming integrated processing module 45.

[0070] The multidimensional data acquisition and topology construction module 41 is deployed within the edge computing processor 13 of the inspection terminal 10. The multidimensional data acquisition and topology construction module 41 sends a positioning request to the BeiDou positioning module 11 via the system's internal bus and simultaneously receives three-dimensional spatial coordinates. and instantaneous velocity vector The multidimensional data acquisition and topology construction module 41 sends a hardware trigger interrupt signal to the sensor module 12 to read the operating status data after analog-to-digital conversion.

[0071] The multi-dimensional data acquisition and topology construction module 41 sends a network status probe to the wireless communication module 14 to obtain the current transmission latency of the TCP / IP protocol stack. and packet loss rate The multi-dimensional data acquisition and topology construction module 41 calls the arithmetic logic unit of the central processing unit to calculate the network quality index. Network Quality Index with three-dimensional spatial coordinates Perform hash mapping and divide the local non-volatile memory into rasterized memory regions to generate and iteratively update the spatial network signal quality topology library.

[0072] The anomaly assessment and cascaded control module 42 is deployed within the edge computing processor 13 and exchanges data with the multidimensional data acquisition and topology construction module 41 through a shared memory mechanism. The anomaly assessment and cascaded control module 42 reads runtime status data, inputs it into the internally encapsulated isolated forest detection operator, and outputs an anomaly severity score. .

[0073] The anomaly assessment and cascading control module 42 is equipped with a hardware configuration distribution unit. This unit determines the severity of the anomaly. When the set system threshold is exceeded, the hardware configuration distribution unit generates a configuration containing the target sampling frequency. The control message is written to the hardware register of the Beidou positioning module 11 via the serial communication interface, thereby modifying the positioning data output cycle. Simultaneously, the hardware configuration distribution unit generates a message containing the target heartbeat transmission interval. The configuration command is sent to the MQTT daemon of the wireless communication module 14 via inter-process communication to reset the count value of the long connection keep-alive timer.

[0074] The blind zone adaptive transmission scheduling module 43 is deployed within the edge computing processor 13. The blind zone adaptive transmission scheduling module 43 reads the three-dimensional spatial coordinates output by the multi-dimensional data acquisition and topology construction module 41. and instantaneous velocity vector Call the linear algebra operation library to calculate the predicted spatial coordinates The blind zone adaptive transmission scheduling module 43 uses predicted spatial coordinates. Perform addressing lookup on the spatial network signal quality topology library to extract the predicted network quality index. .

[0075] The blind zone adaptive transmission scheduling module 43 has a built-in state machine. When the comparison instruction is output... Less than the blind zone threshold When the logical truth value is reached, the state machine switches to the burst transmission state, instructing the wireless communication module 14 to clear the buffer and send a message with an exception marker at full load. When the positioning determination is that the device has completely entered the blind zone, the state machine switches to the suspended state, cuts off the transmission power of the wireless communication module 14, and calls the LZ4 compression algorithm to perform lossless compression on the sensor and positioning data, writing the compressed structured byte stream into the local circular buffer queue in sequence. When the state machine determines that the device is about to leave the blind zone, it performs an RF front-end wake-up operation and establishes a breakpoint resume connection.

[0076] The multi-hop relay routing control module 44 is distributed and deployed at the edge of the inspection terminal 10 and the candidate inspection terminal 30. When the anomaly assessment and cascade control module 42 outputs an extreme value anomaly and the blind zone adaptive transmission scheduling module 43 determines that it is inside the blind zone, the multi-hop relay routing control module 44 in the inspection terminal 10 drives the short-range wireless communication module 15 to broadcast the coordinates of the anomaly occurrence. The request frame.

[0077] The multi-hop relay routing control module 44, distributed within the candidate inspection terminal 30, receives request frames and parses the coordinates of the anomaly occurrence. The multi-hop relay routing control module 44 within the candidate inspection terminal 30, combined with the real-time coordinates output by its own candidate BeiDou positioning module, calculates the spatial Euclidean distance and substitutes it into the relay fitness weight model to calculate the weight value. The multi-hop relay routing control module 44 within the candidate inspection terminal 30 executes an anti-collision contention broadcast protocol, compares the weight values ​​collected within the network, and selects the node with the largest value as the primary relay node. The primary relay node's multi-hop relay routing control module 44 establishes a local transparent tunnel, encapsulates the MQTT messages received from the short-range RF receiving antenna, and routes them to the wide area network transmitting antenna for transparent forwarding.

[0078] The cloud-based streaming integration processing module 45 is deployed in the distributed computing cluster of the cloud server 20. The cloud-based streaming integration processing module 45 receives encrypted data forwarded by the MQTT access gateway 21 and calls the decryption unit to generate a plaintext data stream. The cloud-based streaming integration processing module 45 deploys a BeiDou coordinate correction filter, uses a weighted moving average algorithm to correct coordinate drift, and outputs calibrated standard trajectory points.

[0079] The cloud-based streaming integration processing module 45 deploys a data splitting routing table. Based on this table, high-concurrency time-series runtime status data and standard trajectory points are directed to the time-series database shard nodes of the distributed storage cluster 23, while device attribute data and configuration parameters are directed to the relational database nodes. The cloud-based streaming integration processing module 45 embeds a business logic adaptation layer 24, which calls a pre-built heterogeneous system data transformation matrix. Perform field-level mapping and type conversion operations on the outbound data to generate a data payload that conforms to the standard protocol of the Enterprise Resource Planning (ERP) system, and push it to the external application server via the Ethernet interface.

[0080] Specific application examples: To further illustrate the execution logic and objective operating state of the technical solution of the present invention, the following description is based on a specific application example of a large-scale petrochemical complex.

[0081] The combined unit area contains dense metal pipelines and large storage tanks, and has multiple radio frequency signal shielding areas. Inspection equipment moves along a pre-set route. The Beidou positioning module 11 built into the inspection terminal 10 outputs three-dimensional spatial coordinates at a base sampling frequency of 1Hz, while the sensor module 12 simultaneously collects equipment pressure parameters and hydrogen sulfide gas concentration parameters. The edge computing processor 13 calculates the network quality index of the current coordinates in real time and writes the calculation results to local memory, generating the following... Figure 4 The scattered topological mapping matrix is ​​shown. Figure 4 The two-dimensional coordinate axes in the graph represent the geographical latitude and longitude offset, and the scatter points of different gray levels indicate the corresponding network quality index values, which are used to characterize the correspondence between spatial location and wireless signal strength.

[0082] During the operating period, the inspection terminal 10 moved to the pipeline valve area. The hydrogen sulfide gas concentration parameter output by the sensor module 12 increased from 10 ppm to 45 ppm over three consecutive sampling cycles. The edge computing processor 13 input this sequence data into the isolated forest anomaly detection model, and the calculated anomaly severity score was 0.82. This calculated value is greater than the preset anomaly judgment threshold of 0.60. In response to this comparison result, the edge computing processor 13 wrote a configuration message to the Beidou positioning module 11, increasing its sampling frequency from 1Hz to 5Hz; simultaneously, it sent a command to the wireless communication module 14 to shorten the MQTT protocol heartbeat transmission interval from 60s to 15s. (See attached...) Figure 5 -Appendix Figure 7 As shown in the figure, the system hardware parameters' step response state is recorded after the trigger time node, which objectively indicates that the device increased the trajectory sampling density and connection survival detection frequency in the abnormal area.

[0083] Subsequently, the edge computing processor 13 calculates the predicted spatial coordinates for the next 3 seconds based on the current three-dimensional spatial coordinates and instantaneous velocity vector. By querying the spatial network signal quality topology library, the extracted predicted network quality index is 0.15, which is less than the set blind zone threshold of 0.30. The edge computing processor 13 determines that the inspection terminal 10 is about to enter the densely shielded area of ​​metal pipelines. In the remaining 2.4 seconds before entering the shielded area, the wireless communication module 14 executes a queue scheduling instruction, sending the operating status data packet containing the hydrogen sulfide concentration exceeding the standard flag to the cloud server 20 in burst mode. The system's internal log shows that at the end of the prediction time window, the radio frequency front-end of the wireless communication module 14 automatically suspends, terminating the reconnection operation. The edge computing processor 13 compresses the subsequently collected high-frequency coordinate data and sensor data and writes them to the local circular buffer.

[0084] While the inspection terminal 10 is inside the shielded area, the anomaly severity score remains at 0.82, meeting the joint judgment criteria of extreme anomaly and network blind spot. The edge computing processor 13 triggers a hardware-level interrupt, driving the short-range wireless communication module 15 to radiate a relay broadcast data frame outward at the maximum allowable transmission power. The candidate inspection terminal 30, located outside the shielded area and 25 meters away from the inspection terminal 10, receives this broadcast data frame. The candidate inspection terminal 30, combining its current network quality index of 0.85 and the absolute spatial distance, substitutes it into the mathematical model to calculate its relay fitness weight as the maximum value within the local network, and switches to become the primary relay node.

[0085] The inspection terminal 10 sends the generated MQTT transparent data packet to the candidate inspection terminal 30 via a short-range wireless link. The candidate inspection terminal 30 then forwards the packet via its established TCP / IP link. Test-verified cloud reception logs show that the data packet with the abnormal marker was successfully delivered via the multi-hop relay routing control module 44 during the backbone network interruption of the inspection terminal 10. The difference between the sampling timestamp carried by the data packet and the cloud reception timestamp is 1.15 seconds. The streaming computing unit 22 of the cloud server 20 extracts the transparent data packet, routes it to the distributed storage cluster 23 for storage, and generates a standard data message through the API mapping of the business logic adaptation layer 24, pushing it to the external enterprise resource planning system. Objective operational records of this embodiment show that under complex electromagnetic shielding environments and abnormal operating conditions, the system sequentially triggers location prediction burst transmission, cascaded adjustment of underlying hardware parameters, and edge self-organizing network cooperation mechanisms, maintaining the connectivity of the data link and the transparent transmission of abnormal messages.

Claims

1. A real-time data transmission and processing method for an intelligent inspection system based on BeiDou positioning, characterized in that, Includes the following steps: Obtain the three-dimensional spatial coordinates and instantaneous velocity vector of the BeiDou positioning module, as well as the operational status data collected by the sensor module; The network quality index is calculated based on transmission delay and packet loss rate. The three-dimensional spatial coordinates are mapped to a discrete grid. The network quality feature values ​​are then smoothly updated using an exponentially weighted moving average algorithm to generate and iteratively update the spatial network signal quality topology library. The operational status data is subjected to anomaly detection, and an anomaly severity score is output. The sampling frequency and heartbeat transmission interval are dynamically adjusted based on the score. The predicted spatial coordinates are calculated by combining the three-dimensional spatial coordinates and the instantaneous velocity vector, and the corresponding predicted network quality index is searched in the topology library; When the predicted network quality index is less than the blind zone threshold, burst concurrent transmission is performed before entering the network blind zone. After entering the blind zone, the radio frequency front-end is paused, and the data and coordinates are compressed and cached locally without loss. If the score is greater than the critical anomaly threshold within the blind zone, the three-dimensional spatial coordinates are extracted as the anomaly occurrence coordinates and broadcast via radio frequency. The candidate terminal receiving the broadcast calculates the relay fitness weight based on the difference between the network quality index gain of the candidate terminal's current communication link and the Euclidean distance penalty to the coordinates where the anomaly occurred. The candidate terminal with the highest weight is switched to become the main relay node and forwards the transparent data packet to the cloud.

2. The real-time data transmission and processing method of the intelligent inspection system based on BeiDou positioning according to claim 1, characterized in that, The steps of performing anomaly detection on the operational status data and outputting anomaly severity scores, and dynamically adjusting the sampling frequency and heartbeat transmission interval based on the scores, specifically include: The temporal feature vector of the running status data is input into a locally preset isolated forest anomaly detection model, and the anomaly severity score is calculated and output, normalized to the interval between zero and one. When the severity score of the anomaly exceeds the preset anomaly judgment threshold, a control frame is written to the phase-locked loop frequency divider register of the Beidou positioning module to increase the sampling frequency, and the long connection keep-alive timer variable of the protocol stack is modified to shorten the protocol heartbeat transmission interval.

3. The real-time data transmission and processing method for the intelligent inspection system based on BeiDou positioning according to claim 2, characterized in that, The numerical calculation logic for dynamically adjusting the sampling frequency of the BeiDou positioning module and the protocol heartbeat transmission interval of the wireless communication module includes: The adjusted target sampling frequency value is determined by positively increasing the product term of the base sampling frequency, the frequency gain adjustment constant, and the severity score of the anomaly, and is limited by the maximum sampling frequency limit of the Beidou positioning module's hardware. The adjusted target value for the protocol heartbeat transmission interval is determined by attenuation through multiplying the standard heartbeat transmission interval by an exponential function that is the product of the abnormality severity score and the heartbeat attenuation coefficient raised to a negative power using a natural constant as the base, and is limited by the set minimum heartbeat transmission interval extreme value.

4. The real-time data transmission and processing method of the intelligent inspection system based on BeiDou positioning according to claim 1, characterized in that, The specific steps for performing bursty concurrent transmissions before entering a network blind spot include: When it is determined that the predicted network quality index is less than the blind zone threshold, and the current real-time network quality index is greater than or equal to the blind zone threshold, it is determined that the network is in a critical state before entering the network blind zone. The normal operation log in the transmission buffer of the wireless communication module is cleared, and the operation status data containing abnormal markers and historical unacknowledged data frames are given the highest transmission priority. The available bandwidth is occupied in burst mode and the data in the transmission buffer is transmitted within the remaining time window before the predicted spatial coordinates are reached.

5. The real-time data transmission and processing method for the intelligent inspection system based on BeiDou positioning according to claim 1, characterized in that, After entering the blind zone, when the inspection terminal leaves the blind zone, the process also includes a breakpoint resume recovery step: The predicted spatial coordinates are continuously calculated at a set basic sampling frequency, and the corresponding predicted network quality index is extracted from the spatial network signal quality topology library. When the predicted network quality index recovers to a level greater than or equal to the blind zone threshold, a hardware wake-up level signal is sent to the wireless communication module to restore power to the radio frequency front-end and re-initiate the network connection request; Read the lossless compressed data blocks cached in the local non-volatile memory, prioritize extracting data blocks with abnormal markers for uploading, and then send regular cached data blocks in batches according to the sampling timestamp and monotonically increasing sequence number.

6. The real-time data transmission and processing method for the intelligent inspection system based on BeiDou positioning according to claim 1, characterized in that, It also includes a coordinate correction step based on a cloud server: The cloud server receives the trajectory data stream composed of the three-dimensional spatial coordinates; The weighted moving average model is invoked to perform spatial correction and smoothing on the original three-dimensional spatial coordinates, and the calibrated standard trajectory points are output. The calibrated 3D spatial coordinates at the current time are calculated by summing the products of the 3D spatial coordinates at the current time and the 3D spatial coordinates of multiple consecutive sampling points back to the past with their corresponding weighting coefficients.

7. The real-time data transmission and processing method for the intelligent inspection system based on BeiDou positioning according to claim 1, characterized in that, It also includes data integration steps based on cloud servers: The parsed running status data and the three-dimensional spatial coordinates are written into the time-series database shard nodes of the distributed storage cluster through a distributed streaming computing unit; The heterogeneous system data transformation matrix is ​​defined through the business logic adaptation layer. The original structured data vector inside the system is multiplied with the heterogeneous system data transformation matrix. Field name mapping and renaming, data type forced conversion and unit unification calculation are performed to generate target data messages that conform to the interface specifications of external third-party systems and push them.

Images