A method and system for dynamic formation reconstruction of multiple UUVs based on collaborative sentinel mechanism

Abstract

This invention discloses a method and system for dynamic formation reconfiguration of multiple UUVs based on a collaborative sentinel mechanism, relating to the field of collaborative control technology for underwater unmanned systems. When a UUV transitions from a static sentinel state to a deployed state, the system triggers real-time reconfiguration: locking the affected area and removing the demobilized UUV, calculating the coverage gap of the cluster; dynamic reconfiguration is performed based on a three-layer collaborative architecture, with the global layer selecting a formation template based on the threat heat distribution along the strike line, the task layer assigning functional roles based on UUV capability profiles, and the execution layer calling a static deployment optimization algorithm to generate temporary sentinel positions; simultaneously, a virtual occupancy mechanism and a rolling dynamic adaptation strategy are introduced to avoid frequent reconfiguration caused by short-term demobilization, and to support continuous task response in highly dynamic environments. This invention solves the problem of defense gaps caused by UUV demobilization in traditional static defense formations, significantly improving the system robustness and continuous operation capability of the cluster in high-density, high-speed, and dispersed penetration scenarios.

Description

Technical Field

[0001] This invention relates to the field of cooperative control technology for underwater unmanned systems, and in particular to a method and system for dynamic array reconfiguration of multiple UUVs based on a cooperative sentinel mechanism. Background Technology

[0002] In modern underwater offensive and defensive systems, the coordinated execution of area defense missions by multiple unmanned underwater vehicles (UUVs) swarms has become a crucial means of enhancing underwater situational awareness and rapid interception capabilities. Existing technologies largely focus on static deployment optimization or offline mission planning, such as pre-allocating UUV outposts through genetic algorithms, particle swarm optimization, or coverage models to maximize coverage of the defense area. However, these methods generally assume a relatively stable threat environment and that UUVs remain stationed in fixed positions throughout the entire process, failing to consider the gaps in the defense formation caused by the dynamic deployment of UUVs to perform interception missions in high-speed, random, and multi-target penetration scenarios. Once a high-value UUV leaves its post, its original coverage area will be exposed for an extended period, allowing subsequent penetration targets to exploit the vulnerability and creating a chain reaction of failures. Therefore, maintaining the system-level defense integrity and dynamic robustness of the remaining UUV swarm while ensuring a high success rate for individual interceptions has become a core challenge that urgently needs to be overcome in current multi-UUV swarm coordinated defense.

[0003] To address the aforementioned challenges, prior art document CN115310293A proposes a multi-UUV cooperative detection array optimization method based on ant colony algorithm, which improves coverage efficiency by simulating ant foraging behavior to adjust UUV positions online. CN121274966A discloses a global task planning method for heterogeneous UUV platforms based on full path coverage, emphasizing complete traversal of task paths. However, neither of these methods solves the key problem of "static defense array instability caused by individual task assignment": the former focuses on local optimization of detection coverage and lacks a response mechanism for UUV state switching; the latter focuses on the integrity of path planning and does not involve real-time reconstruction and elastic recovery of array structure. Especially under high saturation attacks, when multiple UUVs are continuously assigned interception tasks, existing solutions cannot dynamically fill defense gaps or adaptively adjust array form according to threat distribution. Therefore, a dynamic cooperative reconstruction mechanism is urgently needed that can detect UUV off-duty events, quickly generate temporary optimal arrays, and automatically restore the initial defense posture after the task ends. Summary of the Invention

[0004] In view of this, this invention proposes a method and system for dynamic multi-UUV array reconfiguration based on a collaborative sentinel mechanism. By constructing a three-layer collaborative architecture of global, task, and execution layers, and combining virtual placement and rolling reconfiguration mechanisms, it effectively solves the coverage gap problem caused by the single-unit withdrawal in static defense systems, and significantly improves the system robustness and continuous operation capability of multi-UUV clusters in high-density, high-speed, and distributed penetration scenarios.

[0005] This invention provides a method for dynamic reconfiguration of multiple UUV formations based on a cooperative sentinel mechanism, comprising the following steps: S1. Real-time monitoring of the status of multiple UUV clusters in the initial defense formation; when any UUV is detected to switch from static sentinel state to mission execution state, assess the impact of its departure on the defense system based on its expected return time and the system dynamic reconstruction threshold. S2. If the evaluation results meet the reconstruction triggering conditions, the physical array reconstruction program is started; otherwise, the original sentry positions are preserved and reconstruction is suppressed through the virtual placeholder mechanism. S3. The physical formation reconstruction procedure includes: generating a temporary formation based on the capability profile of the remaining UUVs and the threat heat distribution on the current strike line through hierarchical decision-making, and issuing sentry post adjustment instructions to the relevant UUVs; S4. During the refactoring process, if a new UUV is detected to be assigned a task, the current process is interrupted and the evaluation and refactoring process is re-executed based on the updated cluster state to achieve rolling dynamic adaptation. S5. After the mission UUV returns to base, perform local self-healing or global repositioning operations dynamically based on the overall cluster status to restore the integrity of the defense system.

[0006] Furthermore, the hierarchical decision-making adopts a three-layer collaborative architecture, specifically including: At the global level, a matching formation template is selected based on the threat heat distribution of targets on the current strike line. The formation template includes a honeycomb contraction type, a linear extension type, or a wingspan enhancement type. At the mission level, based on the capability profile of each UUV, a functional role is assigned to it. The capability profile includes maximum speed, detection radius and remaining endurance. The functional roles include rapid response sentry, core perception sentry or persistent defense sentry. At the execution layer, constrained by the array template and functional roles, the static deployment optimization algorithm is invoked to calculate a set of temporary optimal sentry positions for the new cluster.

[0007] Furthermore, the threat thermal distribution is generated in any of the following ways: Based on the historical location of the target obtained by reconnaissance UUVs along the detection line, or the predicted arrival point density estimate provided by the real-time intelligence module; and when there is a heat peak area in the threat heat distribution, the area is determined to be a high-risk segment, triggering the honeycomb contraction array template, causing UUVs to be deployed in a concentrated manner to the area.

[0008] Furthermore, the capability profile includes maximum speed, detection radius, and remaining endurance; the allocation rules for the functional roles are as follows: If the maximum speed of the UUV is higher than the cluster average and there is sufficient remaining endurance, it will be assigned as a rapid response sentry and deployed at the front of the formation. If the UUV has the largest detection radius, it will be assigned as the core sensing sentry and deployed at the center of the formation; If a UUV has the longest endurance but the lowest speed, it will be assigned as a long-term defensive outpost and deployed in a rear support position.

[0009] Furthermore, the static deployment optimization algorithm aims to minimize the following multi-objective cost function when calculating temporary outposts: ; in, This refers to the area not covered by the strike line. This is the weighted average response time; A penalty term for UUVs whose distance between them is less than a safety threshold; weight. , , Dynamically adjust based on the current formation template.

[0010] Furthermore, the assessment of the impact of absence on the defense system includes: calculating the estimated return time of the assigned UUVs. and the dynamic reconstruction trigger threshold Compare; if If the original sentry post is not returned within the specified time, the virtual post will be retained as a virtual placeholder and the formation reconstruction will be suppressed; if the post does not return within the specified time, the virtual placeholder will be released and reconstruction will be initiated.

[0011] Furthermore, the dynamic reconstruction trigger threshold Calculate using the following formula: ; in, This represents the maximum displacement distance of the UUV in the reconstruction scheme. The average cruising speed under the current hydrological conditions. Let k be the average delay of the underwater communication link, and k be the stability coefficient. This refers to the algorithm response time related to the computational resource load.

[0012] Furthermore, after generating a temporary optimal set of sentry posts, the execution layer issues sentry post adjustment instructions to the UUV using a phased movement strategy: Phase 1: Converge towards the target region at maximum speed; Phase 2: Decelerate and fine-tune as you approach the temporary outpost, while simultaneously executing a dynamic collision avoidance algorithm.

[0013] Furthermore, the static sentinel state includes two sub-states: Original static sentinel: located at the initial optimal sentinel position; Temporary static sentry: Located at the reconstructed temporary sentry post; The system distinguishes between the two to ensure that the return command is only issued to UUVs at temporary outposts.

[0014] Furthermore, the present invention also provides a multi-UUV dynamic formation reconfiguration system based on a cooperative sentinel mechanism, comprising: The status monitoring module is used to monitor the status of multiple UUV clusters in real time. When it detects that a UUV has switched from a static sentinel state to a task execution state, it triggers the evaluation process. Virtual placeholder manager, used for determining the expected return time With dynamic reconstruction trigger threshold The comparison results determine whether to enable virtual placeholders and suppress refactoring; The threat assessment module is used to generate a threat heat map of the strike line based on the historical locations or real-time predicted arrival point density obtained by reconnaissance UUVs along the identified line. The three-layer collaborative reconstruction module includes: a global layer unit, configured to select a formation template based on threat heat distribution; a task layer unit, configured to assign functional roles based on UUV capability profiles; and an execution layer unit, configured to call a static deployment optimization algorithm to calculate a temporary optimal set of sentry posts with the formation template and functional roles as constraints. The instruction issuing module is used to issue phased transit instructions to the UUVs participating in the reconstruction, and to issue return-to-position instructions to the temporary static sentinels after the mission UUVs return to base. The system supports detecting new task assignments during the reconstruction process and triggering a rolling dynamic adaptation process.

[0015] The present invention has the following advantages over the prior art: By constructing a three-tiered collaborative architecture encompassing global, task, and execution layers, the system achieves rapid adaptive reconfiguration of multi-UUV defense formations under dynamic task disturbances. This layered mechanism avoids the high computational overhead of traditional global re-optimization. Under typical underwater communication and computing power constraints, it can complete the entire process from threat perception to new outpost generation within 200 milliseconds, significantly improving response time. More importantly, the system can intelligently match formation templates such as honeycomb contraction, linear extension, or wingspan enhancement based on the real-time threat heat distribution along the strike line. It also assigns functional roles based on the maximum speed, detection radius, and remaining endurance of each UUV, enabling high-speed units to intercept forward, large detection units to perceive centrally, and long-endurance units to defend in the rear, thereby achieving precise coupling of capabilities and tasks under heterogeneous resource conditions.

[0016] Furthermore, this invention introduces a virtual placeholder mechanism. When the estimated return time of an assigned UUV is less than a preset threshold, its original sentry post is marked as a virtual placeholder, preventing full cluster reconfiguration. This effectively suppresses frequent formation oscillations caused by short-duration tasks, reduces the number of reconfiguration commands by more than two-thirds, and significantly extends the system's continuous operational time. Simultaneously, the rolling dynamic adaptation mechanism ensures that if a new interception assignment occurs during the reconfiguration process, the system can immediately interrupt the current process and replan based on the latest cluster state, thereby improving the interception success rate under extreme stress testing and fully verifying the system's robustness under high-saturation attacks.

[0017] In addition, the present invention also designs a complete closed loop of attack, reconstruction and return. After the mission is completed, all UUVs in temporary outposts automatically return to their initial optimal positions, so that the defense system is restored to its full strength.

[0018] In summary, this invention not only solves the fundamental problem of structural vulnerabilities in traditional static defense after UUVs leave their posts, but also achieves a synergistic leap in three dimensions: timeliness, resource efficiency, and system resilience, providing a dynamic and collaborative defense paradigm that can be implemented in complex underwater combat environments. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is an overall process architecture diagram of an embodiment of the present invention; Figure 2 This is a flowchart of an embodiment of the present invention. Detailed Implementation

[0021] The technical solutions of the present invention will be clearly and completely described below with reference to the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0022] like Figure 1As shown, this invention provides a method for dynamic multi-UUV formation reconfiguration based on a collaborative sentinel mechanism. Its core lies in achieving rapid and adaptive temporary formation reconfiguration through a three-layer collaborative architecture when gaps in the defense formation occur due to UUVs being absent from their interception missions. Combined with virtual occupancy and feedback control mechanisms, ineffective reconfiguration is avoided, ensuring the system's continuous operational capability. This method mainly includes the following key steps: S1. Real-time monitoring of the status of multiple UUV clusters in the initial defense formation; when any UUV is detected to switch from static sentinel state to mission execution state, assess the impact of its departure on the defense system based on its expected return time and the system dynamic reconstruction threshold. S2. If the evaluation results meet the reconstruction triggering conditions, the physical array reconstruction procedure is started; otherwise, the original sentry position is preserved and reconstruction is suppressed through the virtual placeholder mechanism. S3. The physical formation reconstruction procedure includes: generating a temporary formation based on the capability profile of the remaining UUVs and the threat heat distribution on the current strike line through hierarchical decision-making, and issuing sentry post adjustment instructions to the relevant UUVs; S4. During the refactoring process, if a new UUV is detected to be assigned a task, the current process is interrupted and the evaluation and refactoring process is re-executed based on the updated cluster state to achieve rolling dynamic adaptation. S5. After the mission UUV returns to base, perform local self-healing or global repositioning operations dynamically based on the overall cluster status to restore the integrity of the defense system.

[0023] like Figure 2 The diagram shown is a flowchart of the present invention. Through modular collaboration, while ensuring the execution of interception tasks, it effectively suppresses frequent formation oscillations caused by short-term tasks, and improves the system robustness of the cluster in a high-saturation, high-dynamic confrontation environment.

[0024] In one implementation, the present invention designs a three-layer collaborative architecture, namely a hierarchical decision-making architecture. The hierarchical decision-making architecture includes: at the global layer, selecting a matching formation template based on the threat heat distribution of targets on the current strike line, wherein the formation template includes a honeycomb contraction type, a linear extension type, or a wingspan-enhanced type; at the mission layer, assigning functional roles to each UUV based on its capability profile, wherein the capability profile includes maximum speed, detection radius, and remaining endurance, and the functional roles include rapid response sentry posts, core perception sentry posts, or persistent defense sentry posts; at the execution layer, using the formation template and functional roles as constraints, invoking a static deployment optimization algorithm to calculate a set of temporarily optimal sentry post sets for the new cluster.

[0025] Specifically, when a UUV changes from a static sentinel state to a dispatched state due to an interception command, it first submits a status update to the status monitoring module. Upon detecting this change, the status monitoring module transmits the status information to the threat assessment module. The latter, combined with intelligence gathered from the identified lines, analyzes the threat heatmap distribution along the current strike line and provides the results to the three-layer collaborative architecture. Simultaneously, the three-layer collaborative architecture sends a request to the virtual placeholder module to assess whether reconfiguration is permitted: if the estimated return time of the UUV is less than a preset threshold, reconfiguration is deemed not permitted, and the system maintains its current formation to avoid unnecessary resource consumption; otherwise, reconfiguration is deemed permitted, triggering subsequent processes.

[0026] Under the reconfigurable branch, the three-layer collaborative architecture selects a formation template (hive / linear / wingspan) based on threat heatmaps, assigns functional roles according to UUV capability profiles, and uses a static deployment algorithm to generate temporary optimal outposts. Finally, the command issuance module issues transit commands to the remaining UUV clusters to complete the formation adjustment. The entire process forms a complete closed loop from state awareness to decision execution and feedback control, achieving rolling dynamic adaptation to continuous mission disturbances and ensuring the stability and efficiency of the multi-UUV cluster defense system in continuous operations.

[0027] In one implementation, the historical location of the target obtained by reconnaissance UUVs along the detection line, or the predicted arrival point density estimate provided by the real-time intelligence module, is used; and when there is a heat peak area in the threat heat distribution, the area is determined to be a high-risk segment, triggering the honeycomb contraction array template, so that the UUVs are concentrated in the area.

[0028] Specifically, the system deploys several reconnaissance UUVs to conduct routine patrols along the identified line. The identified line is located a certain distance ahead of the defense baseline, serving as an early warning front to detect target activity information in advance. Each reconnaissance UUV uploads the following two types of intelligence to the central command node in real time or periodically: Historical location data: recording the horizontal coordinates of targets actually crossing the identified line over a past period. Real-time predicted arrival point: Based on currently detected target motion parameters (speed, heading), extrapolate its expected position across the strike line. and time .

[0029] After receiving the above data, the central command node will target the strike line area. The data is discretized into M equal-width cells. Then, the threat intensity of each cell is calculated using kernel density estimation. in, Let be the i-th historical or predicted position, and N be the total number of samples; For Gaussian kernel function, u is the standardized distance variable in kernel density estimation; h is the bandwidth parameter, which controls the smoothness of the thermal distribution, with a typical value of 1.5 times the cell width. That is, location The threat weights at each location, after normalization, satisfy the following conditions: .

[0030] Ultimately, all A threat heatmap forming the strike line. This heatmap is updated in real time and serves as a direct input for selecting the global layer formation template: if a certain area... If the value exceeds a preset threshold (e.g., 0.05), it is considered a high-risk segment, triggering the "honeycomb contraction" formation; if the heat distribution is uniform, the "linear extension" formation is selected; if the weights on both sides are significantly higher than those in the center, the "wingspan enhancement" formation is activated.

[0031] In one implementation, the functional roles are assigned according to the following rules: if the maximum speed of the UUV is higher than the cluster average and the remaining endurance is sufficient, it is assigned as a rapid response sentry and deployed at the front of the formation; if the UUV has the largest detection radius, it is assigned as a core perception sentry and deployed at the center of the formation; if the UUV has the longest endurance but a lower speed, it is assigned as a persistent defense sentry and deployed at a rear support position.

[0032] Specifically, after receiving the current status information of the UUV cluster, the system first extracts the capability profile triplet from the status message of each UUV: ,in, The maximum speed of the i-th UUV, in m / s; Its acoustic / optical detection radius, in meters; The remaining battery life is estimated based on the current battery level and power consumption model, in seconds.

[0033] Then, calculate the current statistical baseline value of the cluster: Average speed: ,in The number of UUVs participating in the reconstruction; Remaining battery life threshold: set to This is used to determine if the battery life is "sufficient".

[0034] Next, roles will be assigned in the following priority order: Fast-Response Sentinel Filtering: Iterate through all UUVs, and if the following conditions are met... Then mark it as a candidate rapid response sentry. Select the one with the highest speed from among them. ships, Deployed at the forefront of the formation, that is, close to the outer boundary of the strike line and the area of ​​peak threat heat, it is responsible for high-speed interception of newly emerging targets.

[0035] Core Sensing Sentinel Selection: Among the unassigned UUVs, the one with the largest detection radius is selected. If multiple UUVs have the same radius, the one with the longest endurance is selected and assigned as the core sensing sentinel. It is deployed at the geometric center of the formation or near the center of threat, serving as a global situational awareness node to provide collaborative detection data fusion support for other UUVs.

[0036] Persistent Guard Sentinel Allocation: Among the remaining UUVs, if a UUV satisfies... They are prioritized for long-term defensive outposts, deployed in support positions behind the formation, far from the line of attack and close to underwater command and control nodes, to undertake long-term duty, communication relay or emergency backup tasks.

[0037] In one implementation, the static deployment optimization algorithm aims to minimize the following multi-objective cost function when calculating temporary outposts: in, This refers to the area not covered by the strike line. This is the weighted average response time; This is a penalty term for when the distance between UUVs is less than the safety threshold; the weights α, β, and γ are dynamically adjusted according to the current formation template.

[0038] Specifically, after the three-layer collaborative architecture completes the formation template selection (such as honeycomb contraction, linear extension, or wingspan enhancement) and functional role allocation, the execution layer calls the static deployment optimization algorithm within the current strike line range. The above is the remaining A UUV calculates a set of temporary optimal sentry posts. This minimizes the multi-objective cost function J, where each sub-item is defined and calculated as follows: Coverage cost The strike line is discretized into M equidistant sampling points. For each point Calculate whether it is covered by the detection range of at least one UUV: ,but ,in The generated threat heatmap weights ensure that high-risk areas that are not covered are penalized more severely.

[0039] Response time cost For each sampling point Calculate the theoretical response time of the nearest UUV: The weighted average response time is: .

[0040] Robust punishment To avoid collisions or communication interference between UUVs, a safe distance is introduced. (e.g., 80m). For all UUV pairs , If the gap is less than the threshold, an exponential penalty is applied: This form increases sharply when the distance is close to 0, effectively suppressing excessive aggregation.

[0041] The weighting coefficients α, β, and γ are not fixed, but are automatically configured based on the currently selected formation template to reflect different mission intentions: The weights are always equal to 1 to ensure consistency of dimensions. This strategy allows the same optimization framework to be adapted to different action scenarios.

[0042] In one implementation, after the execution layer completes the calculation of the temporary outpost, the transit process of the UUV to the temporary outpost is initiated. To balance mobility efficiency and cluster security, a phased mobility strategy is adopted: Phase 1 (High-speed convergence phase): All UUVs assigned to adjust formation depart from their current positions at maximum sustainable speed. Navigate along a straight line or a pre-planned safe path towards the target temporary outpost area. The focus at this stage is to rapidly reduce the distance to the target area and shorten the defense window. The navigation module receives the target coordinates from central commands in real time. , It generates heading instructions by combining its own positioning information.

[0043] Phase Two (Fine Positioning and Collision Avoidance Phase): When the distance between the UUV and the target temporary outpost is less than a preset threshold... It automatically switches to deceleration mode, and the speed decreases linearly or exponentially. Simultaneously, the dynamic collision avoidance algorithm is activated: each UUV periodically broadcasts its position, velocity vector, and target point via underwater acoustic communication; the local collision avoidance module calculates the local avoidance velocity increment in real time based on an improved Velocity Obstacle (VO) method or an artificial potential field method, ensuring that the distance between any two UUVs is always greater than the safe distance. If a potential conflict is detected, the system will prioritize the passage of UUVs with higher functional roles (such as rapid response sentries).

[0044] In one implementation, the static sentry state includes two sub-states: original static sentry: located at the initial optimal sentry position; temporary static sentry: located at the reconstructed temporary sentry position; by distinguishing between the two, the system ensures that the homing command is only issued to UUVs located at the temporary sentry position.

[0045] To support precise relocation control after the task is completed, the system implements refined management of the "static sentinel state," dividing it into two mutually exclusive sub-states: Original static sentinel: refers to the UUV being located in the globally optimal sentinel position generated by the initial static deployment optimization algorithm. Typically, this is calculated and fixed all at once before the task begins; Temporary static sentinels: These refer to UUVs that have completed formation reconfiguration and are currently stationed at temporary optimal sentinel positions generated by the static deployment optimization algorithm. Its status is explicitly marked after the instruction issuing module completes its task.

[0046] The system maintains a UUV status table, where each record contains the following fields: UUV_ID, current_position, assigned_role, and state_flag (values: ORIG / TEMP / DISPATCHED). When a UUV completes its interception mission, returns, and re-enters static guard status, the status monitoring module marks it as a "temporary static sentinel." Subsequently, the system only iterates through the status table for UUVs with state_flag == TEMP, issuing them instructions to return to their original sentinel positions; while UUVs that were not originally involved in the reconstruction (state_flag == ORIG) remain unaffected and continue to guard their original positions.

[0047] This mechanism avoids unnecessary movement caused by "all units returning to their positions," effectively reduces the total energy consumption during the return phase when resources are limited, and prevents formation chaos caused by erroneous commands.

[0048] In one implementation, to address the continuous penetration threat in a highly dynamic battlefield environment, the system implements a rolling dynamic adaptation mechanism, the specific implementation process of which is as follows: Once the three-layer collaborative refactoring process is initiated, the system enters the "Refactoring in Progress" state and starts a high-priority event listening thread to continuously monitor the instruction stream from the task assignment module. If a new intercept assignment instruction is received during this period, and this instruction causes another UUV to switch from a static sentinel state to a dispatched state, the following interruption and replanning logic will be triggered immediately: Interrupt the current refactoring process: The system immediately terminates the ongoing formation template selection, role assignment, or sentry optimization calculation (regardless of whether it is at the global layer, task layer, or execution layer), releases the relevant computing resources, and discards any temporary sentry plans that have not yet been issued.

[0049] Update cluster state snapshot: The state monitoring module synchronously refreshes the UUV cluster member list, marks newly assigned UUVs as "dynamically removed," and removes them from the current defense logic sequence. At this time, the number of UUVs participating in the reconstruction is... Updated to (or fewer, if there are multiple new assignments).

[0050] Recalculate coverage gaps: Based on the updated UUV ensemble and its capability profile, combined with the latest threat heatmap (which may change due to the emergence of new targets), reassess the uncovered areas along the strike line.

[0051] Perform the three-layer collaborative reconstruction again: with the updated cluster state as input, completely rerun the three-layer collaborative reconstruction. The global layer reselects the formation template based on the latest threat distribution; the task layer reassigns functional roles based on the remaining UUV capabilities; and the execution layer calls the static deployment optimization algorithm to generate a new set of temporary optimal sentry posts.

[0052] Issuing new instructions and continuing monitoring: New transit instructions are only issued to UUVs that are currently in a static state. At the same time, the event listening thread remains active, supporting multiple rounds of continuous interruption and reentry, forming a rolling closed loop of "perception-decision-interruption-re-decision".

[0053] In one implementation, assessing the impact of absence includes calculating the estimated return time of the assigned UUV. and the dynamic reconstruction trigger threshold Compare; if If the original sentry post is not returned within the specified time, the virtual post is retained as a virtual placeholder, and formation reconfiguration is suppressed; if the post does not return within the specified time, the virtual placeholder is released and reconfiguration is initiated. Dynamic reconfiguration trigger threshold. Calculate using the following formula: ;in, This represents the maximum displacement distance of the UUV in the reconstruction scheme. The average cruising speed under the current hydrological conditions. Let k be the average delay of the underwater communication link, and k be the stability coefficient. This refers to the algorithm response time related to the computational resource load.

[0054] Specifically, when the system detects that a UUV has switched from a static sentinel state to a deployed state, the virtual placeholder manager immediately initiates the off-duty impact assessment process. The core of this process is to determine whether this off-duty departure warrants triggering a full-scale formation reconfiguration. This judgment is based on a quantitative comparison of the reconfiguration cost and the risk of a defensive gap. The virtual placeholder assessment process is as follows: Estimated return time Calculation: The task assignment module estimates the estimated return time of the UUV in real time based on the following parameters. The target's current position, velocity vector, and predicted trajectory; the UUV's maximum speed. Based on the current remaining battery power and hydrological conditions, adjust the effective navigation speed; calculate the interception path length. and return route length Preset interception strategies (such as pursuit paths and engagement times); return path distance and ocean current correction models. Typical calculation formulas are as follows: ;in Distance to the interception point To preset the engagement time, such as setting it to 30 seconds, This is the return distance.

[0055] Threshold Comparison and Decision: The system synchronously calls the environment awareness interface to obtain the current operating parameters and calculates the threshold according to the following formula: ;in, The maximum displacement distance of the UUV in the reconstruction scheme is given. If reconstruction is to be performed immediately, the execution layer will pre-calculate the displacement of the UUV that needs to move the longest distance among the remaining UUVs. The average cruising speed under the current hydrological environment is obtained by referring to tables or through CFD simulation based on the current depth, temperature, salinity, and ocean current model. The average delay of the underwater communication link is given by k, which is a stability coefficient and a margin reserved to deal with model errors. Calculate the time taken by the algorithm.

[0056] Decision-making and execution: If If the UUV is temporarily absent from its post, its original outpost is marked as a virtual occupancy, and its detection capability is retained in the threat coverage model. Simultaneously, a "suppression signal" is sent to the status monitoring module and the three-layer collaborative reconstruction module to prevent triggering the normal formation reconstruction process. If the post is determined to be "absent from duty for an extended period," the post will be immediately released, triggering a complete formation reconfiguration process; the system will start a countdown timer. , For safety margin, if the UUV is not in If the UUV returns to its home port, the virtual placeholder will be forcibly released and a reconstruction will be initiated to prevent permanent defense vulnerabilities caused by UUV damage.

[0057] The role of virtual outposts: During the validity period of the virtual outpost, the location is still included in the defense coverage model. That is, in threat assessment and subsequent task assignment, the area is regarded as still being effectively covered, thereby avoiding a chain of reconstructions caused by a single point of absence.

[0058] Timeout handling and refactoring startup: The system continuously monitors the actual status of the UUV; if it is in... If the system successfully returns to base and reverts to a static sentry within the specified time, the virtual marker is automatically cleared, and the defense system seamlessly returns to its original state. If the system fails to return within the specified time (possibly due to mission extension, malfunction, or destruction), the system releases the virtual placeholder and removes the sentry from the coverage model, thereby triggering a complete formation reconstruction process, with the remaining UUVs filling the real defense gaps.

[0059] In one embodiment, the present invention also provides a multi-UUV dynamic formation reconfiguration system based on a collaborative sentinel mechanism, comprising: a status monitoring module for real-time monitoring of the status of the multi-UUV cluster, triggering an evaluation process when a UUV is detected to switch from a static sentinel state to a task execution state; and a virtual placeholder manager for evaluating based on the expected return time. With dynamic reconstruction trigger threshold The comparison results determine whether to enable virtual occupancy and suppress reconstruction; the threat assessment module is used to generate a threat heat map distribution along the strike line based on the historical positions obtained by the reconnaissance UUVs along the identified line or the real-time predicted arrival point density; the three-layer collaborative reconstruction module includes: a global layer unit configured to select a formation template based on the threat heat map distribution; a task layer unit configured to assign functional roles based on UUV capability profiles; an execution layer unit configured to call a static deployment optimization algorithm to calculate a temporary optimal sentry set with the formation template and functional roles as constraints; and an instruction issuance module used to issue phased transit instructions to the UUVs participating in the reconstruction and to issue return instructions to the temporary static sentries after the mission UUVs return; wherein, the system supports detecting new task assignments during the reconstruction process and triggering a rolling dynamic adaptation process.

[0060] Specifically, the multi-UUV dynamic formation reconfiguration system based on the collaborative sentinel mechanism is deployed at the underwater command and control node, and the modules operate collaboratively as follows: During system initialization, all UUVs are assigned to their original optimal sentry positions according to the static deployment optimization algorithm and enter the "original static sentry" state. The status monitoring module continuously receives status heartbeat packets reported by each UUV through the underwater acoustic communication network and maintains the cluster status table in real time.

[0061] When a UUV receives an intercept command and switches to the "dispatched" state, the status monitoring module immediately captures the event and notifies the virtual placeholder manager. The latter then estimates the expected return time based on the task parameters. :like If the original location of the UUV is marked as a virtual sentry post in the sentry post database, a "suppress reconstruction" signal is sent to the threat assessment module and the three-layer collaborative reconstruction module; otherwise, the complete reconstruction process is triggered.

[0062] If reconstruction is permitted, the threat assessment module integrates historical crossing points obtained by reconnaissance UUVs along the identified line with predicted arrival points provided by the real-time intelligence module, and uses kernel density estimation to generate a threat heat map along the strike line. Subsequently, the three-layer collaborative reconstruction module works in sequence: the global layer unit analyzes the heat map distribution pattern and selects a honeycomb contraction, linear extension, or wingspan enhancement template; the mission layer unit assigns rapid response sentry, core perception sentry, or persistent defense sentry roles based on the maximum speed, detection radius, and remaining endurance of each remaining UUV; and the execution layer unit calls the built-in multi-objective optimization algorithm to calculate the temporary optimal sentry post set under the triple constraints of coverage, response, and security.

[0063] Based on this, the command issuing module generates phased transit commands: the first phase commands the UUVs to converge towards the target area at maximum speed; the second phase automatically decelerates within 100 meters of the temporary sentry post and initiates a dynamic collision avoidance algorithm based on the speed obstacle method to ensure the safe entry of the multi-UUV cluster. At the same time, the system updates the status of these UUVs to "temporary static sentry".

[0064] When a deployed UUV completes its mission, returns to base, and reverts to a static sentry position, the status monitoring module detects the return-to-position event. The command issuance module then issues a command to all "temporary static sentries" to return to their original positions. Once all UUVs have returned to their positions, the system clears the virtual placeholder markers, and the defense system is restored to its full strength.

[0065] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A multi-UUV dynamic formation reconfiguration method based on a cooperative sentinel mechanism, characterized in that, Includes the following steps: S1. Real-time monitoring of the status of multiple UUV clusters in the initial defense formation; When any UUV is detected to switch from static sentinel state to mission execution state, the impact of its absence on the defense system is assessed based on its expected return time and the system dynamic reconfiguration threshold. S2. If the evaluation results meet the reconstruction trigger conditions, the physical array reconstruction procedure will be started. Otherwise, its original sentinel position is preserved and reconstruction is suppressed through a virtual placeholder mechanism; S3. The physical formation reconstruction procedure includes: generating a temporary formation based on the capability profile of the remaining UUVs and the threat heat distribution on the current strike line through hierarchical decision-making, and issuing sentry post adjustment instructions to the relevant UUVs; S4. During the refactoring process, if a new UUV is detected to be assigned a task, the current process is interrupted and the evaluation and refactoring process is re-executed based on the updated cluster state to achieve rolling dynamic adaptation. S5. After the mission UUV returns to base, perform local self-healing or global repositioning operations dynamically based on the overall cluster status to restore the integrity of the defense system.

2. The method of claim 1, wherein, The hierarchical decision-making adopts a three-tier collaborative architecture, specifically including: At the global level, a matching formation template is selected based on the threat heat distribution of targets on the current strike line. The formation template includes honeycomb contraction type, linear extension type, or wingspan enhancement type. At the mission level, based on the capability profile of each UUV, a functional role is assigned to it. The capability profile includes maximum speed, detection radius and remaining endurance. The functional roles include rapid response sentry, core perception sentry or persistent defense sentry. At the execution layer, constrained by the array template and functional roles, the static deployment optimization algorithm is invoked to calculate a set of temporary optimal sentry positions for the new cluster.

3. The method of claim 2, wherein, The threat thermal distribution is generated in any of the following ways: Based on the historical location of the target obtained by reconnaissance UUVs along the detection line, or the predicted arrival point density estimate provided by the real-time intelligence module; and when there is a heat peak area in the threat heat distribution, the area is determined to be a high-risk segment, triggering the honeycomb contraction array template, causing UUVs to be deployed in a concentrated manner to the area.

4. The method of claim 2, wherein, The capability profile includes maximum speed, detection radius, and remaining endurance; the allocation rules for the functional roles are as follows: If the maximum speed of the UUV is higher than the cluster average and there is sufficient remaining endurance, it will be assigned as a rapid response sentry and deployed at the front of the formation. If the UUV has the largest detection radius, it will be assigned as the core sensing sentry and deployed at the center of the formation; If a UUV has the longest endurance but the lowest speed, it will be assigned as a long-term defensive outpost and deployed in a rear support position.

5. The method of claim 2, wherein, The static deployment optimization algorithm aims to minimize the following multi-objective cost function when calculating temporary outposts: ； wherein, to penalize the area of uncovered lines; to be the weighted average response time; to be a penalty term for the distance between UUVs being less than a safety threshold; the weight , , adjusted dynamically according to the current formation template.

6. The method according to claim 1, characterized in that, The evaluating the impact of the off-duty on the defense system includes calculating the expected return time of the assigned UUV and comparing with a dynamic reconfiguration trigger threshold ; if , its original sentry bit is kept as a virtual placeholder and the formation reconfiguration is suppressed; if the timeout is not returned, the virtual placeholder is released and the reconfiguration is started.

7. The method according to claim 6, characterized in that, the dynamic reconfiguration trigger threshold is calculated as follows: ； wherein, is the maximum displacement distance of the UUV in the reconstruction scheme, is the average cruising speed under the current hydrological environment, is the average delay of the underwater communication link, k is the stability coefficient, is the algorithm response time related to the computing resource load.

8. The method of claim 2, wherein, After generating a temporary optimal set of sentry posts, the execution layer issues sentry post adjustment instructions to the UUV using a phased movement strategy: Phase 1: Converge towards the target region at maximum speed; Phase 2: Decelerate and fine-tune as you approach the temporary outpost, while simultaneously executing a dynamic collision avoidance algorithm.

9. The method according to claim 1, characterized in that, The static sentinel state includes two sub-states: Original static sentinel: located at the initial optimal sentinel position; Temporary static sentry: Located at the reconstructed temporary sentry post; The system distinguishes between the two to ensure that the return command is only issued to UUVs at temporary outposts.

10. A multi-UUV dynamic formation reconfiguration system based on a cooperative sentinel mechanism, characterized in that, include: The status monitoring module is used to monitor the status of multiple UUV clusters in real time. When it detects that a UUV has switched from a static sentinel state to a task execution state, it triggers the evaluation process. Virtual placeholder manager, used for determining the expected return time With dynamic reconstruction trigger threshold The comparison results determine whether to enable virtual placeholders and suppress refactoring; The threat assessment module is used to generate a threat heat map of the strike line based on the historical locations or real-time predicted arrival point density obtained by reconnaissance UUVs along the identified line. The three-layer collaborative reconstruction module includes: a global layer unit, configured to select a formation template based on threat heat distribution; a task layer unit, configured to assign functional roles based on UUV capability profiles; and an execution layer unit, configured to call a static deployment optimization algorithm to calculate a temporary optimal set of sentry posts with the formation template and functional roles as constraints. The instruction issuing module is used to issue phased transit instructions to the UUVs participating in the reconstruction, and to issue return-to-position instructions to the temporary static sentinels after the mission UUVs return to base. The system supports detecting new task assignments during the refactoring process and triggering a rolling dynamic adaptation process.

Images