Dual-mode driven fishing vessel night intelligent supervision and scheduling system and method

Abstract

The application relates to the technical field of supervised prediction data processing, and discloses a dual-mode driven intelligent supervision and scheduling system and method for fishing vessels at night, which comprises the following steps: a data state sensing unit acquires first and second state data sequences, and extracts spatially measured observation coordinate data of the second sequence within an observation time window when it is determined that the first sequence is interrupted; a divergence arbitration unit generates a space-time residual integral based on the coordinate data and the track prediction data of the first sequence; and a dynamic scheduling management unit injects the integral as an adjustment scalar into a spatial topology function, updates a task occupation label of an available resource node, and outputs a directional management task routing table. The application quantifies the time sequence logical divergence of heterogeneous data streams, eliminates environmental false alarm noise interference, and optimizes resource allocation efficiency.

Description

Technical Field

[0001] This invention belongs to the field of supervised predictive data processing technology, and in particular relates to a dual-modal driven intelligent monitoring and scheduling system and method for fishing vessels at night. Background Technology

[0002] The current maritime traffic administration and supervision system generally adopts a dual-verification architecture consisting of a cooperative communication network and an active detection network to monitor the compliance status of fishing vessels and allocate corresponding law enforcement resources. It determines the legal operation status of target groups by receiving real-time identity displacement messages of surface targets and comparing them with radar cross section reflection signals. The spatiotemporal correlation of multi-source heterogeneous trajectory data streams depends on the logical continuity of dual-modal information in the temporal evolution. Once the cooperative communication network experiences link interruption or signal shielding, the system's perception of the target's compliance status completely degrades to the non-cooperative active detection data source. The dynamic modulation of the radar echo cross section characteristics under complex sea conditions at night by sea wave clutter and meteorological noise results in strong random fluctuations in the signal-to-noise ratio.

[0003] Faced with the pressure of intensive monitoring in low-visibility waters at night, when fishing vessels abnormally disconnect from the cooperative network, the system, lacking continuous cooperative messages, misinterprets short-term radar false alarms caused by wave surges as non-cooperative avoidance navigation events. This leads the command and control center to issue invalid interception and verification instructions to various surface law enforcement units. This not only leaves the scarce nighttime frontline patrol capacity in a state of frantic and blind dispatching, but also, due to the lack of a deep-seated, time-series logical differentiation mechanism within the system, allows genuinely intentional illegal fishing activities to remain hidden among a massive number of sporadic false alarms, reducing the effectiveness of management agencies in... Monitoring the efficiency of decision-making in the workflow increases the systemic technical burden and management cost of administrative resources in the dynamic allocation and precise scheduling stages. To address these shortcomings, conventional improvement attempts usually focus on improving single-point physical indicators, such as increasing the high-frequency gain of radar receiving antennas or setting multi-level static correlation filtering thresholds to eliminate environmental noise. However, the non-stationary evolution of the energy spectrum of dynamic sea surface clutter limits such linear parameter fine-tuning. Blindly increasing the filtering standards leads to the direct loss of real weak target tracks, while relaxing the threshold causes an exponential increase in false alarms. Combining this with conventional open-loop smoothing algorithms is difficult to be effective. Distinguishing between transient step jumps caused by physical clutter and structural trajectory spoofing caused by ship evasive maneuvers ultimately leads to a tradeoff between detection probability and scheduling false alarm rate. Besides the aforementioned improvements targeting single-point physical indicators, multi-source data fusion control methods also have shortcomings. For example, Chinese invention patent application CN119942286A discloses a non-cooperative ship trajectory dynamic completion and monitoring system that fuses remote sensing imagery and AIS data. This system relies on acquiring clear, continuous images with highly recognizable ship hull features and performs multi-frame image feature comparison based on a high-efficiency revisit mechanism. However, in… In low visibility conditions at night and in complex sea conditions with severe winds and waves, optical remote sensing imaging is limited. Meanwhile, radar images such as synthetic aperture radar are highly susceptible to strong modulation by non-stationary dynamic clutter on the sea surface, causing large-scale random distortion and edge fragmentation of the two-dimensional image features of non-cooperative vessels. The matching logic that relies on the overlap of two-dimensional features or texture comparison of such images is substantially invalidated at the underlying mechanism. The inherent long orbit revisit cycle of remote sensing satellites cannot provide high-frequency trajectory supplementation within a tight spatiotemporal window, resulting in open-loop and lagging characteristics in the resource response path. This makes it impossible to adapt to the precise on-demand dispatching and scheduling under the condition of frequent transient false alarms at night.

[0004] Therefore, the technical problem to be solved by this invention is how to design a data processing and topology reconstruction method to extract the spatiotemporal cumulative deviation between the measured trajectory and the virtual inferred trajectory after the cooperative network message flow is interrupted, and to achieve optimal matching of law enforcement resources by dynamically adjusting the scheduling cost parameter. Summary of the Invention

[0005] This invention aims to solve the problems of frequent false alarms and the blind and idle scheduling of administrative law enforcement resources caused by the lack of a time-series deviation product discrimination mechanism in traditional regulatory systems when multimodal data is interrupted.

[0006] In this technical solution, a dual-modal driven intelligent monitoring and dispatching system for fishing vessels at night includes a cooperative modal data access unit, a non-cooperative modal data parsing unit, a data status perception unit, a disagreement arbitration unit, and a dynamic dispatching management unit. The data status awareness unit is interconnected with the cooperative mode data access unit and the non-cooperative mode data parsing unit via a data bus. The disagreement arbitration unit is associated with the data status awareness unit via a data path, and the dynamic scheduling management unit is associated with the disagreement arbitration unit via a signal path. The system performs data scheduling management on globally available managed resource nodes through the following timing control steps: Step 101: The data state perception unit acquires the first state data sequence output by the cooperative mode data access unit and the second state data sequence output by the non-cooperative mode data parsing unit. Step 102: When it is determined that the signal of the first state data sequence has been interrupted, the data state sensing unit extracts the spatial measured observation coordinate data of the second state data sequence within the continuous observation time window. Step 103: The disagreement arbitration unit performs difference calculation based on the space measured observation coordinate data and the trajectory prediction data of the first state data sequence to generate the spatiotemporal residual integral. Step 104: The dynamic scheduling management unit injects the spatiotemporal residual integral as a dynamic adjustment scalar into the target response space topology function for the target regulated entity, updates the task occupancy label of the globally available managed resource nodes, and outputs the directional management task routing table.

[0007] Preferably, in step 102, the data state sensing unit is used to extract spatial measured observation coordinate data through the following sub-steps: Step 1021, when the spatial coordinates and heading vector in the first state data sequence are missing for three consecutive sampling periods, it is determined that a signal interruption has occurred; Step 1022, within the continuous observation time window, the measured scattered echo coordinates in the second state data sequence on the coaxial time axis are continuously retrieved as spatial measured observation coordinate data.

[0008] Preferably, in step 103, the disagreement arbitration unit is used to calculate the spatiotemporal residual integral through the following sub-steps: Step 1031, obtain the spatial coordinates and heading vector of the last frame before the signal interruption of the first state data sequence, and derive the predicted spatial coordinates of each sampling time within the continuous observation time window through a first-order kinematic extrapolation algorithm, as the trajectory prediction data; Step 1032, calculate the Euclidean distance between the spatial measured observation coordinate data and the trajectory prediction data at the same sampling time, and generate the absolute value of the spatiotemporal residual; Step 1033, at the end of the continuous observation time window, sum the absolute values ​​of all spatiotemporal residuals within the continuous observation time window in the time dimension to obtain the spatiotemporal residual integral.

[0009] Preferably, the disagreement arbitration unit is also used to compare the spatiotemporal residual integral with a safety threshold to determine the temporal logic disagreement degree; the disagreement arbitration unit uses the following calculation formula to determine the safety threshold: ,in, As a safety threshold, This serves as the baseline noise level established based on historical nighttime noise data of the spatial environment. To determine the duration of the continuous observation window, the dynamic scheduling management unit outputs a directional management task routing table in step 104 through the following sub-steps: Step 1041: Extract the current residence coordinates of idle managed resource nodes within the managed area. Under the state increment generated by the spatiotemporal residual integral, the spatiotemporal residual integral is used as a penalty weight and superimposed in a monotonically increasing direction onto the distance function of each idle managed resource node moving towards the target managed entity, thus constructing the target response space topology function; Step 1042: With the minimum total scheduling cost parameter as the convergence optimization objective, solve the target response space topology function, lock the resource pairing combination, and output the directional management task routing table.

[0010] Preferably, before constructing the target response space topology function in step 1041, the dynamic scheduling management unit is further used to filter nodes through the following derived sub-steps: Step 10411, obtain a state matrix containing the remaining range, current movement speed, and personnel on-duty status of each idle managed resource node; Step 10412, exclude idle managed resource nodes that do not meet the timeliness constraints under the constraints of the state matrix, and pass the set of available nodes to step 1041 to update the task occupancy tags of the globally available managed resource nodes.

[0011] Preferably, the dynamic scheduling management unit is also used to determine that the target regulated entity has intentional avoidance characteristics under dynamic response where the spatiotemporal residual integral is greater than the safety threshold, and to increase the scheduling priority of the target regulated entity in obtaining globally available managed resource nodes.

[0012] Preferably, the non-cooperative modal data parsing unit is used to perform spatiotemporal filtering on the incoming raw non-cooperative data, filtering out unstable reflection points with a spatial overlap of less than or equal to 80% within three consecutive scanning cycles, and outputting stable reflection points with a spatial overlap of greater than 80% within three consecutive scanning cycles as spatial measured observation coordinate data to the data status sensing unit.

[0013] Preferably, under the trigger condition that the communication link is restored from the signal interruption state and the first state data sequence is received again, the cooperative mode data access unit is used to calculate the coordinate deviation between the coordinates of the first frame after restoration and the coordinates of the second state data sequence at the same timestamp. When the coordinate deviation is stable within a preset threshold range under causal feedback, the continuous observation time window is closed and the conventional monitoring mode dominated by the first state data sequence is restored.

[0014] Preferably, the system also includes: a scheduling center visualization terminal; the scheduling center visualization terminal is associated with the dynamic scheduling management unit through a data bus, and is used to display the spatial topology network mapping of the directional management task routing table, and to issue digital management scheduling instructions to the target managed resource node terminals pointed to by the directional management task routing table.

[0015] A dual-modal driven intelligent monitoring and scheduling method for fishing vessels at night includes the following steps: Step S1: Obtain the first state data sequence output by the cooperative mode data access unit and the second state data sequence output by the non-cooperative mode data parsing unit; Step S2: When it is determined that the signal of the first state data sequence has been interrupted, extract the spatial measured observation coordinate data of the second state data sequence within the continuous observation time window; Step S3: Based on the space-measured observation coordinate data and the trajectory prediction data of the first state data sequence, calculate and generate the spatiotemporal residual integral. Step S4: Inject the spatiotemporal residual integral as a dynamic adjustment scalar into the target response space topology function for the target regulated entity, update the task occupancy label of the globally available managed resource nodes, and output the directed management task routing table.

[0016] Compared with existing technologies, the dual-modal driven intelligent monitoring and scheduling system and method for fishing vessels at night has the following advantages: 1. In the intelligent monitoring and scheduling of fishing vessels at night, the expected spatial state sequence generated by the cooperative modal data access unit and the measured observation coordinate sequence collected by the non-cooperative modal analysis unit are correlated within the same source time window. When the cooperative modal signal is interrupted, the system extracts the heading vector and velocity scalar of the effective state point before the interruption, and generates a virtual inference trajectory within the observation time window using kinematic extrapolation law. By continuously calculating the spatial straight-line distance between the measured observation coordinate sequence and the virtual inference trajectory, a discrete residual value chain that evolves with the time axis is formed. In this way, the sudden concealment state of the water surface target is converted into a continuous and measurable displacement deviation feature, and the dynamic alignment and state divergence determination of heterogeneous spatial information are completed.

[0017] 2. The multi-source heterogeneous data collected by the system interact in the time series. For non-cooperative transient environmental noise caused by meteorological and sea conditions or normal short-term berthing of ships, the resulting spatial coordinate deviations exhibit random and disordered distribution. The residual integral accumulated within the sliding observation time window remains at a low level. Based on this, the system suppresses the issuance of scheduling commands. However, when a surface target shuts down its cooperative communication equipment and performs directional avoidance maneuvers, the spatial straight-line distance between its corresponding virtual inferred coordinates and actual observed coordinates exhibits a monotonically increasing structural deviation, causing the residual integral to exceed the set verification threshold. This establishes a filtering barrier based on subject intent identification at the control layer, preventing false alarms in the transient environment from interfering with subsequent resource allocation.

[0018] 3. The abnormal scheduling weight label output by the divergence arbitration unit is deeply coupled with the available law enforcement node state matrix loaded by the scheduling resource matching and allocation unit. The system extracts the current stationing coordinates of each idle law enforcement node and introduces the spatiotemporal residual integral as a dynamic adjustment variable into the spatial topology function of the water surface target response. The increasing spatiotemporal residual integral drives the scheduling cost parameter to generate monotonically converged, so that the priority of targets with avoidance characteristics to acquire law enforcement resources is spontaneously increased. Based on this, the pairing combination with the lowest cost parameter is locked to output a directional law enforcement task routing table with unique direction. This realizes the control strategy of switching management resources from passive blind patrol to precise target delivery, and improves the overall efficiency of resource planning, task distribution and workflow scheduling of the data processing system under the purpose of administrative supervision. Attached Figure Description

[0019] Figure 1 This is a flowchart of the intelligent nighttime monitoring and scheduling method for fishing vessels using dual-modal timing control according to the present invention; Figure 2 This is a structural diagram of the nighttime intelligent monitoring and scheduling system for fishing vessels based on the dual-modal bus topology of this invention. Detailed Implementation

[0020] The technical solutions of the embodiments of this application will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of this application are within the scope of protection of this application.

[0021] It should be noted that all directional and positional terms used in this invention, such as: up, down, left, right, front, back, vertical, horizontal, inner, outer, top, bottom, transverse, longitudinal, center, etc., are only used to explain the relative positional relationship and connection between components in a specific state. They are only for the convenience of describing this invention and do not require that this invention be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention. In addition, the descriptions involving "first," "second," etc., in this invention are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated.

[0022] In the description of this invention, unless otherwise explicitly specified and limited, the terms installation, connection, and linking should be interpreted broadly. For example, they can refer to fixed connections, detachable connections, or integral connections; they can refer to mechanical connections; they can refer to direct connections or indirect connections through an intermediate medium; they can refer to the internal connection of two components. For those skilled in the art, the specific meaning of the above terms in this invention can be understood according to the specific circumstances.

[0023] In the description of this specification, references to the terms "an embodiment," "some embodiments," "illustrative embodiments," "examples," "specific examples," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example, and the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0024] A dual-modal driven intelligent monitoring and dispatching system for fishing vessels at night includes a cooperative modal data access unit, a non-cooperative modal data parsing unit, a data status perception unit, a disagreement arbitration unit, and a dynamic dispatching management unit. The data status awareness unit is interconnected with the cooperative mode data access unit and the non-cooperative mode data parsing unit via a data bus. The disagreement arbitration unit is associated with the data status awareness unit via a data path, and the dynamic scheduling management unit is associated with the disagreement arbitration unit via a signal path. The system performs data scheduling management on globally available managed resource nodes through the following timing control steps: Step 101: The data state perception unit acquires the first state data sequence output by the cooperative mode data access unit and the second state data sequence output by the non-cooperative mode data parsing unit. Step 102: When it is determined that the signal of the first state data sequence has been interrupted, the data state sensing unit extracts the spatial measured observation coordinate data of the second state data sequence within the continuous observation time window. Step 103: The disagreement arbitration unit performs difference calculation based on the space measured observation coordinate data and the trajectory prediction data of the first state data sequence to generate the spatiotemporal residual integral. Step 104: The dynamic scheduling management unit injects the spatiotemporal residual integral as a dynamic adjustment scalar into the target response space topology function for the target regulated entity, updates the task occupancy label of the globally available managed resource nodes, and outputs the directional management task routing table.

[0025] Preferably, in step 102, the data state sensing unit is used to extract spatial measured observation coordinate data through the following sub-steps: Step 1021, when the spatial coordinates and heading vector in the first state data sequence are missing for three consecutive sampling periods, it is determined that a signal interruption has occurred; Step 1022, within the continuous observation time window, the measured scattered echo coordinates in the second state data sequence on the coaxial time axis are continuously retrieved as spatial measured observation coordinate data.

[0026] Preferably, in step 103, the disagreement arbitration unit is used to calculate the spatiotemporal residual integral through the following sub-steps: Step 1031, obtain the spatial coordinates and heading vector of the last frame before the signal interruption of the first state data sequence, and derive the predicted spatial coordinates of each sampling time within the continuous observation time window through a first-order kinematic extrapolation algorithm, as the trajectory prediction data; Step 1032, calculate the Euclidean distance between the spatial measured observation coordinate data and the trajectory prediction data at the same sampling time, and generate the absolute value of the spatiotemporal residual; Step 1033, at the end of the continuous observation time window, sum the absolute values ​​of all spatiotemporal residuals within the continuous observation time window in the time dimension to obtain the spatiotemporal residual integral.

[0027] Preferably, the disagreement arbitration unit is also used to compare the spatiotemporal residual integral with a safety threshold to determine the temporal logic disagreement degree; the disagreement arbitration unit uses the following calculation formula to determine the safety threshold: ,in, As a safety threshold, This serves as the baseline noise level established based on historical nighttime noise data of the spatial environment. To determine the duration of the continuous observation window, the dynamic scheduling management unit outputs a directional management task routing table in step 104 through the following sub-steps: Step 1041: Extract the current residence coordinates of idle managed resource nodes within the managed area. Under the state increment generated by the spatiotemporal residual integral, the spatiotemporal residual integral is used as a penalty weight and superimposed in a monotonically increasing direction onto the distance function of each idle managed resource node moving towards the target managed entity, thus constructing the target response space topology function; Step 1042: With the minimum total scheduling cost parameter as the convergence optimization objective, solve the target response space topology function, lock the resource pairing combination, and output the directional management task routing table.

[0028] Preferably, before constructing the target response space topology function in step 1041, the dynamic scheduling management unit is further used to filter nodes through the following derived sub-steps: Step 10411, obtain a state matrix containing the remaining range, current movement speed, and personnel on-duty status of each idle managed resource node; Step 10412, exclude idle managed resource nodes that do not meet the timeliness constraints under the constraints of the state matrix, and pass the set of available nodes to step 1041 to update the task occupancy tags of the globally available managed resource nodes.

[0029] Preferably, the dynamic scheduling management unit is also used to determine that the target regulated entity has intentional avoidance characteristics under dynamic response where the spatiotemporal residual integral is greater than the safety threshold, and to increase the scheduling priority of the target regulated entity in obtaining globally available managed resource nodes.

[0030] Preferably, the non-cooperative modal data parsing unit is used to perform spatiotemporal filtering on the incoming raw non-cooperative data, filtering out unstable reflection points with a spatial overlap of less than or equal to 80% within three consecutive scanning cycles, and outputting stable reflection points with a spatial overlap of greater than 80% within three consecutive scanning cycles as spatial measured observation coordinate data to the data status sensing unit.

[0031] Preferably, under the trigger condition that the communication link is restored from the signal interruption state and the first state data sequence is received again, the cooperative mode data access unit is used to calculate the coordinate deviation between the coordinates of the first frame after restoration and the coordinates of the second state data sequence at the same timestamp. When the coordinate deviation is stable within a preset threshold range under causal feedback, the continuous observation time window is closed and the conventional monitoring mode dominated by the first state data sequence is restored.

[0032] Preferably, the system also includes: a scheduling center visualization terminal; the scheduling center visualization terminal is associated with the dynamic scheduling management unit through a data bus, and is used to display the spatial topology network mapping of the directional management task routing table, and to issue digital management scheduling instructions to the target managed resource node terminals pointed to by the directional management task routing table.

[0033] Preferably, a dual-modal driven intelligent monitoring and scheduling method for fishing vessels at night includes the following steps: Step S1: Obtain the first state data sequence output by the cooperative mode data access unit and the second state data sequence output by the non-cooperative mode data parsing unit; Step S2: When it is determined that the signal of the first state data sequence has been interrupted, extract the spatial measured observation coordinate data of the second state data sequence within the continuous observation time window; Step S3: Based on the space-measured observation coordinate data and the trajectory prediction data of the first state data sequence, calculate and generate the spatiotemporal residual integral. Step S4: Inject the spatiotemporal residual integral as a dynamic adjustment scalar into the target response space topology function for the target regulated entity, update the task occupancy label of the globally available managed resource nodes, and output the directed management task routing table.

[0034] Example 1: In a low-visibility nighttime maritime administrative supervision scenario, multiple regulated entities exist within the jurisdictional waters. The administrative law enforcement dispatch and command center allocates scheduling tasks to all available managed resource nodes globally through a data processing system. When some regulated entities shut down their cooperative communication networks, causing a first-state data sequence signal interruption, the second-state data sequence output by the active detection network is easily affected by transient nighttime noise and dynamic modulation from sea clutter caused by complex sea conditions. This lack of data space mapping leads to a break in the administrative data link, making the mechanical overlay display in the traditional supervision architecture unable to distinguish between transient steps caused by sea surface clutter and structural trajectories caused by vessel escape. This results in the back-end data processing center issuing invalid verification instructions to front-line patrol forces, causing the extremely scarce surface law enforcement nodes to enter a state of blind patrol. When the system faces the aforementioned cooperative communication network link interruption, the data state perception unit... The unit acquires the first state data sequence output by the cooperative mode data access unit and the second state data sequence output by the non-cooperative mode data analysis unit through a data bus interaction connection. Based on the two-dimensional Euclidean space discretization mechanism, after discretizing the continuous spatial lattice, set operations are performed. The non-cooperative mode data analysis unit extracts the spatial coordinates of the received discrete reflection points and converts them into a two-dimensional orthogonal grid array according to the set distance and azimuth resolution. It extracts the corresponding target grid array within the continuous scanning period, calculates the ratio of the number of intersection elements to the number of union elements, and obtains the quantized spatial overlap value. In this conversion process, the above-mentioned set distance and azimuth resolution serves as the basic physical scale for discretization mapping. Its value matches the front-end radar hardware detection system. By extracting the pulse width of the current transmitted pulse of the detection radar, it is converted into a determined 15-meter radial distance resolution. At the same time, the antenna beamwidth is extracted and directly mapped to 0.With a tangential azimuth resolution of 5 degrees, the non-cooperative modal data analysis unit uses this defined basic grid scale to forcibly partition Euclidean space, ensuring that each intersection and union operation is based on the actual physical detection granularity. This provides a deterministic input-output mapping path, avoiding distortion in the calculation of two-dimensional spatial feature overlap caused by resolution divergence. Based on this overlap measurement benchmark, after receiving the original non-cooperative data, the non-cooperative modal data analysis unit uses spatiotemporal filtering to remove unstable reflection points with a spatial overlap of less than or equal to 80% within three consecutive scan cycles, and outputs stable reflection points with a spatial overlap of greater than 80% within three consecutive scan cycles as the spatial measured observation coordinate data. When it is determined that the spatial coordinates and heading vector in the first state data sequence are missing for three consecutive sampling cycles, the data state perception unit determines that a signal interruption has occurred, and then locks the last frame before the interruption. The spatial coordinates and heading vector are based on a first-order kinematic uniform linear model. Physical displacement is the integral of the velocity vector with magnitude and direction over time. The data state sensing unit retrieves the position coordinate sequence of five consecutive sampling periods before the signal interruption, calculates the linear spatial distance between adjacent sampling points, divides it by a fixed sampling time interval to obtain the target scalar average velocity, merges the scalar average velocity and heading vector to establish a complete initial velocity vector. Under the constraint of a complete initial velocity vector, the predicted spatial coordinates at each sampling moment within the continuous observation time window are derived using a first-order kinematic extrapolation algorithm, serving as the trajectory prediction data. Simultaneously, within the continuous observation time window, the spatially measured observation coordinate data in the second state data sequence on the coaxial time axis are continuously retrieved. The divergence arbitration unit calculates the Euclidean distance between the spatially measured observation coordinate data and the trajectory prediction data at the same sampling moment within the continuous observation time window, obtaining the absolute value of the spatiotemporal residual.

[0035] At the end of the continuous observation window, the disagreement arbitration unit sums up the absolute values ​​of all spatiotemporal residuals within the continuous observation window along the time dimension to calculate the spatiotemporal residual integral. The dynamic scheduling management unit outputs a routing table for directed management tasks. It extracts the current resident coordinates of idle managed resource nodes within the managed area and calculates the scheduling cost parameters based on the target response space topology function determined by the spatiotemporal residual integral. The calculation formula is as follows: ,in, To make the first The idle managed resource node is scheduled to the first The scheduling cost parameters of each target are subject to the regulatory entity. For the first The current location coordinates of the idle managed resource nodes up to the [number]th [node]. The absolute spatial distance of each target is determined by the latest measured spatial coordinates of the regulated entity. This represents the total number of discrete sampling points within the continuous observation time window; For the first The absolute value of the spatiotemporal residual formed by the predicted trajectory data and the measured spatial coordinate data at each sampling point. To prevent the system-level minimal constant with a denominator of zero, the scheduling cost parameter is physically a dimensionality-reduced pseudo-overhead variable. Its core transformation mechanism is to use the spatiotemporal residual integral representing the ship's deliberate avoidance of determinism as a mathematical attenuation term for physical distance. Under real sea conditions, the absolute spatial distance of node navigation represents basic energy consumption, while the time loss caused by target avoidance behavior constitutes the implicit sunk cost of scheduling flow. The system uses the above topological function to inversely map the discrete deviation accumulation of micro-trajectories into a dynamic discount weight of macro-distance, so that targets with high confidence in escaping intentions present highly compressed virtual distances in the spatial topological network. This eliminates the dimensional barrier between physical spatial scale and administrative decision-making costs, allowing the capacity matching algorithm to preferentially converge towards it.

[0036] Under the aforementioned target response space topology function constraints, the dynamic scheduling management unit determines that the target regulated entity possesses intentional avoidance characteristics when the spatiotemporal residual integral exceeds the safety threshold. Furthermore, by setting a state matrix containing the remaining range, current movement speed, and personnel on-duty status of each idle managed resource node, it excludes nodes that do not meet the timeliness constraints to filter the available node set. Driven by the penalty weights generated by the spatiotemporal residual integral in a monotonically increasing direction, the scheduling cost parameters corresponding to the target regulated entity with avoidance characteristics monotonically converge, spontaneously increasing the target's regulated status. The entity acquires the scheduling priority of globally available managed resource nodes, and solves the target response space topology function with the goal of minimizing the total scheduling cost parameter to lock resource pairings. It updates the task occupancy labels of available resource nodes, outputs a uniquely oriented management task routing table, and maps and displays it on the scheduling center's visualization terminal. When filtering time constraints, the system extracts the current movement rate of each node in the state matrix, calculates the estimated travel time from the current stationary coordinates to the target space's measured observation coordinates, and determines when the estimated travel time exceeds the target escape time limit or the time limit in the state matrix is ​​exceeded. When the remaining range of a node is less than 1.5 times the safety redundancy of the absolute spatial distance, and the personnel's on-duty status is marked as offline, the node is determined not to meet the timeliness constraint and is removed from the set of available nodes. In this state, the disorderly random deviation caused by transient false alarms in the wave environment remains at a low integral and is spontaneously suppressed under the accumulation of sliding time windows. Meanwhile, illegal fishing entities with genuine intent to avoid the situation are accurately captured due to their structural deviation characteristics of trajectory forgery, triggering high-priority and accurate dispatching, allocating transportation resources to complete the control transition. To prevent regulated entities from injecting GPS location spoofing devices the moment communication is restored, The system suffers from a closed-loop vulnerability caused by falsified coordinates, leading to the system accepting false states. Before restoring the normal mode dominated by the first state data sequence, the cooperative modal data access unit calls the control server to calculate the instantaneous heading angular acceleration of the first state sequence recovery node and compares it with the physical kinematic characteristics of the second state sequence obtained by active radar detection at the very end. Only when the dynamic change rates of the two match and there is no step change exceeding the physical and mechanical limits is the causal feedback accepted by the underlying control gateway, thereby cutting off the path for position spoofing signals to intrude into the scheduling logic at the physical self-consistency level.

[0037] Example 2: When the system faces extreme sea state verification conditions where the cooperative communication network link is interrupted and the non-cooperative detection network is modulated by sea surface clutter, the data processing system runs an experimental verification on a control server with a data bus throughput rate of not less than 10Gbps and dual-channel heterogeneous data alignment function. The data used in the experimental verification comes from discrete track sequence datasets collected and solidified through a sea trial physical experimental platform. The ship detection radar corresponding to the non-cooperative data has a ranging accuracy of 15 meters, an angle measurement accuracy of 0.5°, a sampling rate of not less than 1Hz, and is the total number of discrete sampling points within the calibration continuous observation time window. This core variable requires physical constraints on the system's processing load and the timeliness of disagreement arbitration; based on the total number of discrete sampling points. The value of directly determines the calculation cycle of the spatiotemporal residual integral. An excessively large value increases the bus data processing load of the control server and causes a delay in the output of the scheduling routing table. Conversely, an excessively small value cannot effectively overcome the continuous interference cycle of ocean clutter, thus reducing the confidence level of intent recognition. Therefore, when the duration of the sudden noise accompanying the background clutter of the monitored signal is within a frequency domain distribution of 2 to 4 seconds, to ensure the minimum number of samples required for first-order kinematic extrapolation and to filter out false alarms from ocean waves, the total number of discrete sampling points... The number of discrete sampling points is between 8 and 12, and under this typical verification condition, the total number of sampling points is... Set to 10.

[0038] The verification working condition setting includes an input benchmark simulating sea surface wave clutter interference. Gaussian white noise with a signal-to-noise ratio of 15 dB is actively injected into the original second-state data sequence input to the non-cooperative modal data analysis unit, and a random spatial position disturbance with an amplitude of 25 meters and a period of 3 seconds caused by wave surging is added. In the first control sample group without this technical solution, the Gaussian white noise and position disturbance directly cause high-frequency disordered jumps in the spatial measured observation coordinate data, resulting in a false deviation between the observed coordinates at three consecutive discrete sampling points and the trajectory prediction data extrapolated from the cooperative modal data, producing a false positive spatiotemporal residual absolute value with an average value of 34.6 meters, rather than caused by actual ship motion. In the sample group of this invention using this technical solution, the non-cooperative modal data analysis unit uses spatiotemporal filtering to perform spatial overlap verification on stable reflection points within a continuous scanning period. Under a hard constraint of a spatial overlap threshold of 80%, due to the transient instability of wave clutter... The spatial overlap of the reflection point within adjacent periods is only 31.4% to 45.2%, thus falling below the set threshold. The non-cooperative modal data analysis unit filters and eliminates it, and the output real spatial measured observation coordinate data shows a smoothed spatial coordinate waveform. This 80% hard constraint threshold parameter is not a system-defined prior constant, but is derived based on the geometric projection topological characteristics of the entity target. Specifically, the rigid structure characteristics of a typical fishing vessel determine that the theoretical maximum value of the maximum shear offset caused by the physical rotation or translation of the hull divided by the total envelope area within adjacent extremely short periods of radar high-frequency scanning is limited to within 20%. This, in reverse anchoring the underlying geometric laws, ensures that the spatial overlap of the real ship's reflected echo must remain above the physical lower bound of greater than 80%, thus providing a reliable derivation basis for effectively separating the entity target from the continuously collapsing and changing sea clutter of the fluid physical form. At the discrete sampling point time of the first state data sequence... to After a sudden interruption of cooperative communication signals, the absolute value of the spatiotemporal residual locked by the disagreement arbitration unit at time... At that time, it was 2.3 meters. At that time, it was 4.1 meters. At a time of 5.8 meters, the divergence arbitration unit summed the absolute values ​​of the spatiotemporal residuals at all discrete sampling points at the end of the continuous observation time window, calculating a spatiotemporal residual integral of 38.5 meters. Since this value did not exceed the preset verification threshold of 50 meters, the divergence arbitration unit did not generate an abnormal scheduling weight label, thus proving that pure environmental noise cannot trigger scheduling changes. When the regulated entity changes course to avoid the event, as the deviation between the target's movement direction and the trajectory prediction data increases, the absolute values ​​of the spatiotemporal residuals calculated at the discrete sampling points in this invention's sample group show a single increasing trend from 12.4 meters, 26.8 meters to 78.2 meters, ultimately producing a spatiotemporal residual integral of 423.6 meters. This value exceeds the preset verification threshold and generates an abnormal scheduling weight label. To establish the rationality of the numerical range, the control server added a control group with out-of-tolerance parameters for testing. When the total number of discrete sampling points... When set to less than 5, the spatiotemporal residual integral is prone to fluctuation and overload under random noise impact. Its value reaches 62.1 meters even when the target is not avoided, causing the system to output incorrect management commands. Furthermore, when the total number of discrete sampling points... When the value is set to 20, which is greater than the upper limit of the working window, the single-step calculation time of the control server increases from 32.5ms to 245.8ms, indicating that this value range ensures both noise resistance and stability while taking into account the real-time performance of the calculation.

[0039] To quantitatively assess resource scheduling overhead under different avoidance characteristics, the dynamic scheduling management unit introduces the spatiotemporal residual integral as a dynamic adjustment parameter into the spatial topology function of the regulated entity, and calculates the scheduling cost parameter used to determine the priority of capacity allocation. The specific calculation formula is as follows: ,in, To make the first The idle managed resource node is scheduled to the first The scheduling cost parameters of each target are subject to the regulatory entity. For the first The current location coordinates of the idle managed resource nodes up to the [number]th [node]. The absolute spatial distance of the measured spatial coordinates of each target regulated entity. This represents the total number of discrete sampling points within the continuous observation time window. For the first The absolute value of the spatiotemporal residual formed by the predicted trajectory data and the measured spatial coordinate data at each sampling point. To prevent the system constant from having a denominator of zero, the control server ran the above calculation in multiple test sample groups containing different trajectory deviation intensities. When the trajectory deviation intensities were 20 meters, 50 meters, and 100 meters, the denominator term in the spatial topology function monotonically increased as the spatiotemporal residual integral value increased regularly, leading to a decrease in the calculated scheduling cost parameter. The scheduling cost parameters are reduced from 1.45 under the normal baseline to 0.82, 0.41, and 0.18, respectively. The monotonically convergent characteristics of the algorithm drive the resource allocation algorithm to automatically lock the supply and demand pairing with the lowest total cost after excluding idle managed resource nodes that do not meet the range limit. The algorithm also upgrades the corresponding capacity node task occupancy label and outputs a uniquely directional management task routing table. Tests show that by using the technical solution of this invention to plan resources, the system's scheduling hit rate has increased from 58.3% under the existing static threshold triggering method to 96.2%, and the average response latency of administrative scheduling instructions has been shortened from 18.4 minutes to 2.1 minutes. The entire data flow and conversion control is completed entirely by the processor of the control server for memory addressing and bus state switching. This improves the capacity allocation efficiency of the marine supervision and management data processing system under low visibility conditions at night without modifying the surface hardware.

[0040] Example 3: This example combines Figures 1 to 2 This document describes a dual-modal driven intelligent monitoring and scheduling system and method for fishing vessels at night. Figure 1 As shown, steps S1, S2, S3, and S4 are as follows: Step S1 involves acquiring the first state data sequence output by the cooperative mode data access unit and the second state data sequence output by the non-cooperative mode data parsing unit; Step S2 involves extracting the spatial measured observation coordinate data of the second state data sequence within the continuous observation time window when it is determined that the first state data sequence has a signal interruption; Step S3 involves calculating and generating the spatiotemporal residual integral based on the spatial measured observation coordinate data and the trajectory prediction data of the first state data sequence; Step S4 involves injecting the spatiotemporal residual integral as a dynamic adjustment scalar into the target response spatial topology function for the target regulated entity, updating the task occupancy label of the globally available managed resource nodes, and outputting the directed management task routing table.

[0041] like Figure 2As shown, the dual-modal driven intelligent nighttime monitoring and dispatching system for fishing vessels includes a cooperative modal data access unit, a non-cooperative modal data parsing unit, a data status perception unit, a disagreement arbitration unit, a dynamic dispatching management unit, a dispatching center visualization terminal, and a target managed resource node terminal. The data status perception unit is interconnected with the cooperative modal data access unit and the non-cooperative modal data parsing unit via a data bus; the disagreement arbitration unit is associated with the data status perception unit via a data path; and the dynamic dispatching management unit is associated with the disagreement arbitration unit via a signal path. The system acquires a first state data sequence output by the cooperative modal data access unit and a second state data sequence output by the non-cooperative modal data parsing unit through the data status perception unit. When determining the first state data sequence... When a signal interruption occurs, the data status perception unit extracts the spatial measured observation coordinate data of the second state data sequence within the continuous observation time window; the divergence arbitration unit performs difference calculation based on the spatial measured observation coordinate data and the trajectory prediction data of the first state data sequence to generate a spatiotemporal residual integral; the dynamic scheduling management unit injects the spatiotemporal residual integral as a dynamic adjustment scalar into the target response spatial topology function for the target regulated entity, updates the task occupancy label of the globally available managed resource nodes, and outputs a directional management task routing table; the scheduling center visualization terminal is connected to the dynamic scheduling management unit through a data bus to display the spatial topology network mapping of the directional management task routing table and issue digital management scheduling instructions to the target managed resource node terminals pointed to by the directional management task routing table.

[0042] Example 4: During nighttime surface target supervision and management, when the wind and wave level reaches level 4 and is accompanied by gusts, the scattered echoes generated by the swells on the sea surface produce disordered environmental false alarms in the non-cooperative detection network. If the fixed static verification threshold is used, the spatiotemporal deviation residuals caused by the environmental false alarms accumulate within the continuous observation time window, causing the integral of the spatiotemporal residuals to jump beyond the fixed verification boundary. This generates a changing scheduling weight label in the control center, causing the control server to distribute scheduling instructions deviating from the target sea area to idle managed resource nodes. This results in scheduling redundancy of surface transport capacity resources in a specific sea area, and the focus shifts to avoiding missed detections caused by clutter signal coverage of the target. This causes fluctuations in the stability of the transport capacity planning in the scheduling management and control center when the wind and wave intensity changes abruptly.

[0043] To suppress control system anomalies caused by environmental noise fluctuations that are not compatible with fixed boundaries, the control server allocates a sliding time buffer of 100 frames in the processor of the scheduling control gateway. It continuously reads the discrete raw scattered voltage sequences acquired by the radar antenna. The processor extracts the 100 frames of raw voltage sequences from the time buffer, transforms the voltage sequences into a background scattered energy matrix using the amplitude square algorithm, and processes the background scattered energy matrix using a two-dimensional discrete Fourier transform to separate the spatial high-frequency components. It calculates the rolling variance of the spatial high-frequency components on the time axis, and combines this rolling variance with the total number of discrete sampling points within the current continuous observation time window. The real-time verification threshold is dynamically calculated using a linear gain control function. The calculation formula is as follows: ,in, The real-time verification threshold after online calibration. This refers to the dimensionless scaling factor embedded within the control server. This represents the rolling variance of the spatial high-frequency components measured within the sliding time buffer. This is a constant noise floor correction value that characterizes the inherent ranging system error of the radar.

[0044] The control server dynamically adjusts the verification threshold in memory based on the changing rolling variance within the sliding timing buffer. When the non-cooperative mode data parsing unit calculates the spatial overlap of the target reflection points within a continuous scan cycle, the processor loads the discrete grid coordinate array of the target within three consecutive scan cycles into the cache. The spatial overlap is determined by comparing the ratio of the number of intersection elements to the number of union elements in the grid coordinate arrays. If the first state data sequence signal is interrupted, the divergence arbitration unit retrieves the dynamically updated real-time verification threshold. Instead of a statically preset verification threshold, a real-time verification threshold is applied when the rolling variance of the environmental surge increases from 0.15 to 0.85. The value monotonically increases from 35.2 meters to 72.4 meters, making the real-time verification threshold... The spatiotemporal residual integral value formed by pure clutter signals without steering avoidance behavior remains stable below the real-time verification boundary when the data deviation caused by sudden distortion of the sea waves is greater than that caused by sudden distortion of the sea waves. However, when the steering avoidance target undergoes continuous unidirectional position changes, the spatiotemporal residual integral value reaches 315.6 meters and crosses the real-time verification boundary. A uniquely directional management task routing table is output to the scheduling and management terminal. The processor converts the resource allocation routing table and updates the task occupancy label of the available resource nodes. This ensures the stability of task distribution control of the maritime administrative supervision and scheduling system under low visibility conditions at night without changing the surface hardware. In addition, for the trigger condition of communication link interruption recovery under signal interruption, the system also performs a closed-loop check with a preset threshold at this stage. The coordinates of the first frame after recovery are extracted and compared with the non-cooperative measured coordinates at the same timestamp. The preset threshold for this coordinate deviation is not a static empirical value, but is taken from the real-time verification threshold dynamically updated in the aforementioned steps. The lower limit safety envelope is the range of 35.2 meters to 50 meters. The system will only issue an instruction to close the continuous observation time window if the coordinate deviation is absolutely converged within this preset threshold limit within three consecutive cycles, ensuring that the determination of the system's exit from the abnormal flow interruption state has a definite underlying data support.

[0045] Example 5: When the system faces deployment conditions in a new monitored sea area, the control server reads the pre-calibration parameters before receiving data, the data status sensing unit acquires the passive scattering signal of the radar antenna under no-load conditions, and the processor calculates the average value of the original voltage sequence over 600 consecutive scan cycles as a constant noise floor correction value characterizing the inherent system error of the radar. The numerical benchmark is used; the non-cooperative modal data parsing unit extracts the spatial measured coordinate data of a 30-meter-long calibration vessel, and the disagreement arbitration unit compares the measured coordinates with the benchmark coordinates output by the calibration vessel's positioning terminal, calculates the absolute spatial deviation at 200 consecutive sampling points, and the processor calculates the root mean square value of the spatial deviation and the total number of discrete sampling points. Perform scaling to calculate the dimensionless scaling factor. The physical boundaries are initialized using parameters that match the radar hardware, established by the gain control model.

[0046] In a maritime surveillance scenario where the system with the parameters initialized as described above is running, the processor of the scheduling and control gateway initiates trajectory alignment control after calibration. When the first-state data sequence experiences a fault and disconnection, the processor loads a dimensionless scaling factor. With constant noise correction value Dynamically update real-time verification threshold ; Dimensionless scaling factor at the step of the rolling variance of background scattering fluctuations Adjusting the real-time verification threshold Moving in the same monotonically converging direction is used to track the upper limit of energy dissipation of background clutter and control the residual integral value caused by inherent hardware noise. When avoiding unidirectional changes in the target, the spatiotemporal residual integral exceeds the real-time verification boundary, driving the dynamic scheduling management unit to output the directional management task routing table, update the task occupancy label of available resource nodes, and distribute transportation resources to the target water surface location.

[0047] Example 6: When the system faces signal drift caused by the long-term aging of data conversion devices in the non-cooperative mode data parsing unit, the second-state data sequence output by the non-cooperative detection network generates a continuous linear position displacement. This causes a unidirectional spatial position deviation between the spatial measured observation coordinate data calculated under normal conditions without steering avoidance behavior and the trajectory prediction data generated by the cooperative mode data. This spatial deviation caused by physical damage to hardware devices accumulates continuously on the time axis, causing the spatiotemporal residual integral calculated by the divergence arbitration unit to exceed the verification judgment limit and generate false abnormal scheduling weight labels. This drives the dynamic scheduling management unit to calculate the deviated resource pairing scheme based on the incorrect spatial topology overhead and output a directional management task routing table containing the incorrect flow target on the control terminal, reducing the scheduling matching degree of the supervision and management system under low visibility conditions at night.

[0048] To correct residual reconstruction errors caused by hardware aging drift, the data state awareness unit invokes a differential drift compensation operator when receiving the second state data sequence. This operator captures the hardware offset by periodically reading the fixed sea surface beacon reference coordinates. The control server then uses the measured real-time beacon coordinate values ​​to subtract the stored standard coordinate values ​​to determine the current drift amount. And utilize this drift amount The coordinates of each sampling point in the second-state data sequence are corrected to output the corrected spatial measured observation coordinate data. The specific calculation formula is as follows: ,in, To correct the position components in the measured coordinate data of the regulated entity's space, For the corresponding positional components in the original second-state data sequence without accompanying waveform index compensation, This represents the temporal linear drift deviation calculated from the fixed sea surface beacon reference coordinates over a 24-hour period.

[0049] Under the real-time intervention of the differential drift compensation operator, the linear bias caused by device aging in the non-cooperative mode data parsing unit is canceled out before entering the divergence arbitration, so that the corrected spatial measured observation coordinate data reflects the physical position of the water surface target, limiting the accumulation of false residuals in the divergence arbitration unit. This drives the dynamic scheduling management unit to determine the resource pairing combination based on the corrected scheduling cost parameters and automatically update the task occupancy label of available resource nodes. As a result, a uniquely directional management task routing table is output on the scheduling visualization terminal, maintaining the stability of the data processing system's control over the distribution status of managed resource nodes under abnormal signal conditions.

[0050] The embodiments of this application have been described above with reference to the accompanying drawings. Unless otherwise specified, the embodiments and features in the embodiments of this application can be combined with each other. This application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit of this application and the scope of protection of this invention, and all of these forms are within the protection scope of this application.

Claims

1. A dual-modal driven intelligent monitoring and dispatching system for fishing vessels at night, characterized in that, It includes a cooperative modal data access unit, a non-cooperative modal data parsing unit, a data status awareness unit, a disagreement arbitration unit, and a dynamic scheduling management unit. The data status perception unit is interactively connected with the cooperative mode data access unit and the non-cooperative mode data parsing unit through a data bus; the disagreement arbitration unit is associated with the data status perception unit through a data path; and the dynamic scheduling management unit is associated with the disagreement arbitration unit through a signal path. The system manages data scheduling for globally available managed resource nodes through the following timing control steps: Step 101: The data state perception unit acquires the first state data sequence output by the cooperative mode data access unit and the second state data sequence output by the non-cooperative mode data parsing unit. Step 102: When it is determined that the signal of the first state data sequence has been interrupted, the data state sensing unit extracts the spatial measured observation coordinate data of the second state data sequence within the continuous observation time window. Step 103: The disagreement arbitration unit performs difference calculation based on the space measured observation coordinate data and the trajectory prediction data of the first state data sequence to generate the spatiotemporal residual integral. Step 104: The dynamic scheduling management unit injects the spatiotemporal residual integral as a dynamic adjustment scalar into the target response space topology function for the target regulated entity, updates the task occupancy label of the globally available managed resource nodes, and outputs the directional management task routing table.

2. The dual-modal driven intelligent monitoring and dispatching system for fishing vessels at night according to claim 1, characterized in that, In step 102, the data state sensing unit is used to extract spatial measured observation coordinate data through the following sub-steps: Step 1021, when the spatial coordinates and heading vector in the first state data sequence are missing for three consecutive sampling periods, it is determined that a signal interruption has occurred; Step 1022, within the continuous observation time window, the measured scattered echo coordinates in the second state data sequence on the coaxial time axis are continuously retrieved as spatial measured observation coordinate data.

3. The dual-modal driven intelligent monitoring and dispatching system for fishing vessels at night according to claim 1, characterized in that, In step 103, the disagreement arbitration unit calculates the spatiotemporal residual integral through the following sub-steps: Step 1031, obtain the spatial coordinates and heading vector of the last frame before the signal interruption of the first state data sequence, and derive the predicted spatial coordinates of each sampling moment within the continuous observation time window using a first-order kinematic extrapolation algorithm, as the trajectory prediction data; Step 1032, calculate the Euclidean distance between the measured spatial observation coordinate data and the trajectory prediction data at the same sampling moment, and generate the absolute value of the spatiotemporal residual; Step 1033, at the end of the continuous observation time window, sum the absolute values ​​of all spatiotemporal residuals within the continuous observation time window in the time dimension to obtain the spatiotemporal residual integral.

4. The dual-modal driven intelligent monitoring and dispatching system for fishing vessels at night according to claim 1, characterized in that, The disagreement arbitration unit is also used to compare the spatiotemporal residual integral with the safety threshold to determine the temporal logic disagreement degree; the disagreement arbitration unit uses the following calculation formula to determine the safety threshold: ,in, As a safety threshold, This serves as the baseline noise level established based on historical nighttime noise data of the spatial environment. To determine the duration of the continuous observation window, the dynamic scheduling management unit outputs a directional management task routing table in step 104 through the following sub-steps: Step 1041: Extract the current residence coordinates of idle managed resource nodes within the managed area. Under the state increment generated by the spatiotemporal residual integral, the spatiotemporal residual integral is used as a penalty weight and superimposed in a monotonically increasing direction onto the distance function of each idle managed resource node moving towards the target managed entity, thus constructing the target response space topology function; Step 1042: With the minimum total scheduling cost parameter as the convergence optimization objective, solve the target response space topology function, lock the resource pairing combination, and output the directional management task routing table.

5. The dual-modal driven intelligent monitoring and dispatching system for fishing vessels at night according to claim 4, characterized in that, Before constructing the target response space topology function in step 1041, the dynamic scheduling management unit is also used to filter nodes through the following derived sub-steps: Step 10411, obtain a state matrix containing the remaining range, current movement speed, and personnel on-duty status of each idle managed resource node; Step 10412, under the constraints of the state matrix, exclude idle managed resource nodes that do not meet the timeliness constraints, and pass the set of available nodes to step 1041 to update the task occupancy tags of the globally available managed resource nodes.

6. The dual-modal driven intelligent monitoring and dispatching system for fishing vessels at night according to claim 4, characterized in that, The dynamic scheduling management unit is also used to determine whether the target regulated entity has intentional avoidance characteristics under dynamic response conditions where the spatiotemporal residual integral is greater than the safety threshold, and to increase the scheduling priority of the target regulated entity in obtaining globally available managed resource nodes.

7. The dual-modal driven intelligent monitoring and dispatching system for fishing vessels at night according to claim 1, characterized in that, The non-cooperative modal data parsing unit is used to perform spatiotemporal filtering on the incoming raw non-cooperative data, filtering out unstable reflection points with a spatial overlap of less than or equal to 80% within three consecutive scan cycles, and outputting stable reflection points with a spatial overlap of greater than 80% within three consecutive scan cycles as spatial measured observation coordinate data to the data status sensing unit.

8. A dual-modal driven intelligent monitoring and dispatching system for fishing vessels at night, as described in claim 2, is characterized in that, When the communication link is restored from a signal interruption and the first state data sequence is received again, the cooperative mode data access unit calculates the coordinate deviation between the coordinates of the first frame after restoration and the coordinates of the second state data sequence at the same timestamp. When the coordinate deviation stabilizes within a preset threshold range under causal feedback, the continuous observation time window is closed, and the normal monitoring mode dominated by the first state data sequence is restored.

9. A dual-modal driven intelligent monitoring and dispatching system for fishing vessels at night, as described in claim 1, characterized in that, The system also includes: a scheduling center visualization terminal; the scheduling center visualization terminal is connected to the dynamic scheduling management unit through a data bus, and is used to display the spatial topology network mapping of the directional management task routing table, and to issue digital management scheduling instructions to the target managed resource node terminals pointed to by the directional management task routing table.

10. A dual-modal driven intelligent monitoring and scheduling method for fishing vessels at night, used to operate the dual-modal driven intelligent monitoring and scheduling system for fishing vessels at night as described in claim 1, characterized in that, Includes the following steps: Step S1: Obtain the first state data sequence output by the cooperative mode data access unit and the second state data sequence output by the non-cooperative mode data parsing unit; Step S2: When it is determined that the signal of the first state data sequence has been interrupted, extract the spatial measured observation coordinate data of the second state data sequence within the continuous observation time window; Step S3: Based on the space-measured observation coordinate data and the trajectory prediction data of the first state data sequence, calculate and generate the spatiotemporal residual integral. Step S4: Inject the spatiotemporal residual integral as a dynamic adjustment scalar into the target response space topology function for the target regulated entity, update the task occupancy label of the globally available managed resource nodes, and output the directed management task routing table.

Images