Network topology self-healing method and system of distributed diving monitoring system

Abstract

The invention discloses a network topology self-healing method and system for a distributed diving monitoring system, and relates to the field of network communication, and the method comprises the steps: obtaining real-time mechanical parameters; determining a link quality change rule of the distributed sensor nodes of the heave compensation device under different phases of the heave period; calculating a physical interference coefficient of a distributed sensor node, and identifying a high-impedance phase interval and a low-impedance phase interval in a heaving period; in the high impedance state phase interval, executing a phase locking route freezing strategy; executing an unfreezing strategy in the low resistance state phase interval; monitoring a real-time gas pressure value of the energy accumulator; and when the real-time gas pressure value is reduced and the reduction amplitude exceeds a preset abnormal pressure threshold value, judging that the heave compensation device has a mechanical fault, and executing a top-speed self-healing strategy. The problems of communication resource waste and low data transmission efficiency can be effectively solved.

Description

TECHNICAL FIELD

[0001] Embodiments of the present application relate to the field of network communication, and particularly relate to a network topology self-healing method and system of a distributed diving monitoring system. BACKGROUND

[0002] In the field of deep-sea diving operations and ocean engineering, distributed diving monitoring systems are widely used to monitor the state of diving equipment, environmental parameters, and physiological indicators of operating personnel in real time. The system usually adopts a wireless sensor network architecture, which is composed of multiple distributed sensor nodes that collect and transmit data through a self-organizing network. In actual applications, the mother ship will undergo periodic heave motion due to the continuous action of sea waves. In order to compensate for the impact of this motion on diving operations, a heave compensation device is usually equipped. The device is mainly composed of mechanical components such as hydraulic cylinders and accumulators, and can offset the vertical displacement of the mother ship through the extension and retraction of the hydraulic cylinder, thereby maintaining the stable position of the diving equipment relative to the seabed.

[0003] In traditional wireless sensor networks, the self-healing mechanism of network topology mainly relies on heartbeat packet loss detection and received signal strength indication threshold determination. When the system detects that a node continuously loses heartbeat packets or the signal strength is lower than the preset threshold, it will determine that the node is failed or the link is interrupted, and then trigger the topology reconstruction process to find a new communication path through the routing discovery protocol. This self-healing mechanism performs well in static or slowly changing network environments and can effectively deal with node failures, battery depletion, or permanent communication obstacles. However, this threshold-triggered self-healing strategy assumes that changes in network topology are caused by real hardware failures or permanent environmental changes, and does not consider the regular fluctuations in link quality caused by periodic physical structure movements.

[0004] In the heave compensation scenario, the mother ship will undergo periodic heave motion with an amplitude of up to 2.4 meters and a period of about 6 seconds. The sensor nodes attached to the heave compensation device will undergo dramatic position changes with the extension and retraction of the hydraulic cylinder. When the hydraulic cylinder contracts, the node may be blocked by the metal structure, causing signal attenuation; when the hydraulic cylinder extends, the node position changes and the signal strength recovers. This periodic physical movement will cause the communication link quality to present regular good-bad alternation, forming a false failure phenomenon. If the traditional self-healing mechanism is used, the system will misjudge this periodic signal fluctuation as a real node failure, triggering topology reconstruction every few seconds. This causes the topology to repeatedly switch between node drop, reconstruction, node online, and again drop, consuming a large amount of routing discovery messages and topology maintenance overhead, which consumes network bandwidth and node energy, making the system unable to normally transmit effective monitoring data.

[0005] In summary, the existing network topology self-healing method cannot distinguish between the link quality false degradation caused by the periodic motion of the heave compensation device and the real node failure or link interruption, resulting in frequent unnecessary reconstruction of the network topology, causing waste of communication resources and low data transmission efficiency. SUMMARY

[0006] The embodiment of the present application provides a network topology self-healing method and system of a distributed diving monitoring system, which is used to solve the problems of waste of communication resources and low data transmission efficiency.

[0007] To achieve the above-mentioned purpose, the embodiment of the present application adopts the following technical scheme: In a first aspect, a network topology self-healing method of a distributed diving monitoring system is provided, which is applied to a distributed diving monitoring system including a heave compensation device and a distributed sensor node, the heave compensation device including a hydraulic cylinder and an accumulator, and the method includes: obtaining real-time mechanical parameters of the heave compensation device, the real-time mechanical parameters including real-time stroke data of the hydraulic cylinder, real-time extension and contraction speed of the hydraulic cylinder, and real-time gas pressure value of the accumulator; determining the link quality change rule of the distributed sensor node at different phases of the heave cycle of the heave compensation device according to the real-time stroke data and the real-time extension and contraction speed; calculating the physical interference coefficient of the distributed sensor node based on the real-time extension and contraction speed and the real-time stroke data, and identifying the high resistance phase interval and the low resistance phase interval in the heave cycle according to the physical interference coefficient, wherein the high resistance phase interval corresponds to the communication link quality false degradation period, and the low resistance phase interval corresponds to the communication link quality normal period; in the high resistance phase interval, a phase-locked routing freezing strategy is executed to suspend sending all routing discovery messages, ignore routing failure alarms, and locally cache data to be transmitted; in the low resistance phase interval, an unfreezing strategy is executed to remove the routing freezing state and concentrate on burst transmission of locally cached data; in response to a network topology update request, monitoring the real-time gas pressure value of the accumulator; when the real-time gas pressure value decreases and the decrease amplitude exceeds a preset pressure abnormal threshold, determining that the heave compensation device has a mechanical failure, and executing an ultra-fast self-healing strategy to cancel all routing freezing restrictions, reduce the packet loss detection threshold, and trigger network topology reconstruction.

[0008] In a possible implementation manner of the first aspect, the link quality change rule of the distributed sensor node at different phases of the heave cycle of the heave compensation device is determined according to the real-time stroke data and the real-time extension and contraction speed, including: According to the periodic change of the real-time expansion speed, a heave period of the heave compensation device is determined; Within the heave period, signal strength data of the distributed sensor nodes is collected; A time sequence correlation model between the real-time travel data, the real-time expansion speed and the signal strength data is constructed to determine a link quality change rule of the distributed sensor nodes at different phases of the heave period of the heave compensation device.

[0009] In a possible implementation manner of the first aspect, the physical interference coefficient of the distributed sensor nodes is calculated based on the real-time expansion speed and the real-time travel data, and the calculation includes: The absolute value of the real-time expansion speed is compared with a preset speed threshold value; When the absolute value of the real-time expansion speed is greater than the speed threshold value, it is determined that the current is in a high-speed motion state, and a frequency shift influence factor is calculated; According to the real-time travel data, it is determined whether the current position of the hydraulic cylinder is in a metal shielding interval; When the current position of the hydraulic cylinder is in the metal shielding interval, a cutting influence factor is calculated; The frequency shift influence factor and the cutting influence factor are analyzed in coupling degree, and a synergistic action coefficient of the two is calculated; According to the synergistic action coefficient, it is determined whether the frequency shift influence factor and the cutting influence factor are in an activated state at the same time; When the frequency shift influence factor and the cutting influence factor are in the activated state at the same time, the product of the two is calculated as a composite interference strength; When only one influence factor is in the activated state, the value of the activated influence factor is taken as a single interference strength; The composite interference strength or the single interference strength is normalized by a nonlinear mapping function to obtain the physical interference coefficient.

[0010] In a possible implementation manner of the first aspect, according to the physical interference coefficient, a high-resistance phase interval and a low-resistance phase interval in the heave period are identified, and the identification includes: The heave period is divided into a plurality of phase segments, and the average value of the physical interference coefficient in each phase segment is calculated; The average value of the physical interference coefficient is compared with a preset interference threshold value; When the average value of the physical interference coefficient is higher than the preset interference threshold value, the corresponding phase segment is marked as a high-resistance phase interval; When the average value of the physical interference coefficient is lower than or equal to the preset interference threshold value, the corresponding phase segment is marked as a low-resistance phase interval.

[0011] In a possible implementation manner of the first aspect, a phase-locked routing freezing strategy is executed, and the execution includes: According to the heave period and the starting phase of the high resistance state phase interval, the opening time and the closing time of the route freezing window are calculated; At the opening time of the route freezing window, a route table locking flag is set in the network routing protocol layer; Sending all packets and link state advertisement packets is suspended; All received route failure alarms are marked as a to-be-verified state, and a topology reconstruction process is not triggered; The current data packet to be transmitted is stored in a local cache queue; At the closing time of the route freezing window, the route table locking flag is cleared.

[0012] In another possible implementation of the first aspect, the unfreezing strategy is executed, including: According to the heave period and the starting phase of the low resistance state phase interval, the opening time and the duration of the data transmission window are calculated; At the opening time of the data transmission window, the route table locking flag is cleared, and the normal operation of the routing protocol is resumed; According to a priority order, the data packet to be transmitted is read from the local cache queue, and burst transmission is performed; A route discovery packet and a topology update request sent by a neighbor node are received and processed; The route failure alarm marked as the to-be-verified state in the high resistance state phase interval is verified again, and if communication cannot be restored in the low resistance state phase interval, it is determined as a real failure.

[0013] In another possible implementation of the first aspect, after monitoring the real-time gas pressure value of the accumulator, the method further includes: According to the real-time gas pressure value, the equivalent stiffness of the heave compensation device is calculated; The equivalent stiffness is normalized to obtain a gas-liquid stiffness coefficient; An inverse proportional mapping relationship between the gas-liquid stiffness coefficient and a network self-healing sensitivity threshold is established; When the gas-liquid stiffness coefficient increases, the packet loss rate tolerance threshold is increased, and the sensitivity of network self-healing is reduced; When the gas-liquid stiffness coefficient decreases, the packet loss rate tolerance threshold is reduced, and the sensitivity of network self-healing is increased.

[0014] In another possible implementation of the first aspect, when the real-time gas pressure value decreases and the decrease amplitude exceeds a preset pressure abnormality threshold, it is determined that the heave compensation device has a mechanical failure, and an ultra-speed self-healing strategy is executed to cancel all route freezing restrictions, reduce a packet loss detection threshold, and trigger network topology reconstruction, including: The change rate of the real-time gas pressure value of the accumulator is determined; When the real-time gas pressure value decreases by more than the pressure anomaly threshold within the preset time window, a mechanical failure alarm is triggered; Clear all route freeze window settings, cancel phase-locked route freeze strategy; Adjust the packet loss detection threshold to the preset minimum value, so that the continuous loss of a single data packet is determined as node failure; Start the network topology reconstruction process with the highest priority, and search for redundant communication paths; Based on the redundant communication paths, the mechanical failure alarm information is transmitted through the reconstructed network topology.

[0015] In a second aspect, the application provides a distributed diving monitoring system, comprising: The heave compensation device comprises a hydraulic cylinder and an accumulator, the hydraulic cylinder is connected with the accumulator through a hydraulic pipeline, and the accumulator is used to provide pressure compensation for the hydraulic cylinder; The distributed sensor node is installed on the heave compensation device and is used to collect real-time mechanical parameters of the heave compensation device, wherein the real-time mechanical parameters include real-time stroke data of the hydraulic cylinder, real-time extension and contraction speed of the hydraulic cylinder, and real-time gas pressure value of the accumulator; The network gateway node is wirelessly connected with the distributed sensor node and is used to execute the network topology self-healing method of the distributed diving monitoring system.

[0016] In a third aspect, the application provides a machine-readable storage medium, which stores instructions for causing a machine to execute the network topology self-healing method of the distributed diving monitoring system.

[0017] By the technical solution, the intelligent identification and adaptive response to the periodic physical interference are realized by deeply fusing the mechanical motion state of the heave compensation device and the network topology self-healing strategy. The phase-locked routing freezing mechanism effectively distinguishes the false failure caused by the mechanical motion from the real node failure, avoids the problem that the topology reconstruction is triggered once every several seconds in the traditional method, greatly reduces the routing control message overhead, and reduces the node energy consumption. The time division multiplexing communication mode fully utilizes the good link quality in the low resistance state phase interval, realizes efficient data transmission through burst transmission, effectively improves the data transmission success rate, and reduces the transmission delay. The mechanical failure detection mechanism based on the gas pressure of the accumulator realizes cross-layer collaborative perception, and when the real mechanical failure occurs, the network topology reconstruction can be completed in a short time, thereby ensuring the reliability and safety of the monitoring system. The adaptive network self-healing sensitivity adjustment strategy dynamically optimizes according to the system working state, reduces the sensitivity to avoid misjudgment in stable operation, and improves the sensitivity to quickly respond in abnormal state, thereby realizing the best balance between stability and response speed. The overall scheme significantly improves the communication stability and reliability of the distributed submersible monitoring system in the complex marine environment, and provides a strong guarantee for the safety of deep sea operation.

[0018] Other features and advantages of the embodiments of the present application will be described in detail in the subsequent specific embodiments part. BRIEF DESCRIPTION OF DRAWINGS

[0019] Figure 1 A flowchart of a network topology self-healing method of a distributed submersible monitoring system provided by the embodiments of the present application is shown in the figure. Figure 2 A structure diagram of a distributed submersible monitoring system provided by the embodiments of the present application is shown in the figure. DETAILED DESCRIPTION

[0020] In order to make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme in the embodiments of the present application will be described clearly and completely in conjunction with the drawings in the embodiments of the present application. It should be understood that the specific embodiments described herein are only used to illustrate and explain the embodiments of the present application, and are not used to limit the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0021] It should be noted that if the embodiments of the present application involve directional indications (such as up, down, left, right, front, back, etc.), the directional indications are only used to explain the relative position relationship, motion condition, etc. between the components in a certain posture (as shown in the drawings), and if the certain posture changes, the directional indications also change accordingly.

[0022] In addition, if the description of "first", "second" and the like is involved in the embodiments of the present application, the description of "first", "second" and the like is only for the purpose of description, and cannot be understood as indicating or implying the relative importance of the indicated technical features or implicitly indicating the number of the indicated technical features. Therefore, the features limited by "first", "second" can be explicitly or implicitly included at least one of the features. In addition, the technical solutions of various embodiments can be combined with each other, but it must be based on the realization of ordinary skilled in the art, when the combination of technical solutions appears contradictory or cannot be realized, it should be considered that the combination of technical solutions does not exist, nor in the protection scope claimed by the present application.

[0023] Figure 1 The flowchart of the network topology self-healing method of the distributed diving monitoring system according to the embodiments of the present application is schematically shown. As shown in Figure 1 The network topology self-healing method of the distributed diving monitoring system according to the embodiments of the present application is provided, which is applied to the distributed diving monitoring system, the distributed diving monitoring system includes a heave compensation device and a distributed sensor node, the heave compensation device includes a hydraulic cylinder and an accumulator, and the method can include the following steps.

[0024] S110, acquiring real-time mechanical parameters of the heave compensation device, the real-time mechanical parameters including real-time stroke data of the hydraulic cylinder, real-time expansion and contraction speed of the hydraulic cylinder and real-time gas pressure value of the accumulator; S120, determining the link quality change rule of the distributed sensor node at different phases of the heave cycle according to the real-time stroke data and the real-time expansion and contraction speed; S130, calculating the physical interference coefficient of the distributed sensor node based on the real-time expansion and contraction speed and the real-time stroke data, and identifying the high resistance phase interval and the low resistance phase interval in the heave cycle according to the physical interference coefficient, wherein the high resistance phase interval corresponds to the false degradation period of the communication link quality, and the low resistance phase interval corresponds to the normal period of the communication link quality; S140, in the high resistance phase interval, executing a phase-locked routing freezing strategy to suspend sending all routing discovery messages, ignoring routing failure alarm, and locally caching data to be transmitted; S150, in the low resistance phase interval, executing a thawing strategy to remove the routing freezing state, and concentrating burst transmission of locally cached data; S160, in response to a network topology update request, monitoring the real-time gas pressure value of the accumulator; S170, when the real-time gas pressure value decreases and the decrease amplitude exceeds a preset pressure abnormal threshold, determining that the heave compensation device has a mechanical failure, executing an ultra-fast self-healing strategy to cancel all routing freezing restrictions, reduce the packet loss detection threshold, and trigger network topology reconstruction.

[0025] In the distributed diving monitoring system, the real-time stroke data of the hydraulic cylinder is obtained through the displacement sensor installed on the hydraulic cylinder piston rod. The sensor can use magnetostrictive or pull rope displacement sensor, which can measure the extension and retraction position of the hydraulic cylinder with millimeter level precision. The displacement sensor converts the physical displacement into an electrical signal, which is amplified and filtered by a signal conditioning circuit, and then converted into a digital signal by an analog-to-digital converter. The sampling frequency is usually set to 50Hz to ensure that the rapid movement of the hydraulic cylinder can be captured. The real-time extension and retraction speed of the hydraulic cylinder is obtained by numerical differentiation calculation on the continuously collected stroke data, and five-point central difference method is used to reduce the noise influence, and the calculation formula is that the current time speed is equal to the stroke difference between the previous and next sampling points divided by twice the time interval. The real-time gas pressure value of the accumulator is measured by the pressure sensor installed at the gas chamber end of the accumulator. The sensor uses piezoresistive or capacitive pressure sensor. The analog signal output by the pressure sensor is also subjected to signal conditioning and analog-to-digital conversion, and data acquisition is performed at a sampling frequency of 20Hz.

[0026] All sensor data is transmitted to the data acquisition module through CAN bus or RS485 bus. The data acquisition module marks the original data with time stamp and packs the data to form a standardized mechanical parameter data frame. The data frame contains four fields of time stamp, stroke data, extension and retraction speed and gas pressure value, each field is stored in 32-bit floating point format to ensure data accuracy. By collecting these mechanical parameters in real time, accurate physical state information is provided for subsequent link quality analysis and topology self-healing decision, so that the network layer can perceive the motion state of the mechanical layer, and cross-layer collaborative optimization is realized.

[0027] To determine the link quality variation law of the heave compensation device at different phases of the heave cycle, a mapping relationship between mechanical motion and communication quality needs to be established. First, the heave cycle is identified by analyzing the periodic variation of the real-time extension and retraction speed of the hydraulic cylinder. The specific method is to perform fast Fourier transform on the extension and retraction speed to identify the main frequency component in the frequency domain. The period corresponding to the main frequency is the heave cycle, which is usually between 4-8 seconds.

[0028] After identifying the heave period, a complete period is divided into 16 phase segments, each of which accounts for about 6.25% of the period. In each phase segment, the received signal strength indication value reflecting the degree of attenuation of wireless signals in the propagation process is continuously collected by the distributed sensor nodes. The signal strength data collection adopts a sliding window mechanism, and the window length is set to 0.5 seconds. The mean and variance of the signal strength are calculated within the window, the mean reflects the average link quality of the phase segment, and the variance reflects the stability of the link quality. At the same time, the hydraulic cylinder stroke position and extension speed corresponding to the phase segment are recorded, and a three-dimensional data set is established, containing four dimensions of phase angle, stroke position, extension speed and signal strength. Through statistical analysis of the data of multiple heave periods, the least squares method is used to fit the functional relationship between signal strength and stroke position. Experimental data show that when the hydraulic cylinder is in a fully retracted state, the sensor node is blocked by the metal structure, and the signal strength decreases by 15-25 dBm; when the hydraulic cylinder is in the middle position, the signal strength is relatively stable; when the hydraulic cylinder is fully extended, the change in node position causes the signal path to change, and the signal strength fluctuates by 5-10 dBm. In addition, the extension speed also affects the signal quality, when the speed exceeds 0.3 meters per second, the Doppler shift effect causes signal demodulation difficulty, and the bit error rate rises significantly. By establishing this timing correlation model, the link quality state when the hydraulic cylinder is in a specific phase at any future time can be predicted, providing a theoretical basis for the phase locking strategy.

[0029] The calculation of the physical interference coefficient needs to consider the combined effect of the hydraulic cylinder movement speed and position on the communication link. First, determine whether it is currently in a high-speed motion state, compare the absolute value of the real-time extension speed with the speed threshold of 0.25 meters per second, if it exceeds the threshold, it is considered to be in a high-speed motion state. In the high-speed motion state, the frequency shift influence factor is calculated, which reflects the influence of the Doppler effect on signal demodulation. The speed is normalized to the 0-1 interval, and when the speed reaches 0.5 meters per second, the frequency shift influence factor approaches 1. According to the real-time stroke data, determine whether the hydraulic cylinder is in the metal shielding interval, and determine the shielding interval through pre-calibration, which is usually located in the front 20% and rear 15% of the stroke range. In these intervals, there is a metal structure blocking between the sensor node and the gateway node.

[0030] When it is determined to be in the shielding interval, the cutting influence factor is calculated, which reflects the attenuation effect of metal shielding on signal propagation. The cutting influence factor is calculated according to the depth position of the node in the shielding interval, the closer to the center of the shielding interval, the larger the cutting influence factor, and the maximum value is 1. Then analyze the synergistic effect of the frequency shift influence factor and the cutting influence factor, when both are in an active state, i.e. the frequency shift influence factor is greater than 0.3 and the cutting influence factor is greater than 0.2, the two interference mechanisms produce a superposition effect, at this time the product of the two is calculated and a synergistic gain coefficient of 1.5 times is applied, to get the combined interference strength.

[0031] If only one factor is activated, the factor value is directly used as the single interference strength. Finally, the interference strength is nonlinearly mapped by the hyperbolic tangent function to compress the value into the interval of 0-1, and the final physical interference coefficient is obtained. This coefficient can quantitatively represent the comprehensive interference degree of mechanical motion on the communication link at the current moment. When identifying the high and low resistance state phase intervals based on the physical interference coefficient, the time average of the physical interference coefficient in each phase segment is calculated, and the average value is compared with the interference threshold 0.4. The phase segment higher than the threshold is marked as the high resistance state phase interval, indicating that the communication link quality is severely affected by mechanical motion in this period. The phase segment lower than or equal to the threshold is marked as the low resistance state phase interval, indicating that the communication condition is relatively good.

[0032] Executing the phase-locked routing freeze strategy in the high resistance state phase interval can avoid frequent topology reconstruction. According to the identified heave period and the starting phase angle of the high resistance state phase interval, combined with the current system time, the accurate opening time of the next routing freeze window is calculated. When the system time reaches the opening time, a routing table lock flag is set in the routing layer of the network protocol stack, which prevents any modification operation of the routing table. At the same time, the sending of all active routing discovery packets, including route request packets and link state advertisement packets, is suspended. These packets are used to detect network topology and maintain routing information under normal circumstances, but sending them in the high resistance state phase interval will fail due to poor link quality, causing unnecessary network overhead. For routing failure alarms received from other nodes during this period, the topology reconstruction process is not triggered immediately, but these alarms are marked as pending verification status and recorded in a temporary queue, because these alarms are likely to be false failures caused by mechanical motion.

[0033] For the data packets generated by the application layer, they are not discarded but stored in a local buffer queue. The queue uses a circular buffer structure, and its capacity is configured according to the node memory size, usually set to be able to buffer the data generated in a complete heave period. The data packets in the buffer queue are sorted according to priority and timestamp, with high-priority monitoring data such as physiological parameter abnormal alarms at the front of the queue. The closing time of the routing freeze window is calculated according to the duration of the high resistance state phase interval, and is usually closed when the hydraulic cylinder moves to the starting position of the low resistance state phase interval. Through this phase-locked mechanism, routing maintenance activities are actively suspended during periods of inevitable link quality degradation, avoiding frequent topology reconstruction triggered by false failures in traditional methods, significantly reducing control packet overhead and node energy consumption, while ensuring data non-loss through data buffering mechanism.

[0034] The unfreezing strategy is executed in the low resistance phase interval to restore normal communication and transmit the buffered data. According to the heave period and the starting phase angle of the low resistance phase interval, the opening time of the data transmission window is calculated, which corresponds to the phase position of the hydraulic cylinder moving to the good link quality. At the same time, the effective duration of the transmission window is determined according to the duration of the low resistance phase interval, which is usually 2-4 seconds.

[0035] At the opening time of the transmission window, the routing table lock flag is cleared, the normal operation of the routing protocol is restored, and the routing table update and route discovery are allowed. Immediately read the data packets to be transmitted from the local buffer queue according to the priority order, and send them in burst transmission mode. The interval between data packets is shortened to the minimum frame interval during burst transmission, making full use of the good link quality in the low resistance phase interval and improving channel utilization. For high priority data packets, a retransmission mechanism is used to ensure reliable transmission, with a maximum of two retransmissions if no acknowledgement is received within the transmission window. At the same time, start receiving and processing route discovery messages and topology update requests from neighbor nodes, update the local routing table, and maintain the accuracy of the network topology.

[0036] For route failure alarms marked as to-be-verified state in the high resistance phase interval, secondary verification is performed in the low resistance phase interval. The specific method is to send probe messages to the nodes involved in the alarm. If no response is received after sending three consecutive probe messages within the transmission window, it is determined that the node has indeed failed, and the topology reconstruction process for the node is triggered. If a response is received, it is confirmed as a false failure, the alarm record is cleared, and the original route remains unchanged. Through this unfreezing strategy, data transmission and topology maintenance are completed in a concentrated manner during the period of good link quality, achieving a time-division multiplexing communication mode that ensures data transmission efficiency and avoids unnecessary topology reconstruction, allowing the network to operate stably in a periodic mechanical motion environment.

[0037] The real-time gas pressure value of the accumulator is monitored in response to network topology update requests, which is used to determine whether the heave compensation device is in normal working condition. As an energy storage element of the hydraulic system, the gas pressure of the accumulator directly reflects the compensation ability and mechanical health status of the system. In normal working condition, the gas pressure of the accumulator should be maintained within the design pressure range, usually 12-18 MPa, and the pressure value fluctuates slightly within this range with the extension and retraction of the hydraulic cylinder, with a fluctuation amplitude generally not exceeding 2 MPa. The trend of pressure value change is continuously monitored, and the pressure change rate is calculated using the sliding window method with a window length of 30 seconds. The slope of the fitted straight line is the pressure change rate. At the same time, the standard deviation of the pressure value is calculated to evaluate the severity of pressure fluctuation.

[0038] Based on the real-time gas pressure value, the equivalent stiffness of the heave compensation device can be calculated, which reflects the ability of the system to resist external load changes. The calculation takes into account the gas polytropic index of the accumulator and the effective action area of the hydraulic cylinder. Normalize the equivalent stiffness to get the gas-liquid stiffness coefficient, which varies between 0 and 1, and is about 0.6-0.8 under normal working conditions. Establish an inverse proportional mapping relationship between the gas-liquid stiffness coefficient and the network self-healing sensitivity threshold. When the gas-liquid stiffness coefficient is large, it means that the system has strong compensation ability and the mechanical movement is stable. At this time, the packet loss tolerance threshold can be increased to reduce the sensitivity of network self-healing, avoiding overreaction to normal periodic link fluctuations. Conversely, when the gas-liquid stiffness coefficient decreases, it means that the system's compensation ability has decreased and abnormal movement patterns may occur. At this time, the packet loss tolerance threshold should be reduced to increase the sensitivity of network self-healing, and real faults can be discovered and responded to in a timely manner. Through this adaptive adjustment mechanism, the network topology self-healing strategy can be dynamically optimized according to the real-time state of the mechanical system, ensuring system stability while improving fault response speed.

[0039] When the real-time gas pressure value of the accumulator is detected to be abnormally low, it is determined that the heave compensation device has a mechanical failure and the ultra-fast self-healing strategy is executed. The pressure anomaly is determined based on two conditions: one is that the absolute value of the pressure is below the safe lower limit, usually set to 10 MPa; the second is that the pressure drop rate exceeds the abnormal threshold, which is calculated by calculating the pressure change amplitude within a preset time window. If the pressure drops by more than 3 MPa within 60 seconds, it is considered abnormal.

[0040] Pressure anomalies are usually caused by serious mechanical failures such as accumulator gas leaks, hydraulic oil leaks, or hydraulic line ruptures. These failures can cause the heave compensation device to fail, causing the hydraulic cylinder's movement pattern to change dramatically, and the spatial position and movement trajectory of the sensor node to deviate completely from the normal state. The original phase locking strategy is no longer applicable. Once a mechanical failure alarm is triggered, all route freeze window settings are immediately cleared, the phase-locked route freeze strategy is canceled, and the continuous operation state of the routing protocol is restored. The packet loss detection threshold is adjusted from the normal value of 5 consecutive lost data packets to the minimum value of 1 consecutive lost data packet, making the node failure determination extremely sensitive and able to quickly identify real communication interruptions caused by mechanical failures.

[0041] The network topology reconfiguration process with the highest priority is started, which uses a flooding routing discovery algorithm. The gateway node broadcasts a routing request message to all neighboring nodes. Each node receives the request and forwards and replies to the routing response message. Through multiple rounds of interaction, a complete network topology graph is established. During the topology reconstruction process, redundant communication paths are searched, and paths with fewer hops and stable link quality are preferentially selected as the main path. At the same time, a backup path is established to improve reliability. After the reconstruction is completed, the mechanical fault alarm information is transmitted to the monitoring center through the newly established network topology. The alarm information includes detailed parameters such as fault type, pressure data, and timestamp, providing a basis for fault diagnosis and emergency handling. The execution time of the ultra-fast self-healing strategy is usually within 2-3 seconds, ensuring that network communication can quickly recover after a mechanical failure occurs, and ensuring the continuous operation of the monitoring system and the timely transmission of safety alarms.

[0042] In this embodiment, the mechanical motion state of the heave compensation device is deeply integrated with the network topology self-healing strategy, achieving intelligent identification and adaptive response to periodic physical disturbances. The phase-locked routing freezing mechanism effectively distinguishes between false failures caused by mechanical motion and real node failures, avoiding the problem of triggering topology reconstruction every few seconds in traditional methods, significantly reducing routing control message overhead, and reducing node energy consumption. The time-division multiplexing communication mode fully utilizes the good link quality in the low resistance phase interval, achieving efficient data transmission through burst transmission, effectively improving data transmission success rate and reducing transmission delay. The mechanical fault detection mechanism based on accumulator gas pressure realizes cross-layer collaborative perception. When a real mechanical failure occurs, the network topology can be reconstructed in a short time, ensuring the reliability and safety of the monitoring system. The adaptive network self-healing sensitivity adjustment strategy dynamically optimizes according to the system working state, reducing sensitivity to avoid misjudgment during stable operation, and increasing sensitivity to quickly respond in abnormal state, achieving the best balance between stability and response speed. The overall scheme significantly improves the communication stability and reliability of the distributed submersible monitoring system in complex marine environments, providing a strong guarantee for deep sea operation safety.

[0043] In one embodiment of the present embodiment, the link quality variation law of the distributed sensor nodes of the heave compensation device at different phases of the heave cycle is determined according to real-time travel data and real-time stretching speed, including the following steps: S210, determining the heave cycle of the heave compensation device according to the periodic variation of the real-time stretching speed; S220, collecting signal strength data of the distributed sensor nodes within the heave cycle; S230, constructing a time sequence correlation model between real-time travel data, real-time stretching speed and signal strength data to determine the link quality variation law of the distributed sensor nodes of the heave compensation device at different phases of the heave cycle.

[0044] The heave period of the heave compensation device needs to be determined by periodically analyzing the real-time extension and contraction speed of the hydraulic cylinder. In the marine environment, the heave motion of the mother ship caused by wave action has obvious periodic characteristics, which will be directly transmitted to the motion of the hydraulic cylinder of the heave compensation device. By continuously collecting the extension and contraction speed data of the hydraulic cylinder, a time series data set is established, and the sampling duration covers at least 5 complete wave periods to ensure statistical effectiveness.

[0045] The collected speed time series is subjected to fast Fourier transform to convert the time domain signal to the frequency domain for analysis. In the frequency domain, the frequency component with the most concentrated energy corresponds to the main frequency of the heave motion, and the reciprocal of the main frequency is the heave period. In order to improve the accuracy of period identification, the original signal is windowed using the Hanning window function to reduce the spectral leakage effect. In actual sea conditions, the heave period usually varies between 4 and 8 seconds, and the period differs under different sea conditions, so the period parameter needs to be updated in real time. The period identification result is verified by autocorrelation function analysis, the correlation coefficient of the speed sequence under different time delays is calculated, and the time delay corresponding to the first peak of the correlation coefficient corresponds to the heave period. This method can accurately capture the periodic characteristics of the heave motion, providing a time reference for subsequent phase division and link quality analysis, ensuring that the phase locking strategy can be accurately synchronized with the actual mechanical motion.

[0046] After determining the heave period, the signal strength data of the distributed sensor nodes needs to be continuously collected within a complete period to establish a link quality archive. The signal strength data is obtained through the wireless transceiver module of the sensor node, which automatically measures the received signal strength indication value when receiving data packets. This value is expressed in dBm and represents the signal power. In order to obtain comprehensive link quality characteristics, data collection is performed for at least 10 consecutive heave periods, with a sampling frequency of 20Hz in each period to ensure that signal changes during rapid hydraulic cylinder movement can be captured. The time stamp, hydraulic cylinder stroke position and extension and contraction speed corresponding to each signal strength sampling point are recorded synchronously during the collection process to form a multi-dimensional associated data set.

[0047] Considering the random fading characteristics of the wireless channel, the original signal strength data is preprocessed using median filtering to remove burst interference noise. At the same time, the signal quality indicators of each sampling point are calculated, including signal-to-noise ratio and bit error rate, which can more comprehensively reflect the communication quality of the link. One heave period is equally divided into 16 phase segments, and the mean, maximum, minimum and standard deviation of the signal strength are calculated in each phase segment to establish a phase-signal strength statistical table. Through the superposition and averaging of multi-period data, the influence of random factors is eliminated, and the stable periodic variation mode is extracted to provide a reliable data basis for constructing the time series correlation model.

[0048] The construction of the time series correlation model needs to mine the internal relationship between real-time stroke data, real-time extension speed and signal strength data. A prediction model is established by using multiple regression analysis method, taking signal strength as the dependent variable and hydraulic cylinder stroke position and extension speed as the independent variables. The independent influence and interaction of the two independent variables on signal strength are considered during model construction. The stroke position mainly affects the signal propagation path by changing the spatial geometric relationship and shielding state between nodes and gateways, while the extension speed affects the frequency stability of the signal through the Doppler effect.

[0049] Through training of historical data, the model parameters are determined, and the regression coefficients are solved by least squares method. In model verification, independent test data set is used to evaluate the prediction accuracy. In the model, the phase angle is introduced as the time variable, the heave cycle is mapped to the phase space of 0 to 360 degrees, and the periodic function relationship between the phase angle and the link quality is established. Through analysis, it is found that in the phase angle 0-90 degrees and 270-360 degrees interval, corresponding to the fast contraction and extension stage of the hydraulic cylinder, the signal strength fluctuates sharply and the overall level is low; in the phase angle 90-270 degrees interval, corresponding to the slow movement or transient stay stage of the hydraulic cylinder, the signal strength is relatively stable and at a high level.

[0050] Based on the time series correlation model, the link quality state at future time can be predicted in real time according to the current stroke data and extension speed. The model provides a decision basis for the phase-locked routing freezing strategy, so that the opening and closing time of the routing freezing window can accurately match the actual change of link quality, realizing predictive network topology management.

[0051] This embodiment realizes the quantitative characterization of the complex relationship between mechanical motion and communication link quality by establishing a complete process of heave cycle identification, signal strength collection and time series correlation modeling. The cycle identification method based on fast Fourier transform and autocorrelation analysis can accurately capture the periodic characteristics of heave motion in marine environment, providing an accurate time reference for subsequent analysis. The multi-cycle, high-frequency signal strength data collection establishes a comprehensive link quality archive, and through statistical analysis, the influence of random interference is eliminated, and the stable periodic change rule is extracted. The time series correlation model successfully reveals the causal relationship between the stroke position, extension speed and signal strength of the hydraulic cylinder, so that the network layer can predict the link quality change trend according to the real-time state of the mechanical layer. This method changes the traditional passive response topology management to active prediction management, effectively improves the time matching accuracy of the routing freezing window, avoids the wrong decision of executing routing freezing in good link quality period or attempting communication in poor link quality period, significantly improves the network resource utilization efficiency and communication reliability, and provides scientific theoretical support and technical guarantee for the stable operation of distributed submersible monitoring system in complex marine environment.

[0052] In one embodiment of the present embodiment, based on the real-time stretching speed and real-time stroke data, the physical interference coefficient of the distributed sensor node is calculated, including the following steps: S310, comparing the absolute value of the real-time stretching speed with the preset speed threshold value; S320, when the absolute value of the real-time stretching speed is greater than the speed threshold value, it is determined that the current is in a high-speed motion state, and a frequency shift influence factor is calculated; S330, according to the real-time stroke data, it is judged whether the current position of the hydraulic cylinder is in the metal shielding interval; S340, when the current position of the hydraulic cylinder is in the metal shielding interval, a cutting influence factor is calculated; S350, coupling degree analysis is performed on the frequency shift influence factor and the cutting influence factor, and a synergistic action coefficient of the two is calculated; S360, according to the synergistic action coefficient, it is judged whether the frequency shift influence factor and the cutting influence factor are in an activated state at the same time; S370, when the frequency shift influence factor and the cutting influence factor are in an activated state at the same time, the product of the two is calculated as a composite interference strength; S380, when only one influence factor is in an activated state, the numerical value of the activated influence factor is taken as a single interference strength; S390, the composite interference strength or the single interference strength is normalized by a nonlinear mapping function to obtain a physical interference coefficient.

[0053] The motion state of the hydraulic cylinder needs to be accurately compared with the preset speed threshold value. The real-time stretching speed is obtained by differentiating the continuously collected stroke data. In order to reduce the influence of measurement noise and quantization error, five-point central difference method is used to calculate the instantaneous speed. This method can effectively suppress high-frequency noise by performing weighted difference operation on the stroke data of the current sampling point and the two points before and after it. The calculated speed value can be positive or negative, positive indicating hydraulic cylinder extension motion and negative indicating contraction motion, so the absolute value is taken to obtain the size of the motion speed.

[0054] In the wireless sensor network of 2.4 GHz frequency band, when the relative motion speed of the node exceeds 0.25 meters per second, the Doppler shift generated will cause the tracking error of the carrier synchronization circuit at the receiving end, affecting the signal demodulation performance. The speed threshold is usually set to 0.25 meters per second, which can effectively distinguish between low-speed stable motion and high-speed intense motion. The comparison process is realized by using a digital comparator. When the absolute value of the real-time stretching speed is greater than the speed threshold, a high-level signal is output to trigger the high-speed motion state flag, which is used as the enabling condition for the subsequent frequency shift influence factor calculation. In order to avoid the state jitter caused by the fluctuation of the speed near the threshold, a Schmidt trigger mechanism is introduced, different upper and lower thresholds are set, a hysteresis comparison characteristic is formed, and the stability and reliability of the state determination are improved.

[0055] When it is determined that the hydraulic cylinder is in a high-speed motion state, the frequency shift influence factor needs to be calculated to quantify the influence of the Doppler effect on the communication link. The Doppler effect is a physical phenomenon that the frequency of the received signal is shifted relative to the transmitted frequency due to the relative motion between the transmitting node and the receiving node. In the distributed diving monitoring system, the sensor node is fixedly installed on the piston rod or cylinder body of the hydraulic cylinder. When the hydraulic cylinder is stretched at high speed, the node produces radial motion relative to the fixedly installed gateway node, causing Doppler shift.

[0056] The calculation of the frequency shift influence factor needs to consider the frequency tracking ability of the receiver and the robustness of the demodulation algorithm. The speed is mapped to the influence factor interval of 0 to 1 by using normalization processing. In the specific calculation, first, the absolute value of the real-time stretching speed is subtracted from the speed threshold to obtain the speed increment exceeding the threshold, and then divided by the reference speed range to obtain the initial normalized value. Considering that the frequency shift influence is not linearly increased, but at low speed the influence is small, and at high speed the influence increases sharply, a quadratic function is used for nonlinear mapping, the initial normalized value is squared, so that the frequency shift influence factor grows faster at high speed. When the speed reaches a high level, the frequency shift influence factor is close to 1, indicating that the influence of the Doppler effect on the communication reaches a serious level, and the link quality deteriorates significantly.

[0057] Judging whether the current position of the hydraulic cylinder is in the metal shielding interval needs to establish the mapping relationship between the stroke position and the shielding state. In the mechanical structure of the heave compensation device, the cylinder body, piston rod and surrounding support frame of the hydraulic cylinder and other metal components will form physical shielding to the wireless signal propagation. Through field survey and three-dimensional modeling analysis, the specific position range of the shielding interval is determined. When the hydraulic cylinder is in the fully retracted state, the piston rod is almost completely retracted into the cylinder body, and the sensor node installed on the piston rod is surrounded by the thick metal cylinder body, the direct path between the sensor node and the gateway node is completely blocked, and the signal can only be transmitted through diffraction or multipath reflection, resulting in serious signal attenuation.

[0058] In the first 20% of the hydraulic cylinder stroke, the sensor node is in deep occlusion state, and the signal strength is significantly attenuated compared to the unoccluded state. When the hydraulic cylinder extends to more than 80% of the stroke position, the end of the piston rod approaches the installation base of the hydraulic cylinder, and the metal structure of the base and the hydraulic pipeline will partially occlude the signal, so the last 15% of the stroke is also defined as the occlusion interval. The judgment process compares the current stroke position with the pre-marked occlusion interval boundary by reading the real-time stroke data. If the stroke position is less than 20% of the maximum stroke or greater than 85% of the maximum stroke, it is determined that it is currently in the metal occlusion interval, and the occlusion state flag is set to true, otherwise the flag is false. The judgment result is used as the trigger condition for the calculation of the cutting influence factor, to ensure that the corresponding interference assessment is only performed when there is indeed metal occlusion.

[0059] When it is confirmed that the current position of the hydraulic cylinder is in the metal occlusion interval, the cutting influence factor is calculated to quantify the blocking effect of the metal structure on signal propagation. The naming of the cutting influence factor is derived from the cutting effect of the metal occlusion on the electromagnetic wave propagation path, which causes the direct signal to be blocked and the communication link quality to drop sharply. The calculation of the cutting influence factor needs to consider the specific position depth of the node in the occlusion interval, because the occlusion degree varies with the position.

[0060] In the front-end occlusion interval, the occlusion degree increases as the stroke position decreases, and when the stroke position approaches 0, the node is completely surrounded by the cylinder body, and the occlusion is the most serious. Linear interpolation method is used to calculate the cutting influence factor in the front-end occlusion interval, the stroke position is normalized to the interval of 0 to 1, when the stroke position is 0, the cutting influence factor is 1, when the stroke position is 20% of the maximum stroke, the cutting influence factor is smaller, and the intermediate position is interpolated in a linear relationship. In the rear-end occlusion interval, the occlusion degree is relatively light and changes relatively gently, the cutting influence factor changes in a small range, and linear interpolation is also used for calculation. Considering that metal occlusion not only causes signal power attenuation, but also causes multipath propagation and delay spread, leading to inter-symbol interference, a multipath fading correction term is introduced into the cutting influence factor. The final cutting influence factor comprehensively reflects the direct attenuation effect of metal occlusion on signal propagation and the indirect interference caused by multipath effect, providing accurate occlusion influence evaluation for subsequent interference strength calculation.

[0061] The coupling degree analysis of frequency shift influence factor and cutting influence factor is used to reveal the interaction relationship between the two physical interference mechanisms. In actual systems, Doppler frequency shift and metal occlusion do not act independently, but may occur simultaneously and produce a synergistic effect. When the hydraulic cylinder moves at high speed in the occlusion interval, the sensor node is not only blocked by the metal structure, but also produces frequency shift due to rapid movement, and the superposition of the two interferences leads to a deterioration of the link quality that exceeds the simple addition of single interference.

[0062] The calculation of the synergy coefficient is based on the numerical value and the activation state of the two influence factors. First, it is determined whether the frequency shift influence factor exceeds the activation threshold, which corresponds to a certain motion speed, at which point the Doppler frequency shift begins to have a noticeable impact on the demodulation performance. At the same time, it is determined whether the cutting influence factor exceeds the activation threshold, which corresponds to a mild occlusion state. When both influence factors exceed their respective activation thresholds, it is considered that the two interference mechanisms are activated simultaneously, and the synergy coefficient is calculated. The synergy coefficient reflects the coupling strength of the two interferences, and is calculated using the geometric mean method, i.e., the square root of the product of the two influence factors, to obtain the basic coupling strength, and then multiplied by the synergy gain coefficient, which is determined through experiments and reflects the phenomenon that the superposition effect of double interference is greater than the linear accumulation.

[0063] If only one influence factor exceeds the activation threshold, the synergy coefficient is set to 0, indicating that there is no synergy effect, and the interference strength is determined only by the activated single factor. Through the coupling degree analysis, the worst communication environment can be accurately identified, i.e., the period when the hydraulic cylinder is moving at high speed within the occlusion interval, providing key decision basis for the phase locking strategy.

[0064] According to the synergy coefficient, it is determined whether the frequency shift influence factor and the cutting influence factor are in an activated state simultaneously, which is the key branch for selecting the interference strength calculation method. The determination of the activated state is based on whether the numerical value of the influence factor exceeds the pre-set activation threshold. The judgment logic uses AND gate operation, and only when both conditions are met simultaneously, it is determined that both are in an activated state. The result of this determination determines whether to use the composite interference strength calculation or the single interference strength calculation.

[0065] The physical scenario corresponding to the simultaneous activation state is that the hydraulic cylinder is moving at high speed within the occlusion interval, at which time the sensor node faces the worst communication environment, with the signal being severely attenuated by the metal structure and the frequency mismatch caused by the rapid displacement, and the signal processing at the receiving end faces double challenges. Experimental data show that the packet loss rate under the simultaneous activation state is significantly higher than that under the single interference state. Through the clear activation state determination, different severity of interference scenarios can be distinguished, providing accurate classification basis for the quantitative calculation of the interference strength. If the determination result is simultaneous activation, the composite interference strength calculation is performed; if the determination result is non-simultaneous activation, the single interference strength calculation is performed, ensuring that the calculation of the physical interference coefficient can accurately reflect the actual link quality condition.

[0066] When both frequency shift and cutting impact factors are active, the product of the two is calculated as the composite interference strength, and a synergy gain is applied to reflect the superimposed effect of double interference. The calculation of the composite interference strength first directly multiplies the numerical values of the two impact factors to obtain a basic composite value. This multiplication operation reflects the interaction of the two interference mechanisms, and when either factor increases, the composite interference strength will significantly increase. Considering that actual measurements have found that the degree of damage to link quality when double interference acts simultaneously is greater than the simple multiplication of the effects of the two individual interferences, there is a synergistic degradation effect, so a synergy gain coefficient is multiplied by the basic composite value.

[0067] This gain coefficient is obtained by comparing the measured data in single interference and double interference scenarios. When both types of interference exist simultaneously, the link quality degradation speed significantly accelerates. In order to prevent exceeding the upper limit of physical significance, a saturation limiting mechanism is set to retain a certain dynamic range for distinguishing extreme adverse conditions. The composite interference strength accurately quantifies the degree of link quality degradation of the hydraulic cylinder when moving at high speed in the occlusion interval. The closer this value is to 1, the worse the communication environment, and the link is almost unusable, requiring strict route freezing strategies to avoid invalid communication attempts and resource waste.

[0068] When only one impact factor is active, the numerical value of the active impact factor is directly taken as the single interference strength, without the need for composite calculation and synergy gain processing. Single activation corresponds to two typical scenarios: one is when the hydraulic cylinder moves at high speed in the non-occlusion interval, at which time only the frequency shift impact factor is active, and the interference is mainly due to the Doppler effect; the other is when the hydraulic cylinder moves at low speed or is stationary in the occlusion interval, at which time only the cutting impact factor is active, and the interference is mainly due to metal occlusion.

[0069] In the first scenario, although there is a frequency offset, there is no metal occlusion, and the signal power is still sufficient. The automatic frequency control circuit at the receiving end can track and compensate for the frequency offset within a certain range, and the communication quality may decline but can still maintain basic data transmission. In the second scenario, although the signal is attenuated by the metal structure, there is no frequency mismatch problem because the node is stationary or moving at low speed, and the receiving end can partially compensate for the effects of signal attenuation by increasing the receiving sensitivity and using forward error correction coding and other technical means. The numerical value of the single interference strength is directly equal to the numerical value of the active impact factor. Compared with the composite interference strength, the numerical value of the single interference strength is usually smaller, reflecting the actual situation that the link quality is deteriorated but to a relatively lesser extent under single interference mechanism. By distinguishing between single interference and composite interference, more refined link quality assessment can be achieved, avoiding the use of a uniform conservative strategy for all interference scenarios, while ensuring communication reliability and improving network resource utilization efficiency.

[0070] The composite interference intensity or the single interference intensity is normalized by a nonlinear mapping function to obtain a final physical interference coefficient, which is used as a quantitative index for subsequent high-low resistance state phase interval identification. The selection of the nonlinear mapping function needs to consider the nonlinear relationship between the interference intensity and the actual link quality. When the interference is light, the link quality decreases slowly and the system has a certain tolerance. When the interference exceeds a certain critical point, the link quality deteriorates rapidly and communication is almost impossible.

[0071] The hyperbolic tangent function is used as the mapping function, which has S-shaped curve characteristics and can smoothly map the input value to the interval of 0 to 1. In the middle region, it has a large slope and is sensitive to changes in interference intensity. At both ends, it tends to be saturated, avoiding the excessive influence of extreme values. The mapping function maps different levels of interference intensity to corresponding physical interference coefficients. When the interference intensity is small, the physical interference coefficient is small, indicating that the interference is light and the link quality is acceptable. When the interference intensity is large, the physical interference coefficient is large, indicating that the interference is serious and the link quality is deteriorating.

[0072] The normalized physical interference coefficient continuously changes between 0 and 1. The larger the value, the more serious the physical interference and the worse the communication link quality. This coefficient comprehensively reflects the combined effects of Doppler shift and metal shielding on communication, providing accurate link quality evaluation for network topology self-healing strategies. The routing freeze decision can be based on the real physical state rather than a simple signal strength threshold, significantly improving the accuracy and effectiveness of the self-healing strategy.

[0073] The embodiment realizes accurate evaluation of communication interference caused by mechanical movement of the hydraulic cylinder. The speed threshold comparison and Schmidt trigger mechanism effectively identify high-speed movement state, avoiding misjudgment caused by state jitter. The frequency shift influence factor and the cutting influence factor respectively quantify the influence degree of the two physical interference mechanisms of Doppler effect and metal shielding, and the calculation method considers the nonlinear characteristics and position dependence, with significantly better evaluation accuracy than the simple linear model. The synergistic gain coefficient accurately reflects the link quality status in the worst communication scenario, making the interference strength evaluation more in line with the actual situation. The calculation method distinguishes between composite interference and single interference, realizing refined link quality classification and avoiding resource waste caused by adopting a unified conservative strategy for all scenarios. The nonlinear mapping function converts the interference strength into a physical interference coefficient, which continuously changes in the interval of 0 to 1, and can accurately reflect the gradual change process of the link quality, providing a quantitative basis for identifying the high-resistance state phase interval and the low-resistance state phase interval. The overall scheme converts complex physical processes into calculable mathematical models, enabling the network topology self-healing strategy to make accurate decisions based on real physical states, significantly reducing the misjudgment rate compared to the traditional signal strength threshold method, significantly improving the time matching accuracy of the routing freezing strategy, and effectively improving the communication stability and reliability of the distributed submersible monitoring system in complex mechanical motion environments.

[0074] In one of the embodiments of the present embodiment, the high-resistance state phase interval and the low-resistance state phase interval in the heave cycle are identified according to the physical interference coefficient, including the following steps: S410, dividing the heave cycle into multiple phase segments and calculating the average value of the physical interference coefficient in each phase segment; S420, comparing the average value of the physical interference coefficient with the preset interference threshold; S430, when the average value of the physical interference coefficient is higher than the preset interference threshold, the corresponding phase segment is marked as a high-resistance state phase interval; S440, when the average value of the physical interference coefficient is lower than or equal to the preset interference threshold, the corresponding phase segment is marked as a low-resistance state phase interval.

[0075] Dividing the heave cycle into multiple phase segments is the basis for realizing refined link quality management. The heave cycle is identified by the foregoing method and is usually between 4 to 8 seconds. In order to balance the time resolution and calculation complexity, a complete heave cycle is equally divided into 16 phase segments, each corresponding to a phase angle of 22.5 degrees, with a time span of about 0.25 to 0.5 seconds. The phase segment division uses phase angle as the reference rather than absolute time, because the heave cycle may change with sea conditions, and the division based on phase angle can ensure that the correspondence between the phase segment and the hydraulic cylinder movement state remains consistent under different cycle lengths.

[0076] In each phase segment, the real-time calculated physical interference coefficient values are continuously collected, with a sampling frequency of 20 Hz, ensuring that a sufficient number of samples are obtained within each phase segment, typically containing 5 to 10 physical interference coefficient sampling points. Statistical analysis is performed on these sampling points, and the arithmetic mean is calculated as the representative interference level for that phase segment. Before calculating the mean, outlier detection is performed using the standard deviation criterion to remove obvious outliers that may be caused by transient electromagnetic interference or sensor reading errors. These outliers are removed to improve the representativeness of the mean. At the same time, the standard deviation of the physical interference coefficient within each phase segment is calculated, which reflects the degree of fluctuation of the interference level in that phase segment. A smaller standard deviation indicates a stable interference state, while a larger standard deviation indicates an unstable interference state. By superimposing and statistically analyzing data from multiple heave cycles, a phase segment-average interference coefficient table is established, which records the typical interference levels corresponding to each of the 16 phase segments, providing a data basis for subsequent threshold comparison and state marking.

[0077] The average physical interference coefficient of each phase segment is compared with the preset interference threshold to determine the link quality state of that phase segment. The setting of the interference threshold is a key parameter for identifying high and low resistance state phase intervals, and needs to consider the performance indicators and business requirements of the communication system. Through experimental testing, the packet loss rate, transmission delay, and throughput of different physical interference coefficients are analyzed to establish a quantitative relationship between the interference coefficient and the communication quality.

[0078] When the physical interference coefficient is low, the packet loss rate is small, the transmission delay is within the normal range, and the communication quality meets the monitoring data transmission requirements. When the physical interference coefficient is at a medium level, the packet loss rate increases, and the communication quality begins to deteriorate but can still maintain basic transmission. When the physical interference coefficient is high, the packet loss rate increases significantly, and the communication is almost unusable. Based on these experimental results, the interference threshold is set to 0.4, which serves as the dividing line between high and low resistance states, effectively distinguishing between acceptable and unacceptable communication quality states.

[0079] The comparison process is implemented using a digital comparator to numerically compare the average interference coefficient of the phase segment with the threshold value, and the comparison result is output as the basis for state marking. To improve the robustness of the identification, a time continuity constraint is introduced, requiring that at least two adjacent phase segments have average interference coefficients exceeding the threshold value before marking these phase segments as high resistance state, avoiding false positives caused by accidental fluctuations in a single phase segment. At the same time, adaptive adjustment of the interference threshold is considered, based on historical statistical data and current sea conditions, to fine-tune the threshold within a certain range, allowing the high and low resistance state identification to adapt to different working environments. In severe sea conditions, the threshold is appropriately lowered to improve protection, while in calm sea conditions, the threshold is appropriately raised to increase communication opportunities.

[0080] When the average value of the physical interference coefficient of a phase segment is higher than the preset interference threshold, the phase segment is marked as a high-resistance state phase interval, indicating that the communication link quality is in a deteriorated state and is not suitable for normal data transmission and routing maintenance activities in this period. The high-resistance state marking process maintains a phase state table in the network management module, which contains 16 entries corresponding to the 16 phase segments of the heave cycle. Each entry contains a phase angle range, an average interference coefficient, a state mark, and a time stamp field. When a phase segment is determined to be in a high-resistance state, the state mark field of the corresponding entry is set to a high-resistance state, and the time stamp of the marking time is recorded.

[0081] High-resistance state phase intervals are usually concentrated in the stages of rapid contraction and extension of the hydraulic cylinder. In these stages, the hydraulic cylinder moves at high speed and may be in the occlusion interval, and the physical interference coefficient increases significantly. The phase segment marked as a high-resistance state triggers the execution of the phase-locked routing freezing strategy when the same phase is reached in the next heave cycle. The system calculates the opening time of the routing freezing window in advance and completes the preparation work such as routing table locking, message sending suspension, and data caching before the time comes. The high-resistance state marking is also used to generate a link quality prediction curve, which takes the phase angle as the horizontal axis and the physical interference coefficient as the vertical axis, directly showing the trend of link quality change in the heave cycle, helping the system administrator to understand the network operation state. By accurately marking the high-resistance state phase interval, the network topology self-healing strategy can predictively avoid the link quality deterioration period and concentrate communication activities in the period with good link quality, significantly improving the communication success rate and network resource utilization efficiency.

[0082] When the average value of the physical interference coefficient of a phase segment is lower than or equal to the preset interference threshold, the phase segment is marked as a low-resistance state phase interval, indicating that the communication link quality is in a normal or good state and is suitable for data transmission and network maintenance activities in this period. Low-resistance state marking is also performed in the phase state table, setting the state mark field of the corresponding phase segment to a low-resistance state and recording the marking time. Low-resistance state phase intervals usually correspond to stages where the hydraulic cylinder moves slowly or is in a transient state. In these stages, the hydraulic cylinder is located at the intermediate stroke position, there is no serious metal occlusion, the movement speed is relatively low, the Doppler effect is small, and the physical interference coefficient remains at a low level.

[0083] The phase section marked as low resistance state is a time window for the system to concentrate data transmission, and in these time periods, a thawing strategy is executed to read the accumulated data to be transmitted in the high resistance state phase interval from the local cache queue and quickly send it using a burst transmission mode. The duration of the low resistance state phase interval directly affects the throughput of data transmission. Through statistical analysis, the typical duration of the low resistance state phase interval is 2 to 4 seconds, and during this period, data is transmitted at the maximum transmission rate to empty the accumulated data packets in the cache queue. The low resistance state marker is also used to schedule the routing discovery and topology update activities, and only in the low resistance state phase interval, the routing request message and the link state advertisement message are sent to ensure that the control message can be successfully received and processed to maintain the accuracy of the network topology. By accurately identifying and marking the low resistance state phase interval, time division multiplexing of communication resources is achieved, maximizing data transmission efficiency in limited good communication periods, while ensuring the normal operation of the network protocol, so that the distributed submersible monitoring system can maintain stable and reliable communication performance in a periodic mechanical interference environment.

[0084] The present embodiment realizes fine timing analysis of link quality by dividing the heave period into 16 phase sections and calculating the average physical interference coefficient of each phase section. Based on the comparison and determination of the interference threshold, the high resistance state and low resistance state phase intervals are accurately identified. The marking of the high resistance state phase interval enables the system to predictively avoid the link quality deterioration period, and through the phase-locked routing freeze strategy, it avoids unnecessary transmission attempts in poor communication conditions, greatly reducing the routing control message overhead and node energy consumption. The marking of the low resistance state phase interval provides a time window for concentrated data transmission, and through the burst transmission mode, the accumulated data is sent in the low resistance state period, effectively improving the data transmission success rate. The phase section division and state marking mechanism realizes time division multiplexing of communication resources, fully utilizes the limited good communication period, and significantly improves the network throughput. The adaptive interference threshold adjustment enables the identification strategy to adapt to different sea conditions, providing stronger protection in harsh environments and more communication opportunities in calm environments, significantly enhancing the adaptability of the system. The overall scheme provides accurate phase state information for the network topology self-healing strategy, enabling the routing freeze and thaw operations to be accurately synchronized with the mechanical movement period of the hydraulic cylinder, realizing deep collaboration between the mechanical layer and the network layer, and significantly improving the communication stability and reliability of the distributed submersible monitoring system in complex marine environments.

[0085] In one embodiment of the present embodiment, the phase-locked routing freeze strategy is executed, including the following steps: S510, according to the heave period and the start phase of the high resistance state phase interval, calculating the opening time and closing time of the routing freeze window; S520, at the opening time of the routing freeze window, setting a routing table lock flag at the network routing protocol layer; S530, pause sending all packets and link state advertisement packets; S540, mark all received route invalid alarm as to-be-verified state, and do not trigger topology reconstruction process; S550, store current to-be-transmitted data packet to local cache queue; S560, clear route table locking flag at the closing time of route freezing window.

[0086] The opening and closing time of the route freezing window needs to accurately synchronize the phase change of the heave cycle and the control timing of the network protocol. First, the current system time is obtained as the calculation reference, which is synchronized through the network time protocol to ensure the consistency of the clocks of each node. According to the identified heave cycle, the cycle length is determined. Through the marked high resistance phase interval, the starting phase angle and the ending phase angle are obtained. The phase angle is converted into time offset, and the calculation formula is time offset equals phase angle divided by 360 degrees and multiplied by heave cycle.

[0087] By monitoring the real-time stroke data and extension and retraction speed of the hydraulic cylinder, the current phase position is tracked in real time, and the phase-locked loop technology is used to ensure the accuracy of phase tracking. When it is detected that the current phase is about to enter the high resistance phase interval, the opening time of the route freezing window is calculated in advance by a certain time, and this advance is used to compensate for the calculation delay and instruction execution time, to ensure that the route freezing operation takes effect when the high resistance phase interval really starts. The calculation formula of the opening time is the current time plus the remaining time to reach the starting phase of the high resistance state minus the advance. The closing time is calculated based on the ending phase of the high resistance phase interval, and the phase angle to time conversion is also used, and the closing time equals the opening time plus the duration of the high resistance phase interval.

[0088] Considering that the heave cycle may have slight fluctuations, the sliding average method is used to smooth the cycle length of the recent several cycles, and the smoothed cycle length is used for time calculation to improve the stability of the calculation. After the calculation is completed, the opening time and the closing time are stored in the timer register, and the timer counts down, and when the countdown is zero, the corresponding freezing or unfreezing operation is triggered to ensure the accuracy of time control.

[0089] At the opening time of the route freezing window, the route table locking flag needs to be set in the network routing protocol layer to prevent any modification operation on the route table. The route table is a core data structure maintained by the network layer, which records the next hop information, path cost and link state of the arriving nodes to the target nodes and other key parameters. In the normal operating state, the route protocol will dynamically update the route table according to the received route update packet, link state change and topology discovery result to adapt to the changes of network topology.

[0090] The route table locking flag is a Boolean variable stored in the global state register of the routing protocol module, and the initial value is false. When the timer triggers the route freezing window opening event, the interrupt service program sets the route table locking flag to true, and the setting of the flag uses atomic operation to ensure the safety of operation in a multi-threaded environment. After the route table locking flag takes effect, all route table write operations of the routing protocol module will first check the flag state, and if the flag is true, the write operation will be rejected and an error code indicating that the operation is rejected will be returned. In specific implementation, the flag checking code is added at the entrance of the route table update function to form a protection mechanism.

[0091] The route table locking does not affect the read operation of the route table, and the data forwarding module can still query the route table for packet forwarding, ensuring that the communication on the established path is not affected. The locking mechanism also prevents the route aging timer from deleting route table entries. Because in the high resistance state phase interval, even if some route entries exceed the aging time, it may be a false failure caused by temporary degradation of link quality and should not be deleted immediately. By setting the route table locking flag, the modification of the route table triggered by false judgment of link failure during the period of link quality degradation is effectively prevented, the stability of the network topology is maintained, and the control overhead and communication interruption caused by frequent topology reconstruction are avoided.

[0092] Pausing the sending of all route discovery packets and link state advertisement packets is a key control measure of the phase locking route freezing strategy. Route discovery packets include route request packets and route response packets, which are used to detect network topology and establish route paths. Link state advertisement packets are used to propagate the neighbor relationship and link quality information of nodes in link state routing protocols, and each node sends the packet periodically or when the link state changes, so that all nodes in the network can obtain a global topology view.

[0093] In the high resistance state phase interval, due to the degradation of link quality, the success rate of sending these control packets is extremely low, and even if they are sent, they are likely to be lost due to signal attenuation or frequency mismatch, causing ineffective energy consumption and channel occupation. The implementation of pausing the sending adds a frozen state check in the packet sending function of the routing protocol module, and when it detects that the route freezing window is in the open state, all route discovery packets and link state advertisement packets are intercepted, and the packet does not enter the sending queue, but returns a sending delay state code. For periodically sent link state advertisement packets, pause the trigger of the timer to avoid sending requests during freezing. For event-triggered route discovery packets, record the triggering event in the pending queue and wait for the freezing window to close before processing.

[0094] It should be noted that the suspension of sending is only for routing control packets, and does not affect the reception and processing of data packets. The node can still receive packets from other nodes, maintaining the ability to perceive the network state. By suspending the sending of routing control packets, invalid communication attempts during periods of poor link quality are avoided, saving valuable energy resources and channel bandwidth, and reserving resource space for efficient data transmission in the high-resistance state phase interval.

[0095] Marking all received route failure alarms as a to-be-verified state without immediately triggering the topology reconstruction process is a key mechanism for distinguishing between false failures and real failures. Route failure alarms are usually triggered by the following situations: no acknowledgement is received after continuously sending data packets for multiple times, the heartbeat packet of a neighbor node times out without being received, or a link disconnection notification packet sent by another node is received. In traditional network topology self-healing mechanisms, once a route failure is detected, the topology reconstruction process is immediately started, the failed route is deleted, and a new communication path is established through route discovery.

[0096] However, in the heave compensation scenario, the link quality degradation in the high-resistance state phase interval is temporary, determined by the periodic characteristics of mechanical motion. When the hydraulic cylinder moves to the low-resistance state phase interval, the link quality will automatically recover, and the route failure at this time is a false failure, which does not require topology reconstruction. The to-be-verified state is implemented by maintaining a route failure record table in the routing protocol module, which includes fields such as the target address of the failed route, the failure time, the failure reason, and the verification state. When a route failure alarm is received within the routing freeze window, no operation of deleting the route and triggering reconstruction is performed, but a new record is created in the route failure record table, the verification state field is set to to-be-verified, and the failure time and reason are recorded.

[0097] A verification timer is also set, with a timeout time set to one complete heave cycle, ensuring enough time to cross the high-resistance state phase interval and enter the low-resistance state phase interval for verification. In the low-resistance state phase interval, the system performs a second verification on all routes in the to-be-verified record table. The specific method is to send a probe packet to the target node of the failed route. If a response is received, it is confirmed as a false failure, the record is cleared, and the original route is maintained. If multiple consecutive probes have no response, it is confirmed as a real failure, and route deletion and topology reconstruction are performed. Through the to-be-verified mechanism, frequent topology reconstruction caused by periodic link fluctuations is effectively avoided, significantly reducing control overhead and network jitter.

[0098] Storing the current data packet to be transmitted to the local cache queue is an important measure to ensure that data is not lost. In the high resistance state phase interval, due to the suspension of message sending and the deterioration of link quality, the monitoring data generated by the application layer cannot be sent immediately. If these data are discarded directly, the monitoring information will be missing, affecting the monitoring function of the system. The local cache queue is a buffer allocated in the memory of each sensor node, implemented using a circular queue data structure, with a fixed capacity limit, usually configured according to the available memory size of the node.

[0099] The queue includes three control variables: head pointer, tail pointer, and counter. The head pointer points to the earliest data packet in the queue, the tail pointer points to the next available storage location, and the counter records the number of data packets in the queue. When the application layer generates a new data packet that needs to be sent, the network layer first checks the routing freeze state. If it is in the freeze state, it calls the enqueue function of the cache queue to write the data packet to the position pointed to by the tail pointer, then increments the tail pointer by one and takes the modulus of the queue length, and increments the counter by one. Data packets in the queue are sorted by priority and timestamp. High-priority data such as physiological parameter abnormal alarm and device fault alarm are placed at the front of the queue to ensure that they are sent first after unfreezing. For data packets with the same priority, they are arranged in the order of timestamp, following the first-in-first-out principle.

[0100] When the cache queue is full, a selective discard strategy is used to discard the data packet with the lowest priority and the earliest timestamp to make room for new data, and the packet loss event is recorded for statistical analysis. The cache queue also implements data compression for periodic monitoring data. If the values of consecutive data packets change very little, they can be compressed to reduce storage space usage. Through the local cache mechanism, it is ensured that the data generated during the period of link quality deterioration will not be lost, and the monitoring data can be sent in a concentrated manner when the low resistance state phase interval arrives, ensuring the integrity and continuity of the monitoring data.

[0101] The routing table lock flag is cleared at the closing time of the routing freeze window, restoring the normal operation state of the routing protocol. The closing time corresponds to the end time of the high resistance state phase interval, at which time the hydraulic cylinder motion enters the low resistance state phase interval and the link quality begins to recover. The timer triggers an interrupt when it reaches the closing time, and the interrupt service program performs the clearing operation of the routing table lock flag, setting the flag from true to false. This operation also uses atomic operations to ensure thread safety.

[0102] After the route table lock flag is cleared, the route protocol module resumes the write permission to the route table, and can modify the route table according to the newly received route update information. The route aging timer also resumes the work, and cleans up the expired route items. The clearing operation also triggers a series of subsequent actions, including checking the accumulated pending route event queue during the freezing period, processing the events, such as processing the delayed route discovery request, sending the paused link state advertisement packet, etc. At the same time, the sending process of the cached data is started, the data packets are read from the local cache queue according to the priority order, and the network layer sending function is called for transmission.

[0103] In order to fully utilize the good link quality in the low resistance state phase interval, a burst transmission mode is adopted, a plurality of data packets are continuously sent, and the interval between the data packets is shortened to the minimum interframe interval, so as to improve the channel utilization. The sending success rate is monitored during the sending process. If the success rate is lower than expected, it may indicate that the link quality has not fully recovered, at which time the sending rate is reduced or the sending of part of the data is delayed. Clearing the route table lock flag means that a complete cycle of the phase-locked route freezing strategy ends, the system enters the normal communication stage of the low resistance state phase interval, and is ready for the freezing operation of the next heave cycle, realizing the periodic switching of freezing and unfreezing, so that the network communication and mechanical movement are accurately synchronized.

[0104] The present embodiment realizes the accurate synchronization of network protocol control and mechanical movement of the heave compensation device by accurately calculating the opening and closing time of the route freezing window, and ensures that the route freezing operation can accurately cover the high resistance state phase interval. The setting of the route table lock flag effectively prevents the route table modification triggered by misjudgment during the link quality degradation period, greatly improves the network topology stability, and avoids the communication interruption caused by frequent topology reconstruction. The measure of pausing the sending of route control packets avoids the invalid transmission attempt in the poor communication environment, significantly reduces the route control packet overhead and node energy consumption, and prolongs the working life of the battery-powered node. The to-be-verified mechanism of the route failure alarm successfully distinguishes between false failure and real failure, greatly reduces the misjudgment rate of false failure, reduces the number of unnecessary topology reconstruction, and basically eliminates network jitter. The local cache queue ensures that the monitoring data generated in the high resistance state phase interval will not be lost, and the data integrity is effectively guaranteed. The cache capacity can cover the typical high resistance state duration. The timely clearing of the route table lock flag ensures that the normal communication can be quickly restored in the low resistance state phase interval, and the maximum effective time window is obtained for centralized data transmission. The overall phase-locked route freezing strategy realizes the deep cooperation between the network layer and the physical layer, changes the passive response type of topology management to active prediction type, and significantly improves the communication stability, reliability and efficiency of the distributed diving monitoring system in the periodic mechanical interference environment.

[0105] In one of the implementations of the embodiment, a thawing strategy is executed, including the following steps: S610, according to the heave period and the start phase of the low resistance state phase interval, calculating the opening time and the duration of the data transmission window; S620, at the opening time of the data transmission window, clearing the routing table lock flag and restoring the normal operation of the routing protocol; S630, according to the priority order, reading the data packets to be transmitted from the local cache queue and performing burst transmission; S640, receiving and processing the routing discovery packet and the topology update request sent by the neighbor node; S650, performing secondary verification on the route failure alarm marked as the to-be-verified state in the high resistance state phase interval, and if communication cannot be restored in the low resistance state phase interval, it is determined as a real failure.

[0106] The opening time and the duration of the data transmission window need to accurately grasp the time characteristics of the low resistance state phase interval. The data transmission window refers to the effective time period for data transmission and network maintenance in the low resistance state phase interval, and the accurate calculation of this time period directly affects the data transmission efficiency and network resource utilization. First, according to the marked low resistance state phase interval, the start phase angle and the end phase angle are obtained. Through the heave period length, the phase angle is converted into a time offset.

[0107] The phase-locked loop technology is used to track the phase position of the current hydraulic cylinder in real time, and the remaining time to reach the start phase of the low resistance state is predicted by monitoring the real-time stroke data and the extension speed. The opening time of the data transmission window is set to the start time of the low resistance state phase interval, and no advance is needed because the link quality has been restored and transmission can start immediately. The calculation formula of the opening time is the current time plus the remaining time to reach the start phase of the low resistance state. The calculation of the duration needs to consider several factors, first of all, the theoretical duration of the low resistance state phase interval, and secondly, a certain safety margin should be reserved to avoid data transmission at the end of the phase interval, resulting in transmission being interrupted by the next high resistance state phase interval.

[0108] The safety margin is used to compensate for phase tracking errors and transmission delays. Therefore, the duration of the data transmission window is equal to the duration of the low resistance state phase interval minus the safety margin. Considering the fluctuation of the heave period, an adaptive adjustment mechanism is adopted to dynamically correct the duration according to the measured data of the recent several periods. If the period is detected to be shortened, the duration is correspondingly shortened, and if the period is detected to be lengthened, the duration is appropriately lengthened, ensuring that the data transmission window is always in a period of good link quality. After calculation, the opening time and the duration are stored in the timer register, and the timer triggers the thawing operation at the opening time and triggers the transmission window closing operation at the end of the duration, preparing for the next frozen period.

[0109] The first step of the unfreezing strategy is to clear the routing table lock flag at the opening time of the data transmission window and resume the normal operation of the routing protocol. When the timer reaches the opening time of the data transmission window, an interrupt service program is triggered, which first performs the clearing operation of the routing table lock flag, sets the flag from true to false, and uses an atomic operation to ensure safety in a multi-threaded environment. After the routing table lock flag is cleared, the routing protocol module immediately resumes full access to the routing table, including reading, writing, updating, and deleting operations.

[0110] The various functional modules of the routing protocol begin to operate normally, the routing update processing module checks the cache of the routing update packets received during the freezing period, processes these packets, and updates the local routing table. The routing aging timer resumes work and checks the routing table for routes that exceed the aging time, deleting routes that are indeed invalid. The link state advertisement timer resumes triggering and prepares to send the link state information of the node to announce the current connection state and link quality parameters to neighboring nodes. The routing discovery module resumes its response capability and can process routing request packets from other nodes to participate in the establishment of routing paths.

[0111] The recovery operation also includes restarting the heartbeat packet sending mechanism to send heartbeat packets to neighboring nodes to maintain neighbor relationships and ensure that the online status of neighboring nodes can be detected in a timely manner. At the same time, the received signal strength monitoring function is activated to collect real-time signal strength data and verify whether the link quality has indeed recovered to the expected level. If the signal strength is below the preset minimum threshold, it indicates that the link quality recovery is not ideal, at which time an alarm will be triggered and the subsequent transmission strategy will be adjusted to reduce the transmission rate or delay the transmission of some non-critical data. After the routing protocol resumes normal operation, the network enters an active state, and full information exchange can be performed between nodes to complete the network maintenance tasks accumulated during the freezing period and create a good network environment for efficient data transmission.

[0112] Reading data packets from the local cache queue in priority order and performing burst transmission is a key measure to fully utilize the low resistance phase interval. The local cache queue stores all the data packets accumulated during the high resistance phase interval, which have been sorted by priority and timestamp. When burst transmission starts, the counter of the cache queue is first checked to determine the number of data packets in the current queue. If the queue is empty, the transmission step is skipped, and if the queue is not empty, the transmission process is started.

[0113] The transmission process starts from the head of the queue, reads the data packet with the highest priority, calls the sending function of the network layer to encapsulate the data packet into a network frame and sends it through the wireless transceiver module. After sending, the head pointer is incremented by one and the queue length is taken modulo, the counter is decremented by one, and then the next data packet is read immediately to continue sending, forming a continuous burst transmission mode. The characteristic of burst transmission is that the transmission interval between data packets is very short, only the minimum frame interval, which is specified by the wireless communication protocol, is maintained to ensure that the receiving end has enough time to process the previous data packet. By shortening the transmission interval, more data packets can be sent in a limited data transmission window, significantly improving channel utilization.

[0114] During burst transmission, the success rate is monitored in real time, and the success rate is calculated by counting the number of received acknowledgement messages and the number of sent data packets. If the success rate is high, it indicates that the link quality is good, and the high-speed burst transmission is continued; if the success rate is at a medium level, it indicates that the link quality is general, and the transmission rate is appropriately reduced to increase the interval between data packets; if the success rate is low, it indicates that the link quality is poor, and the burst transmission is suspended, and the remaining data packets are retained in the cache queue, waiting for the next low-resistance phase interval.

[0115] For high-priority critical data such as device fault alarm, physiological parameter anomaly, etc., the acknowledgement retransmission mechanism is adopted, and after sending, the acknowledgement message from the receiving end is waited for, if no acknowledgement is received within the timeout time, retransmission is performed, to ensure the reliable transmission of critical data. For low-priority periodic monitoring data, the best-effort transmission strategy is adopted, without waiting for acknowledgement, if the transmission fails, it is directly discarded, because there will be new monitoring data generated subsequently, and the value of old data decreases. Burst transmission continues until the data transmission window is about to end or the cache queue is empty, if there are remaining data in the queue at the end of the window, these data are retained in the queue, waiting for the next low-resistance phase interval to continue transmission.

[0116] Receiving and processing the route discovery message and topology update request sent by the neighbor node is an important link to maintain the accuracy of the network topology. In the low-resistance phase interval, the communication ability of each node is restored, and normal network protocol interaction can be carried out. The route discovery message includes route request message and route response message, when a node needs to establish a route to a specific target, it will broadcast a route request message, which contains source node address, target node address, sequence number and hop count, etc.

[0117] After receiving the route request message, the node first checks whether it is a duplicate message. If the same request has been processed, the message is discarded to avoid repeated processing. If it is a new request, the node checks whether it is the target node. If it is, a route response message is generated and sent to the source node along the reverse path to establish a bidirectional route. If it is not, the hop count is incremented and the path information in the message is updated, and then the request message is forwarded to continue propagation. During the forwarding process, the node records the reverse route to the source node in the routing table to prepare for forwarding of subsequent route response messages.

[0118] The topology update request is usually initiated by a gateway node or cluster head node to collect global topology information of the network. The request message contains parameters such as topology version number, update type, and data collection range. After receiving the topology update request, the node collects local topology information, including neighbor node list, link quality parameters, and residual energy, and encapsulates the information into a topology update response message to send to the requesting node. The processes of processing route discovery messages and topology update requests consume certain computing resources and communication bandwidth, so it is necessary to reasonably control the processing priority to ensure that the main task of data transmission is not affected. A priority scheduling mechanism is adopted, with data packet transmission set as the highest priority and route protocol message processing set as medium priority. Route protocol messages are processed during data transmission gaps to coordinate data transmission and network maintenance. By timely processing route discovery messages and topology update requests, the network topology information is updated in a timely manner to provide accurate routing information for subsequent data transmission, improving the overall performance and reliability of the network.

[0119] The secondary verification of route failure alarms marked as to-be-verified state in the high-resistance state phase interval is a key step to distinguish false failures from real failures. Route failure alarms received in the high-resistance state phase interval are marked as to-be-verified state and recorded in the route failure record table. These alarms may be false failures caused by temporary degradation of link quality, or real node failures or permanent link interruptions. Secondary verification is performed in the low-resistance state phase interval, at which time the link quality has recovered, and if the route is indeed failed, it should be able to be found through verification.

[0120] The verification process first traverses the route failure record table, reads all records in the to-be-verified state, and performs a verification operation on each record. The verification method is to send a probe packet to the target node of the failed route. The probe packet uses a special packet type identifier and contains information such as the source node address, target node address, probe sequence number, and timestamp. After sending the probe packet, a verification timer is started, and the target node's response is waited for during this period. If a probe response packet from the target node is received within the timeout period, it means that the route is actually available, and the previous failure alarm is a false failure. At this time, the corresponding record in the route failure record table is cleared, the original route table entry is kept unchanged, and a false failure event is recorded in the log for statistical analysis.

[0121] If no response is received within the timeout period, a second probe is performed, the probe packet is repeatedly sent, and the response is waited for. If there is still no response, a third probe is performed. After multiple consecutive probes have no response, it is determined that the route is actually failed, which is a real failure. At this time, the route deletion operation is performed, the route entry is deleted from the route table, and the topology reconstruction process is triggered to start the route discovery mechanism to find an alternative path. The determination of a real failure also generates a failure report, records detailed information such as the failure node address, failure time, and failure type, and sends it to the gateway node or monitoring center, so that the system administrator can perform failure diagnosis and maintenance.

[0122] The secondary verification mechanism effectively avoids misjudgment caused by periodic link fluctuations, ensures that only real failures will trigger topology reconstruction, and greatly reduces unnecessary network overhead. The verification process is performed in the second half of the data transmission window, at which time most of the cached data has been sent, there is sufficient time and bandwidth resources for verification operations, and it will not affect the main task of data transmission.

[0123] The embodiment realizes accurate matching of data transmission and low resistance state phase interval by accurately calculating the opening time and duration of the data transmission window, ensures that data transmission is always performed in a period of good link quality. Clearing the routing table lock flag in time and restoring the normal operation of the routing protocol gives the data transmission the maximum effective time window. The burst transmission mode fully utilizes the good link quality of the low resistance state phase interval, significantly improves the channel utilization rate by shortening the data packet sending interval, and successfully clears the buffer queue within the transmission window. The data transmission success rate is greatly improved, and the reliable transmission of key data is guaranteed by the acknowledgement and retransmission mechanism. Timely processing of routing discovery messages and topology update requests ensures the accuracy of network topology information, significantly reduces the routing table update delay, and significantly improves the response speed of the network to topology changes. The secondary verification mechanism successfully distinguishes between false failures and real failures, greatly reduces the false failure misjudgment rate, avoids a large amount of unnecessary topology reconstruction, and significantly reduces network control overhead. Real failures can be discovered and processed in time within the low resistance state phase interval, and the fault response time is greatly shortened. The overall thawing strategy realizes the organic combination of data transmission, network maintenance and fault diagnosis, efficiently completes multiple tasks within the limited low resistance state phase interval, and significantly improves the data transmission efficiency, network stability and fault handling capability of the distributed submarine monitoring system, providing a strong guarantee for the reliable operation of the system in complex marine environments.

[0124] In one of the embodiments of the present embodiment, after monitoring the real-time gas pressure value of the accumulator, the following steps are further included: S710, calculating the equivalent stiffness of the heave compensation device according to the real-time gas pressure value; S720, normalizing the equivalent stiffness to obtain a gas-liquid stiffness coefficient; S730, establishing an inverse proportional mapping relationship between the gas-liquid stiffness coefficient and the network self-healing sensitivity threshold; S740, when the gas-liquid stiffness coefficient increases, the packet loss rate tolerance threshold is increased, and the sensitivity of network self-healing is reduced; S750, when the gas-liquid stiffness coefficient decreases, the packet loss rate tolerance threshold is reduced, and the sensitivity of network self-healing is increased.

[0125] The equivalent stiffness of the heave compensation device is calculated according to the real-time gas pressure value of the accumulator, which is used to quantify the compensation ability and mechanical stability of the system. The equivalent stiffness of the heave compensation device reflects the ability of the system to resist external load changes, and the greater the stiffness, the more stable the system and the stronger the regularity of mechanical movement. The calculation of the equivalent stiffness is based on the gas state equation of the accumulator and the mechanical properties of the hydraulic cylinder. The accumulator, as an energy storage element, follows a variable process relationship between gas pressure and volume. When hydraulic oil enters or exits the accumulator, the gas is compressed or expanded, and the pressure changes accordingly.

[0126] The formula for calculating equivalent stiffness is: stiffness equals the pressure change divided by the displacement change. In practice, the real-time gas pressure value of the accumulator is first obtained, acquired via a pressure sensor. Then, the derivative of pressure with respect to volume is calculated. According to the polytropic process equation, pressure raised to the nth power multiplied by volume is a constant, where n is the polytropic exponent, typically taken within a certain range for accumulators, representing the actual process between isothermal and adiabatic processes. Differentiating this equation yields the rate of change of pressure with respect to volume. Combining this with the effective working area of ​​the hydraulic cylinder, the volume change is converted into a displacement change of the piston rod, thus obtaining the equivalent stiffness.

[0127] The calculation process requires obtaining parameters such as the initial charging pressure, gas volume, and current operating pressure of the accumulator. These parameters are calibrated and stored in the configuration file during system installation and commissioning. Changes in real-time gas pressure directly affect the equivalent stiffness. Higher pressures result in greater gas compression and increased stiffness; lower pressures lead to gas expansion and decreased stiffness. The calculated equivalent stiffness provides the foundational data for subsequent normalization processing.

[0128] Equivalent stiffness is normalized to obtain gas-liquid stiffness coefficients, enabling the stiffness parameters to be correlated with network parameters under a unified dimension. The purpose of normalization is to convert the equivalent stiffness, which has physical units, into dimensionless coefficients, facilitating the establishment of a mapping relationship with the network's self-healing sensitivity threshold. The normalization method employs maximum-minimum normalization. First, the normal operating range of the equivalent stiffness is determined. Through statistical analysis of long-term operational data of the heave compensation device under different sea states, the minimum and maximum values ​​of the equivalent stiffness are determined.

[0129] The normalization formula is that the gas-liquid stiffness coefficient equals the current equivalent stiffness minus the minimum value, divided by the difference between the maximum and minimum values. This formula maps the equivalent stiffness to a range of 0 to 1. When the equivalent stiffness is at its minimum, the gas-liquid stiffness coefficient is 0, indicating the weakest system compensation capability; when the equivalent stiffness is at its maximum, the gas-liquid stiffness coefficient is 1, indicating the strongest system compensation capability; when the equivalent stiffness is in the middle range, the gas-liquid stiffness coefficient is approximately 0.5, indicating the system is in normal operating condition. The normalization process also includes boundary constraints: if the calculated coefficient is less than 0, it is forcibly set to 0; if it is greater than 1, it is forcibly set to 1, to prevent coefficients from exceeding limits due to abnormal data.

[0130] The physical meaning of the gas-liquid stiffness coefficient is to reflect the mechanical stability and motion regularity of the heave compensation device. The larger the coefficient, the more stable the mechanical motion, the more obvious the periodic characteristics, and the more regular the periodic fluctuations of link quality. At this time, the network topology self-healing strategy can rely more on the phase locking mechanism. The smaller the coefficient, the less stable the mechanical motion, and there may be abnormalities or failures. The fluctuations of link quality are irregular, and at this time, the sensitivity of network self-healing needs to be improved to respond to real link changes in a timely manner.

[0131] An inverse proportional mapping relationship is established between the gas-liquid stiffness coefficient and the network self-healing sensitivity threshold to achieve adaptive coordination of mechanical state and network strategy. The network self-healing sensitivity threshold is mainly reflected in the packet loss rate tolerance threshold parameter, which determines the strictness of the system in determining link failure. The lower the threshold, the higher the sensitivity, and the easier it is to determine link failure and trigger topology reconstruction. The higher the threshold, the lower the sensitivity, and the greater the tolerance to link quality fluctuations.

[0132] The establishment of the inverse proportional mapping relationship is based on the following logic: when the gas-liquid stiffness coefficient is large, it means that the heave compensation device is working stably, and the periodicity of mechanical motion is strong. The fluctuations of link quality are mainly caused by predictable periodic mechanical motion, and the possibility of false failure is large. At this time, the sensitivity of network self-healing should be reduced, and the packet loss rate tolerance threshold should be increased to avoid overreaction to periodic fluctuations. When the gas-liquid stiffness coefficient is small, it means that the heave compensation device may have abnormalities, and the mechanical motion is unstable. The fluctuations of link quality may be caused by real failures, and at this time, the sensitivity of network self-healing should be improved, and the packet loss rate tolerance threshold should be reduced to timely discover and handle real failures.

[0133] The mapping relationship adopts a linear inverse proportional function, and the packet loss rate tolerance threshold is equal to the maximum threshold minus the gas-liquid stiffness coefficient multiplied by the threshold adjustment range. Through this mapping relationship, the change of the gas-liquid stiffness coefficient can dynamically adjust the sensitivity of the network self-healing strategy, realizing the deep coupling of the mechanical layer state and the network layer strategy, and enabling the network topology self-healing to adaptively adjust according to the real-time changes of the physical environment.

[0134] When the gas-liquid stiffness coefficient increases, the packet loss rate tolerance threshold is increased to reduce the sensitivity of network self-healing, avoiding overreaction to link fluctuations caused by stable periodic motion. The increase of the gas-liquid stiffness coefficient usually occurs when the gas pressure of the accumulator rises, such as after the system completes the inflation maintenance, or when the sea conditions change from bad to calm, and the external load changes decrease. At this time, the compensation ability of the heave compensation device is enhanced, and the motion of the hydraulic cylinder is more stable and regular, and the periodic characteristics of mechanical motion are more obvious.

[0135] In this case, the periodic fluctuation of link quality is predictable and manageable, and the phase-locked route freezing strategy can effectively cope with it, without frequent triggering of topology reconstruction. The specific implementation of increasing the packet loss rate tolerance threshold is to calculate a new threshold value according to the established mapping relationship. The adjustment of the threshold is realized by modifying the configuration parameter in the network protocol module, which is referenced in the link quality monitoring function to determine whether to trigger a link failure alarm. After the threshold is increased, the system's tolerance to packet loss is increased, and a link failure will only be determined when the continuous packet loss rate exceeds the higher threshold, reducing false failure misjudgments.

[0136] At the same time, the increase of the threshold will also affect the setting of the route aging time, and the aging time will be correspondingly prolonged, giving the route a longer survival period and avoiding the deletion of still valid routes due to temporary communication interruption. The effect of reducing the network self-healing sensitivity is to reduce the number of unnecessary topology reconstruction, reduce the network control overhead and node energy consumption, and improve the stability of the network, so that the system can make more full use of the phase-locked mechanism to cope with periodic interference, maximize resource utilization efficiency on the premise of ensuring communication quality.

[0137] When the gas-liquid stiffness coefficient decreases, the packet loss rate tolerance threshold is reduced to increase the sensitivity of network self-healing and timely discover and handle possible real faults. The decrease of the gas-liquid stiffness coefficient usually occurs when the accumulator gas pressure decreases, and the possible reasons include gas leakage, hydraulic oil leakage, aging of the sealing element, or long-term operation of the system without maintenance, etc. At this time, the compensation capacity of the heave compensation device decreases, and the movement of the hydraulic cylinder may be irregular, and the periodicity of mechanical movement is weakened or even disappears.

[0138] In this case, the fluctuation of link quality no longer completely follows the predictable periodicity, and may contain real link interruption caused by mechanical failure, the effectiveness of the phase-locked route freezing strategy is reduced, and the sensitivity of network self-healing needs to be increased to actively perform topology reconstruction. The specific implementation of reducing the packet loss rate tolerance threshold is also based on the mapping relationship to calculate a new threshold value. The adjustment of the threshold takes effect immediately, and the link quality monitoring function uses the new threshold to determine the link failure, and when the continuous packet loss rate exceeds the lower threshold, the link failure alarm and topology reconstruction process are triggered more quickly.

[0139] After the threshold is reduced, the system responds more quickly to link quality degradation, can find problems and take measures at the initial stage of failure, and avoid the expansion of the failure to affect more nodes. At the same time, the route aging time is shortened accordingly, and the long-time un-updated route is cleaned up faster, keeping the accuracy of the routing table. The effect of improving the network self-healing sensitivity is to enhance the detection and response ability of the system to real faults, greatly shorten the fault discovery time, improve the timeliness of topology reconstruction, and the network can recover to normal communication faster, reduce the impact of failure on monitoring function, and improve the reliability and security of the system.

[0140] The present embodiment realizes the quantitative evaluation of the working state of the mechanical system by calculating the equivalent stiffness of the heave compensation device and normalizing it to the gas-liquid stiffness coefficient. The inverse proportional mapping relationship between the gas-liquid stiffness coefficient and the network self-healing sensitivity threshold is established, realizing the adaptive cooperation of the mechanical layer and the network layer, so that the network topology self-healing strategy can be dynamically adjusted according to the real-time changes of the physical environment. When the gas-liquid stiffness coefficient increases, the network self-healing sensitivity is reduced by increasing the packet loss rate tolerance threshold, avoiding excessive reaction to stable periodic motion, significantly reducing the number of unnecessary topology reconstruction, and greatly reducing the network control overhead and node energy consumption. When the gas-liquid stiffness coefficient decreases, the network self-healing sensitivity is improved by reducing the packet loss rate tolerance threshold, enhancing the detection ability of real faults, significantly shortening the fault discovery time, greatly improving the fault response speed, and significantly improving the system reliability. The adaptive adjustment mechanism enables the network topology self-healing strategy to achieve the best balance between stability and response speed, maintaining network stability in normal working state, reducing unnecessary overhead, and responding quickly to abnormal state and handling faults in time. The overall scheme realizes cross-layer collaborative optimization, introduces physical state information of the mechanical system into network protocol decision-making, breaks the limitation of traditional network protocol relying only on network layer information, and significantly improves the adaptability and robustness of the distributed submarine monitoring system in complex marine environment, providing a strong guarantee for the long-term stable operation of the system.

[0141] In one of the embodiments of the present embodiment, when the real-time gas pressure value decreases and the decrease amplitude exceeds the preset pressure abnormal threshold, it is determined that the heave compensation device has a mechanical failure, and an ultra-fast self-healing strategy is executed to cancel all route freezing restrictions, reduce the packet loss detection threshold, and trigger network topology reconstruction, including the following steps: S810, determining the change rate of the real-time gas pressure value of the accumulator; S820, triggering a mechanical failure alarm when the decrease amplitude of the real-time gas pressure value within the preset time window exceeds the pressure abnormal threshold; S830, clearing all route freezing window settings and canceling the phase-locked route freezing strategy; S840, adjust the packet loss detection threshold to a preset minimum value, so that the loss of a single packet is determined as node failure; S850, start the network topology reconstruction process of the highest priority, and search for a redundant communication path; S860, transmit the mechanical failure alarm information through the reconstructed network topology based on the redundant communication path.

[0142] The rate of change of the real-time gas pressure value of the accumulator is a key indicator for detecting mechanical failure. The pressure change rate reflects the speed of change of the accumulator gas pressure over time, and can timely discover abnormal pressure decline trends. The change rate is calculated using a sliding window difference method. First, a time series buffer of pressure data is established, which stores the collected pressure data in the recent period. The time window length is selected as the reference time period for change rate calculation, which can capture the rapid change of pressure and filter out the interference of transient fluctuations.

[0143] When calculating the change rate, the pressure value at the current time and the pressure value before the time window are read, and the difference between the two is divided by the time interval to obtain the average change rate. To improve the accuracy of change rate calculation, a multi-point averaging method is used, which takes several consecutive sampling points at the current time and the window time, respectively, to calculate the average values, and then differentiates the two average values to reduce the influence of single-point measurement error. The sign of the change rate indicates the direction of pressure change, a positive value indicates pressure rise, a negative value indicates pressure drop, and the absolute value indicates the degree of change.

[0144] Under normal working conditions, the accumulator gas pressure fluctuates periodically with the extension and retraction of the hydraulic cylinder, and the absolute value of the change rate is relatively small and regular. When serious mechanical failures such as gas leakage, large hydraulic oil leakage, or pipeline rupture occur, the pressure will continuously and rapidly decrease, and the absolute value of the change rate will significantly increase. By monitoring the pressure change rate in real time, abnormal trends can be discovered in the early stages of failure, providing time for timely response measures. The change rate data is also used for trend prediction. By linearly fitting the change rates of multiple time windows, the pressure value at the future time is predicted. If the prediction result shows that the pressure will drop below the safety lower limit, an early warning is triggered to inform the operator to check and maintain.

[0145] When the real-time gas pressure value decreases by more than the pressure anomaly threshold within the preset time window, a mechanical failure alarm is triggered, and an emergency response process is started. The pressure anomaly threshold is set based on the safety operation requirements of the heave compensation device and the failure feature analysis. Through statistical analysis of historical failure data, the threshold parameter that can distinguish between normal fluctuations and abnormal decreases is determined. The preset time window covers several heave periods, which can effectively distinguish between periodic fluctuations and sustained decreases. Within the time window, the total decrease in pressure is calculated, which is the pressure value at the beginning of the window minus the pressure value at the end of the window.

[0146] The decision logic uses double conditions, one is that the pressure decrease exceeds the set threshold, and the other is that the absolute value of the pressure change rate exceeds the set threshold. Both conditions must be met at the same time to trigger the mechanical failure alarm, avoiding false alarms caused by transient interference or measurement errors. After the mechanical failure alarm is triggered, a failure event record is immediately generated, including detailed parameters such as failure time, pressure decrease, pressure change rate, and current pressure value, and is stored in a local log file. At the same time, the audible and visual alarm device is activated, and an alarm signal is sent to the on-site operator through the buzzer and LED indicator light.

[0147] The failure alarm information is sent to the monitoring center through the network, and the alarm message uses the highest priority to ensure timely delivery. The contents of the alarm information include failure type identification, failure level, device number, geographic location, and recommended treatment measures, etc. fields, providing comprehensive failure information for remote monitoring personnel. The triggering of the mechanical failure alarm indicates that the working state of the heave compensation device has a serious abnormality, and the original phase-locked route freezing strategy is no longer applicable, and needs to be immediately switched to the ultra-fast self-healing mode to ensure that network communication can continue to run under mechanical failure conditions, and to ensure the transmission of monitoring data and alarm information.

[0148] All route freezing window settings are cleared and the phase-locked route freezing strategy is canceled, restoring the network protocol to a continuous running state. In normal working mode, the phase-locked route freezing strategy periodically opens and closes the route freezing window according to the phase change of the heave period, freezes route maintenance activities in the high resistance state phase interval, and restores normal communication in the low resistance state phase interval. However, when a mechanical failure occurs, the motion law of the heave compensation device is disrupted, and the hydraulic cylinder may experience abnormal rapid sinking, irregular swinging, or complete failure, etc. Mechanical movement is no longer predictable and periodic, and the phase-locked strategy has lost its application basis.

[0149] The operation of clearing the route freezing window setting includes stopping all phase-locked related timers, including the route freezing window opening timer, the closing timer and the data transmission window timer, to prevent these timers from continuing to trigger freezing or unfreezing operations in the fault state. The set route table locking flag is cleared, which is forcibly set to false to ensure that the route protocol can update the route table at any time. The message sending pause restriction is cancelled to restore the normal sending of route discovery messages and link state advertisement messages, so that the network can timely perceive the topology change and perform route update. The route invalidation record in the to-be-verified state is cleared, and invalidation processing is directly performed on all to-be-verified routes to delete these route items and trigger re-discovery, avoiding delay of topology reconstruction due to continued waiting for verification in the fault state.

[0150] After the phase-locked strategy is cancelled, the network protocol switches to the traditional continuous operation mode, and no longer performs periodic freezing and unfreezing according to the mechanical motion phase, but continuously performs link quality monitoring, route maintenance and topology management. In this mode, the network responds to link quality changes more timely and can quickly adapt to the dramatic changes in the communication environment caused by mechanical failure, but at the same time, it will also increase the control message overhead and energy consumption, which is the price paid to ensure communication reliability in the fault emergency state.

[0151] The packet loss detection threshold is adjusted to a preset minimum value, so that the continuous loss of a single data packet is determined as node failure, greatly improving the sensitivity of network self-healing. In the normal operation mode, the packet loss detection threshold is adjusted adaptively according to the gas-liquid stiffness coefficient, and is usually set to determine link failure after the continuous loss of several data packets. This setting can tolerate a certain degree of packet loss and avoid frequent triggering of topology reconstruction due to temporary communication interruption. However, in the mechanical failure state, the communication environment is extremely harsh and unpredictable, and the position and posture of the sensor node may change dramatically, and the link quality may deteriorate rapidly or even be completely interrupted in a short time.

[0152] In this case, a very high fault detection sensitivity is required to timely discover link failure and start topology reconstruction to avoid long-term interruption of communication due to delayed response. The preset minimum value is set to 1, that is, the continuous loss of a single data packet is determined as node failure or link interruption. The implementation of this setting is to modify the packet loss counter threshold parameter in the network protocol module. The packet loss detection mechanism uses a sliding window method to maintain a sequence number queue of sent data packets and the corresponding confirmation state. After sending a data packet, a confirmation timeout timer is started. If the confirmation message from the receiving end is not received within the timeout time, it is determined that the data packet is lost, and the packet loss counter is incremented by one. When the packet loss counter reaches the threshold, link failure determination is triggered immediately, a route invalidation alarm is generated, the corresponding route item is deleted, and the route discovery process is started to find an alternative path.

[0153] The extremely low packet loss detection threshold makes the network extremely sensitive to link quality degradation, enabling fault detection and topology reconstruction initiation within a short time after the failure occurs, thereby minimizing communication interruption time. This high sensitivity setting is necessary in emergency situations of mechanical failure, ensuring that critical fault alarm information and monitoring data can be transmitted in a timely manner through a quickly reconstructed network topology, thereby ensuring the safety monitoring function of the system.

[0154] The highest priority network topology reconstruction process is started and the redundant communication path is searched to quickly establish a stable and reliable communication network. Network topology reconstruction refers to the process of re-discovering the network topology and establishing new routing paths after detecting node failure or link interruption. In the mechanical failure state, due to the poor communication environment and the possibility of multiple links failing simultaneously, the difficulty and urgency of topology reconstruction are significantly increased. The highest priority topology reconstruction process uses an active full-network flooding routing discovery algorithm. The gateway node acts as the initiator of topology reconstruction and broadcasts a routing request message to all neighbor nodes. The message contains information such as reconstruction sequence number, source node address, hop limit, and timestamp.

[0155] The node receiving the routing request immediately processes it. First, it checks whether the same sequence number has been processed before. If so, it is discarded to avoid duplicate processing. If not, it increments the hop count, updates the path information, and then forwards the request to all neighbor nodes. During the forwarding process, each node records the source node of the request and establishes a reverse route to prepare for the subsequent routing response message forwarding. The routing request message spreads throughout the network and eventually reaches all reachable nodes. Each node generates a routing response message upon receiving the request and sends it back to the gateway node along the reverse path. The gateway node collects all routing responses, analyzes the network topology, and identifies all available communication paths.

[0156] The search for redundant communication paths is an important goal of topology reconstruction. Redundant paths refer to multiple non-intersecting or partially intersecting paths to the same target node. When the primary path fails, it can quickly switch to the backup path, improving communication reliability. The search algorithm uses a multi-path routing protocol to maintain multiple path records for each target node in the routing table, including the primary path and at least one backup path. When selecting a path, multiple indicators such as hop count, link quality, and node remaining energy are considered. The path with fewer hops, better link quality, and sufficient node energy is preferred as the primary path, and the path with low overlap with the primary path nodes is selected as the backup path to ensure the independence of the primary and backup paths. After topology reconstruction is complete, the network establishes a new routing table containing the primary and backup paths to all reachable nodes, providing a reliable routing foundation for subsequent data transmission.

[0157] The mechanical fault alarm information is transmitted through the reconstructed network topology based on the redundant communication path, ensuring that the key alarm information can be reliably delivered to the monitoring center. The mechanical fault alarm information is the core data of system safety monitoring, containing important content such as fault type, fault parameter, device state and suggested measures, and must be transmitted to the monitoring center with the highest reliability and shortest delay, so that the operator can take timely response measures. The transmission of alarm information adopts a multi-path concurrent transmission strategy, using the searched redundant communication path to send copies of the alarm message through the main path and backup path at the same time, improving the transmission success rate.

[0158] The alarm message adopts a special message type identifier and has the highest processing priority in the network protocol stack. Each relay node forwards the alarm message immediately after receiving it without queuing. A redundancy check code is added during message encapsulation, and a cyclic redundancy check algorithm is used to generate a checksum. The receiving end verifies the message integrity and detects bit errors in the transmission process by checking the checksum. To further improve reliability, the alarm message uses an acknowledgement retransmission mechanism. The sending end starts an acknowledgement timer after sending the message. If no acknowledgement message is received from the monitoring center within the timeout period, the message is retransmitted through the backup path.

[0159] The monitoring center sends an acknowledgement message immediately after receiving the alarm message. The acknowledgement message is also transmitted through multiple paths to ensure that the sending end can receive it. After the alarm information is transmitted, the monitoring software at the monitoring center parses the alarm content and displays the fault alarm on the monitoring interface, including information such as the location of the faulty device, fault type, severity and handling suggestions. At the same time, audible and visual alarms and SMS notifications are triggered to ensure that the on-duty personnel can promptly learn about the fault situation. Through multi-path concurrent transmission of redundant communication paths, even in a harsh communication environment caused by mechanical failure, the reliable delivery of key alarm information is still guaranteed, providing timely and accurate information support for emergency handling of faults.

[0160] The present embodiment realizes early detection of mechanical failure by monitoring the change rate of the accumulator gas pressure in real time, greatly shortens the fault detection time, and gains valuable time for emergency response. The dual judgment conditions of pressure drop amplitude and change rate accurately identify mechanical failure and avoid false alarms caused by normal fluctuations. Timely removal of the route freezing window setting and cancellation of the phase locking strategy enable the network to quickly switch to the emergency operation mode, ensuring that the network can immediately adapt to the communication environment under the fault state. Reducing the packet loss detection threshold to the minimum value greatly improves the fault detection sensitivity, significantly shortens the link failure detection time, and greatly improves the topology reconstruction response speed. The highest priority topology reconstruction process and the redundant path search mechanism complete the network topology reconstruction in a short time, establish a redundant routing table containing the main path and the backup path, and significantly improve the network reliability. The multi-path concurrent transmission strategy based on the redundant communication path ensures the high reliable transmission of the mechanical failure alarm information, and even in the extreme case of multiple link failures, the reliable delivery of critical information is still guaranteed. The execution of the ultra-fast self-healing strategy enables the network to quickly recover the communication capability after the mechanical failure, greatly shortens the communication interruption time, and significantly improves the fault tolerance and emergency response capability of the system, providing reliable protection for the continuous and safe operation of the distributed submersible monitoring system under the condition of mechanical failure, and effectively preventing the monitoring blind area and safety hazards caused by communication interruption.

[0161] Reference Figure 2 The present application also provides a distributed submersible monitoring system, comprising: A heave compensation device comprising a hydraulic cylinder and an accumulator, the hydraulic cylinder being connected to the accumulator through a hydraulic pipeline, and the accumulator being used to provide pressure compensation for the hydraulic cylinder; A distributed sensor node installed on the heave compensation device, used to collect real-time mechanical parameters of the heave compensation device, the real-time mechanical parameters including real-time stroke data of the hydraulic cylinder, real-time extension and contraction speed of the hydraulic cylinder, and real-time gas pressure value of the accumulator; A network gateway node wirelessly connected to the distributed sensor node, used to execute the network topology self-healing method of the distributed submersible monitoring system.

[0162] The present application also provides a machine-readable storage medium having instructions stored thereon, the instructions being used to cause a machine to execute the network topology self-healing method of the distributed submersible monitoring system.

[0163] Those skilled in the art will appreciate that embodiments of the application can be readily used as a method, a system or a computer program product. Accordingly, the application can take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the application can take the form of a computer program product on one or more computer readable storage media (including, but not limited to, disk memory, CD-ROMs, optical storage devices, etc.) embodying computer readable program code.

[0164] The application is described herein with reference to the Figures, which illustrate the described embodiments. The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks or in conjunction with the flowcharts described herein. Figure 1 one or more flowcharts and / or blocks can represent code, circuits, processes, procedures, modules, segments or the like which can implement one or more specifications Figure 1 one or more flowcharts and / or blocks can represent code, circuits, processes, procedures, modules, segments or the like which can implement one or more specifications

[0165] The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks or in conjunction with the flowcharts described herein. Figure 1 one or more flowcharts and / or blocks can represent code, circuits, processes, procedures, modules, segments or the like which can implement one or more specifications Figure 1 one or more flowcharts and / or blocks can represent code, circuits, processes, procedures, modules, segments or the like which can implement one or more specifications

[0166] The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks or in conjunction with the flowcharts described herein. Figure 1 one or more flowcharts and / or blocks can represent code, circuits, processes, procedures, modules, segments or the like which can implement one or more specifications Figure 1 one or more flowcharts and / or blocks can represent code, circuits, processes, procedures, modules, segments or the like which can implement one or more specifications

[0167] In a typical configuration, a computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory.

[0168] The memory can include non-persistent memory and / or persistent memory, such as flash memory, read-only memory (ROM), and / or the like, in a computer readable medium. The memory is an example of computer readable media.

[0169] Computer-readable media includes permanent and non-permanent, movable and non-movable media that can be implemented by any method or technology to store information. The information can be computer-readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassette, magnetic tape disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information accessible by a computing device. According to the definition herein, computer-readable media does not include transitory media such as modulated data signals and carriers.

[0170] It should also be noted that the terms "comprising", "containing", or any other variant thereof are intended to cover non-exclusive inclusions, so that a process, method, article or apparatus that includes a list of elements does not only include those elements, but also includes other elements not explicitly listed, or further includes elements inherent in such a process, method, article or apparatus. Without more limitations, the element defined by the statement "comprising a" does not exclude the presence of additional identical elements in the process, method, article or apparatus that includes the element.

[0171] The above is only an embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application can have various modifications and changes. Any modification, equivalent replacement, improvement, etc. within the spirit and principle of the present application shall be included in the scope of claims of the present application.

Claims

1. A network topology self-healing method for a distributed underwater monitoring system, characterized in that, An application in a distributed diving monitoring system, the distributed diving monitoring system including a heave compensation device and distributed sensor nodes, the heave compensation device including a hydraulic cylinder and an accumulator, the method comprising: The real-time mechanical parameters of the heave compensation device are obtained, including the real-time stroke data of the hydraulic cylinder, the real-time extension and retraction speed of the hydraulic cylinder, and the real-time gas pressure value of the accumulator. Based on real-time travel data and real-time extension / retraction speed, the link quality variation law of the distributed sensor nodes of the heave compensation device under different phases of the heave cycle is determined. Based on real-time scaling speed and real-time travel data, the physical interference coefficient of the distributed sensor nodes is calculated, and the high-resistivity phase interval and low-resistivity phase interval in the heave cycle are identified according to the physical interference coefficient. The high-resistivity phase interval corresponds to the period of pseudo-deterioration of communication link quality, and the low-resistivity phase interval corresponds to the period of normal communication link quality. Within the high-impedance phase interval, a phase-locked route freezing strategy is implemented to suspend the transmission of all route discovery messages, ignore route failure alarms, and cache the data to be transmitted locally. Within the low-impedance phase interval, an unfreezing strategy is executed to release the route from the frozen state and concentrate burst transmission of locally cached data; In response to network topology update requests, monitor the real-time gas pressure value of the accumulator; When the real-time gas pressure value drops and the drop exceeds the preset pressure anomaly threshold, it is determined that the heave compensation device has a mechanical failure. The rapid self-healing strategy is then executed to cancel all route freeze restrictions, lower the packet loss detection threshold, and trigger network topology reconstruction.

2. The method according to claim 1, characterized in that, Based on real-time travel data and real-time extension / retraction speed, the link quality variation law of the distributed sensor nodes of the heave compensation device under different phases of the heave cycle is determined, including: The heave period of the heave compensation device is determined based on the periodic change of the real-time extension and retraction speed. During the heave-descent cycle, signal strength data from distributed sensor nodes are collected; A time-series correlation model is constructed between real-time travel data, real-time scaling speed, and signal strength data to determine the link quality variation law of the distributed sensor nodes of the heave compensation device under different phases of the heave cycle.

3. The method according to claim 1, characterized in that, Based on real-time scaling speed and real-time travel data, the physical interference coefficient of the distributed sensor nodes is calculated, including: The absolute value of the real-time scaling speed is compared with a preset speed threshold; When the absolute value of the real-time scaling speed is greater than the speed threshold, it is determined that the current state is high-speed motion, and the frequency shift influence factor is calculated. Based on real-time stroke data, determine whether the current position of the hydraulic cylinder is within the metal obstruction zone; Calculate the cutting influence factor when the current position of the hydraulic cylinder is in the metal obstruction zone; The coupling degree of the frequency shift influence factor and the cutting influence factor is analyzed, and the synergistic effect coefficient of the two is calculated. Based on the synergistic effect coefficient, determine whether the frequency shift influence factor and the cutting influence factor are simultaneously in an active state; When both the frequency shift influence factor and the cut influence factor are activated, their product is calculated as the composite interference intensity. When only one influencing factor is active, the value of that active influencing factor is taken as the single interference intensity. The physical interference coefficient is obtained by normalizing the composite interference intensity or the single interference intensity through a nonlinear mapping function.

4. The method according to claim 1, characterized in that, Based on the physical interference coefficient, the high-resistivity phase interval and the low-resistivity phase interval in the heave cycle are identified, including: The heave cycle is divided into multiple phase segments, and the average value of the physical disturbance coefficient within each phase segment is calculated. The average value of the physical interference coefficient is compared with the preset interference threshold; When the average value of the physical interference coefficient is higher than the preset interference threshold, the corresponding phase segment is marked as the high-resistivity phase interval. When the average value of the physical interference coefficient is lower than or equal to the preset interference threshold, the corresponding phase segment is marked as the low-resistance phase interval.

5. The method according to claim 1, characterized in that, Implement the phase-locked route freezing policy, including: Calculate the opening and closing times of the route freezing window based on the heave period and the initial phase of the high-resistivity phase interval; At the moment the route freeze window opens, set the route table lock flag at the network routing protocol layer; Suspend the transmission of all messages and link status announcement messages; Mark all received route failure alarms as pending verification and do not trigger the topology reconfiguration process; Store the currently pending data packet to the local cache queue; Clear the routing table lock flag when the routing freeze window closes.

6. The method according to claim 5, characterized in that, Implement the unfreezing strategy, including: Calculate the opening time and duration of the data transmission window based on the heave period and the initial phase of the low-resistivity phase interval; At the moment the data transmission window opens, clear the routing table lock flag and restore the normal operation of the routing protocol; Data packets to be transmitted are read from the local cache queue according to priority and transmitted in bursts. Receive and process route discovery messages and topology update requests sent by neighboring nodes; For route failure alarms marked as pending verification in the high-impedance phase interval, a secondary verification is performed. If communication still cannot be restored in the low-impedance phase interval, it is determined to be a real fault.

7. The method according to claim 1, characterized in that, After monitoring the real-time gas pressure value of the accumulator, the following is also included: Calculate the equivalent stiffness of the heave compensation device based on the real-time gas pressure value; The equivalent stiffness is normalized to obtain the gas-liquid stiffness coefficient; Establish an inverse proportional mapping relationship between the gas-liquid stiffness coefficient and the network self-healing sensitivity threshold; When the gas-liquid stiffness coefficient increases, the packet loss rate tolerance threshold is increased, and the sensitivity of network self-healing is reduced. When the gas-liquid stiffness coefficient decreases, the packet loss rate tolerance threshold is reduced, and the network's self-healing sensitivity is improved.

8. The method according to claim 1, characterized in that, When the real-time gas pressure value drops and the drop exceeds the preset pressure anomaly threshold, a mechanical failure is determined in the heave compensation device. A rapid self-healing strategy is then executed to remove all route freeze restrictions, lower the packet loss detection threshold, and trigger network topology reconstruction, including: Determine the rate of change of the real-time gas pressure value of the accumulator; When the real-time gas pressure value drops beyond the pressure anomaly threshold within a preset time window, a mechanical fault alarm is triggered. Clear all route freeze window settings and cancel the phase lock route freeze policy; Adjust the packet loss detection threshold to a preset minimum value so that the node is determined to be faulty if a single data packet is lost consecutively. Initiate the highest priority network topology reconfiguration process and search for redundant communication paths; Based on redundant communication paths, mechanical fault alarm information is transmitted through the reconstructed network topology.

9. A distributed diving monitoring system, characterized in that, include: The heave compensation device includes a hydraulic cylinder and an accumulator. The hydraulic cylinder and the accumulator are connected by a hydraulic pipeline. The accumulator is used to provide pressure compensation for the hydraulic cylinder. Distributed sensor nodes are installed on the heave compensation device to collect real-time mechanical parameters of the heave compensation device. The real-time mechanical parameters include real-time stroke data of the hydraulic cylinder, real-time extension and retraction speed of the hydraulic cylinder, and real-time gas pressure value of the accumulator. A network gateway node, wirelessly connected to distributed sensor nodes, is used to perform the network topology self-healing method of the distributed diving monitoring system as described in any one of claims 1 to 8.

10. A machine-readable storage medium, characterized in that, The machine-readable storage medium stores instructions for causing the machine to perform the network topology self-healing method of the distributed diving monitoring system according to any one of claims 1 to 8.

Images