An underwater unmanned vehicle cluster patrol area allocation method and system

Abstract

The application discloses an underwater unmanned vehicle cluster patrol area distribution method and system, and relates to the technical field of underwater unmanned cluster task distribution. First, based on the ocean current field, the submarine topography and the power and sensor capacity of each UUV, an equivalent patrol length model considering the ocean current resistance, terrain shelter and heterogeneous detection performance is constructed; a non-overlapping patrol boundary is generated through a virtual coverage force field, and a triple check mechanism is introduced in the distributed negotiation to reduce the dependence on underwater acoustic communication; at the same time, the distribution strategy is dynamically optimized in combination with the actual coverage effect, and when the UUV fails, the adjacent unit takes over the area according to the capacity and locally reconstructs the force field. The scheme does not depend on the prior distribution of the target, and effectively overcomes the problems of coverage blind area, resource imbalance and the like caused by ignoring the underwater environment constraints in the traditional method by strengthening the physical feasibility of the cluster itself in the complex underwater environment, and significantly improves the patrol robustness, persistence and anti-destroying capacity of the cluster in the complex marine environment.

Description

Technical Field

[0001] This invention relates to the field of underwater unmanned swarm mission allocation technology, and in particular to a method and system for allocating patrol areas for underwater unmanned vehicle swarms. Background Technology

[0002] In the collaborative execution of blockade or patrol missions by underwater unmanned vehicle (UUV) swarms, efficiently allocating patrol areas for each UUV to maximize the probability of detecting potential crossing targets is a key technical challenge in the field of underwater collaborative control. Existing methods mostly divide areas based on the prior probability density function (PDF) of the target crossing location, prioritizing resource deployment in high-probability areas. However, in real-world adversarial environments, enemy paths are highly uncertain and they actively avoid known patrol patterns, making the pre-set PDF difficult to obtain, prone to failure, or even maliciously deceiving. Furthermore, these methods generally ignore the unique physical constraints of the underwater environment, such as the asymmetric impact of ocean currents on UUV energy consumption, the obstruction effect of seabed topography on sonar detection, and the low bandwidth and high latency characteristics of underwater acoustic communication. This results in allocations that, while mathematically optimal, cannot be implemented in practice due to rapid battery depletion by UUVs operating against the current, lack of coverage in key areas, or communication negotiation failures, severely weakening the robustness and sustainability of swarm missions.

[0003] To address the aforementioned challenges, existing research has attempted to improve the system by introducing optimization algorithms or formation mechanisms. For example, CN120764801A proposes a dynamic task planning method for UUV swarms based on a multi-objective genetic algorithm, adjusting sub-region weights through target density; CN119024846A discloses a multi-UUV formation transformation method based on energy consumption constraints, employing the Hungarian algorithm to achieve low-energy formation transformation and minimize displacement costs; CN111930116B discloses a large-scale UUV swarm formation formation method based on a grid method, utilizing gridding and particle swarm optimization to achieve ordered formation of large-scale swarms. Although these solutions improve task efficiency or formation stability in specific scenarios, their core still relies on external target distribution information or only focuses on geometric configuration, failing to deeply couple ocean currents, terrain, sensor performance, and communication constraints into the patrol area allocation model, and lacking dynamic feedback and self-correction capabilities for execution effects. Therefore, there is an urgent need for a UUV swarm patrol area allocation technology that does not rely on prior target distribution, embeds underwater physical constraints, and possesses dynamic evolution capabilities. Summary of the Invention

[0004] In view of this, the present invention proposes a method and system for allocating patrol areas for underwater unmanned vehicle swarms. By enhancing the physical feasibility and coverage robustness of the swarm itself in complex underwater environments, it effectively overcomes the problems of coverage blind spots, resource imbalances, and mission vulnerability caused by neglecting underwater environmental constraints in traditional methods, and significantly improves the patrol robustness, sustainability, and survivability of the swarm in complex marine environments.

[0005] This invention provides a method for allocating patrol areas for an underwater unmanned vehicle swarm, comprising the following steps: S1. Acquire underwater environmental current data, digital elevation model (DEM), and status information and sensor performance parameters of each UUV in the target water area; S2. Based on the underwater environment ocean current data, DEM and the status information of each UUV and sensor performance parameters, calculate the equivalent patrol length of each UUV. The calculation model of the equivalent patrol length quantifies the asymmetric impact of ocean current on UUV navigation energy consumption and compensates for terrain obstruction and sensor performance differences. S3. Construct a virtual coverage force field based on the equivalent patrol length of each UUV, and determine the patrol boundary between adjacent UUVs according to the virtual coverage force field; S4. During the distributed negotiation process, multi-dimensional security checks are performed on the boundary adjustment command only when the boundary change exceeds the preset trigger threshold. S5. Control each UUV to perform patrol tasks according to the verified patrol boundaries, and dynamically correct the equivalent patrol length and patrol boundaries based on the actual coverage performance feedback. S6. When a UUV failure is detected, the regional takeover and local force field reconstruction process is triggered.

[0006] Furthermore, the status information of each UUV includes its location, remaining battery power, speed, and heading; The equivalent patrol length of each UUV The calculation formula is: ; in, For dynamic safety margin, and satisfying , Based on the safety margin, This is the uncertainty scaling factor. Uncertainty in ocean current prediction; The sensor performance factor is given by the following condition: , Let be the effective sonar detection width of the i-th UUV. The base width; ρ is the remaining power, and c is the energy consumption coefficient. Let γ be the speed, and γ be the drag exponent with γ = 3.0. The energy consumption coefficient of ocean currents. For the i-th UUV position, The ocean current velocity at the location of the UUV. Let the angle between the i-th UUV's heading and the ocean current be... For losses during round-trip voyages, It is a terrain shading penalty term, and satisfies... ,in, This is the terrain shadow penalty coefficient. For DEM and sonar heading Calculate the sonar beam obstruction area.

[0007] Furthermore, the intensity function of the virtual covering force field for: in, Let x be the intensity of the coverage force field generated by the i-th UUV at position x. The ocean current attenuation coefficient is... The radius of force field diffusion and L is the total length of the control line, and N is the total number of UUVs. Let i be the position of the i-th UUV.

[0008] Furthermore, the patrol boundary The solution formula is: ; in, , Let be the ocean current attenuation factors at the locations of the i-th and j-th UUVs, respectively. This is the ocean current offset coefficient. For prediction time windows.

[0009] Furthermore, the multi-dimensional security verification includes: Geometric verification: Verify whether the boundary position in the verification command is within the physical reach of the UUV; Capability verification: Verify whether the adjusted patrol interval length exceeds the preset threshold of the UUV equivalent patrol length; Historical verification: Verify whether the variation of the boundary position relative to the historical boundary is within the allowable range; If any verification fails, discard the boundary adjustment instruction and trigger a security alert.

[0010] Furthermore, the dynamic correction includes: Calculate relative deviation ,in For actual coverage effectiveness, To achieve the expected coverage effectiveness; when When the percentage is >15%, update the equivalent patrol length: λ=0.2; based on Reconstruct the virtual overlay force field and update adjacent boundaries.

[0011] Furthermore, the ocean current data is obtained through the following methods: Load ocean current maps from a pre-stored marine environment database; Local ocean current data is sampled in real time by an ADCP device mounted on a UUV and updated after being filtered by a sliding window. When ocean current data is missing, the ocean current-related items are set to zero, degenerating into a mode without ocean current compensation.

[0012] Furthermore, the distributed negotiation process employs a layered negotiation mechanism: UUVs are clustered into several subgroups based on geographical location, with each subgroup having a size of 3 to 5. Each seed group elects the UUV with the strongest election ability as its representative; Virtual coverage force field equalization is performed within the subgroup, and the multi-dimensional security verification is performed between subgroups by representing the UUV exchange endpoint coordinates. When the number of UUVs in a subgroup exceeds the threshold or the gap continues to exceed the limit, dynamic subgroup re-partitioning is triggered.

[0013] Furthermore, the actual coverage effectiveness The number of effective UUV detection events per unit time is calculated. It is executed once after each round of patrol mission or every 30 minutes. When the feedback correction amount is less than 5% for three consecutive rounds, dynamic correction is paused to reduce computational overhead.

[0014] Furthermore, the present invention also provides an underwater unmanned vehicle (UUV) swarm patrol area allocation system, comprising: a data acquisition unit for acquiring ocean current data of the target water area and the status information of each UUV; a force field construction unit for calculating the equivalent patrol length based on the ocean current data and the status information of each UUV, and constructing a virtual coverage force field associated with the equivalent patrol length and the ocean current data; a boundary negotiation unit for determining the patrol boundary between adjacent UUVs according to the virtual coverage force field, and performing multi-dimensional security verification on the boundary adjustment command during the distributed negotiation process; and a dynamic optimization unit for dynamically correcting the equivalent patrol length and patrol boundary based on the actual coverage performance feedback of each UUV.

[0015] The present invention has the following advantages over the prior art: (1) Significantly improve the mission survival rate and endurance balance of UUV clusters in strong ocean current environment. By introducing the ocean current direction performance consumption compensation term in the calculation of equivalent patrol length, UUVs sailing against the current are automatically assigned a shorter patrol interval, avoiding premature power exhaustion due to increased energy consumption, effectively improving the overall power consumption time of the cluster, and achieving a balance between the energy and mission load of the cluster.

[0016] (2) Achieving robust, blind-spot-free coverage of the entire control line. This invention does not rely on the external target probability density function PDF to resist enemy path deception and sudden environmental changes. Instead, it uses terrain shadow penalty terms and sensor performance factors to calibrate the patrol area, ensuring that high-risk terrain areas and weak detection areas receive sufficient coverage resources. At the same time, the dynamic feedback correction mechanism automatically adjusts the force field parameters and reconstructs the boundary when it detects that the actual coverage performance deviation exceeds 15%, and has the ability to adapt online to unmodeled disturbances such as sudden ocean current changes and thermocline interference.

[0017] (3) To ensure the security and efficiency of negotiation under low-bandwidth underwater communication conditions, a triple lightweight security verification mechanism and event-triggered distributed negotiation are designed to 100% intercept malicious tampering or conflict commands without increasing significant communication overhead, while effectively reducing the communication frequency between subgroup representatives. This mechanism is specifically optimized for the high latency and low throughput characteristics of underwater acoustic channels, ensuring that the cluster can still achieve consensus and reliable allocation in an anti-interference environment.

[0018] (4) Enhance the task continuity and resilience of the cluster system. When a single UUV fails due to a fault or power depletion, the system automatically triggers a regional takeover strategy based on the equivalent patrol length ratio. The neighboring UUVs share the patrol range according to their capabilities, and only partially reconstruct the virtual coverage force field to avoid communication storms and task interruptions caused by global redistribution, thus significantly improving the survival and task resilience of the cluster in long-term deployment. 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 a diagram illustrating the overall allocation process architecture of an embodiment of the present invention; Figure 2 This is an execution 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 1 As shown, the present invention provides a method for allocating patrol areas for an underwater unmanned vehicle swarm, comprising the following steps: S1. Acquire underwater environmental current data, digital elevation model (DEM), and status information and sensor performance parameters of each UUV in the target water area.

[0023] The acquisition of ocean current data employs a priori and real-time fusion strategy, balancing broad coverage with local accuracy: First, a large-scale ocean current map pre-stored in the marine environment database of the shore-based system is loaded as the global background field; simultaneously, at least one UUV in the cluster is equipped with an Acoustic Doppler Current Profiler (ADCP), which collects a three-dimensional current profile of its location at fixed time intervals during navigation, preferably set to every 10 seconds; the ADCP sampling data is filtered by a sliding window variance and used to dynamically correct the values ​​of the ocean current map in local areas, generating a spatiotemporally consistent fused ocean current field; when the ADCP fails or communication is interrupted, resulting in the loss of local ocean current data, the system automatically sets the ocean current-related terms in the force field model to zero, degenerating into a no-current-compensation mode to ensure that the allocation process is not interrupted.

[0024] S2. Based on the underwater environment ocean current data, DEM and the status information of each UUV and sensor performance parameters, calculate the equivalent patrol length of each UUV. The calculation model of the equivalent patrol length quantifies the asymmetric impact of ocean current on UUV navigation energy consumption and compensates for terrain obstruction and sensor performance differences. Based on this, an equivalent patrol length is constructed through a ternary mapping relationship of energy, environment, and task. This is achieved through a hierarchical modeling approach driven by physical constraints, determining the equivalent patrol length for each UUV. The calculation formula is: ; in, For dynamic safety margin, and satisfying , Based on the safety margin, This is the uncertainty scaling factor. Uncertainty in ocean current prediction; The sensor performance factor is given by the following condition: , Let be the effective sonar detection width of the i-th UUV. The base width; ρ is the remaining power, and c is the energy consumption coefficient. Let γ be the speed, and γ be the drag exponent with γ = 3.0. The energy consumption coefficient of ocean currents. For the i-th UUV position, The ocean current velocity at the location of the UUV. Let the angle between the i-th UUV's heading and the ocean current be... For losses during round-trip voyages, It is a terrain shading penalty term, and satisfies... ,in, This is the terrain shadow penalty coefficient. For DEM and sonar heading Calculate the sonar beam obstruction area.

[0025] Specifically, UUVs are located in When performing patrol missions, its energy consumption per unit distance is proportional to the product of propulsion and speed; and the propulsion must overcome hydrostatic resistance. The additional drag caused by ocean currents. This is because ocean currents only have an angle with the UUV's course. Effective counteracting components are generated, therefore an ocean current projection term is introduced. and with coefficient This characterizes its amplification effect on energy consumption. Therefore, the energy consumption per unit length is precisely expressed as... When UUVs travel against the current ( Energy consumption increases significantly during downstream flow ( ) Energy consumption is reduced, thus truly reflecting the asymmetric physical characteristics of the underwater environment, where it is difficult to navigate against the current and easy to navigate with the current.

[0026] Secondly, to compensate for the sensor performance differences in heterogeneous UUVs, a sensor performance factor is introduced. ,in Let be the effective sonar detection width of the i-th UUV. This is a preset baseline width, such as the minimum sonar width in the cluster. This factor maps UUVs with different detection capabilities to a unified mission efficiency scale. The wider the detection width, the larger the lateral sea area that can be covered per unit patrol length, and the stronger its equivalent mission capability.

[0027] Finally, to compensate for the obstruction effect of seabed topography on sonar detection, a digital elevation model (DEM) and the UUV sonar beam pointing angle were used as the basis for the detection. The effective unobstructed projected area of ​​the sonar main lobe along the patrol path is calculated using ray tracing or geometric occlusion algorithms, thereby deriving the detection performance loss caused by terrain occlusion. Specifically, the ray tracing or geometric occlusion algorithm can be given the UUV depth. and sonar beam angle The projection range of its sonar main lobe on the seabed is: Within this projection range, the DEM data is sampled one-dimensionally along the heading direction to obtain the seabed height sequence H(x); if a point x satisfies Where Δh is the minimum sonar detection range, the point is considered to be in the acoustic blind zone. This loss is penalized by a term... The form is deducted from the total available patrol length, of which To cover the area, This is based on experience weighting. Through this mechanism, the system automatically avoids assigning high-value patrol zones to acoustic blind spots.

[0028] In this invention, UUVs are reduced from ideal moving bodies to finite resource carriers under strong environmental constraints, and physically reliable capability weights are provided through a differentiable mathematical form. Equivalent patrol length. This does not refer to the actual physical distance that a UUV can navigate, but rather the equivalent length of a mission that a UUV can effectively undertake under current conditions, after comprehensively considering underwater environmental constraints and platform capabilities. Its core concept is to transform the finite resource of electricity into a comparable and allocable mission capability indicator through an environment-platform coupling model.

[0029] S3. Construct a virtual coverage force field based on the equivalent patrol length of each UUV, and determine the patrol boundary between adjacent UUVs according to the virtual coverage force field; the intensity function of the virtual coverage force field. for: ; in, Let x be the intensity of the coverage force field generated by the i-th UUV at position x. The ocean current attenuation coefficient is... The radius of force field diffusion and L is the total length of the control line, and N is the total number of UUVs. This is the location of the i-th UUV. The patrol boundary. The solution formula is: ; in, , Let be the ocean current attenuation factors at the locations of the i-th and j-th UUVs, respectively. This is the ocean current offset coefficient. For the prediction time window. Ocean current offset coefficient. It is used to compensate for the prediction time window The feedforward gain factor of the influence of the internal ocean current on the boundary position should be used to shift the patrol boundary in the direction of the ocean current in advance when the ocean current is strong, so as to prevent the coverage gap from being generated due to the target drifting with the current. η The value can be That is, the ratio of the maximum speed of the ocean current to the maximum speed of the UUV; or ,in, This refers to the sensor response delay.

[0030] Specifically, discrete UUV capability indicators, i.e., equivalent patrol length The force field is mapped to an influence distribution in a continuous space, thereby automatically generating non-overlapping, capability-adaptive patrol boundaries through a force field balance mechanism. Its implementation involves two tightly coupled sub-processes: virtual coverage force field construction and patrol boundary solution.

[0031] The virtual covering force field intensity function is defined as: ; By leveraging the triple physical intuitions of capability centralization, environmental degradation, and spatial diffusion, a differentiable and superimposed task influence proxy model is constructed to quantify the effective patrol capability that the i-th UUV can provide at location x. Capability centralization refers to the ability of the UUV to patrol effectively at its current location. As the center of the force field, its maximum influence is equal to its equivalent patrol length. This reflects the allocation principle of "the stronger the capability, the larger the dominant area." Environmental degradation essentially introduces an ocean current attenuation factor. This indicates that at position x, if there exists an angle between the x-axis and the UUV's heading... For ocean currents, their effective coverage capacity decreases exponentially with the intensity of the current. Spatial diffusion refers to the use of a Gaussian kernel. Simulate the natural decay of coverage influence, where the diffusion radius The total length of the control line L and the total number of UUVs N are jointly determined to ensure that the force field scale matches the mission scale. That is, the larger the cluster, the smaller the influence range of a single UUV, thus avoiding excessive overlap of force fields.

[0032] The solution mechanism and dynamic compensation of the patrol boundary: the patrol boundary between adjacent UUVi and j Defined as the position where the force field strengths of the two forces are equal, i.e., satisfying... By taking the logarithm of the above equation and simplifying it, we can obtain the analytical solution: ; in Let be the ocean current attenuation factor for the i-th UUV at its own location. The first two terms... and The midpoint term constitutes the equilibrium point of the static force field. For the ideal boundary under no environmental disturbance; for several terms Introducing a capability asymmetry correction, if the ocean current is stronger at the location of the i-th UUV... The boundary then shifts towards i, allowing it to cover a smaller interval; the third term is ocean current dynamic offset compensation, considered within the prediction time window. Inland, ocean currents will drive UUVs or targets to move, so the boundary should be shifted in advance along the direction of the ocean current. This enables proactive allocation and avoids the rapid invalidation of initial allocations due to environmental dynamics.

[0033] This step integrates the individual capability modeling of S2 into the cluster collaborative execution level. By constructing a physically reliable, mathematically concise, and computationally efficient virtual force field model, the complex multi-constraint allocation problem is transformed into an elegant field theory solution process, achieving the goals of non-overlapping allocation, ocean current attenuation and dynamic offset to automatically adapt the boundary to the underwater flow field, coverage effectiveness assessment and parameter correction.

[0034] S4. During the distributed negotiation process, multi-dimensional security verification is performed on the boundary adjustment command only when the boundary change exceeds the preset trigger threshold.

[0035] Specifically, this invention addresses the challenge of enabling underwater swarms to efficiently and securely achieve patrol boundary consistency in a low-bandwidth, high-latency, and interference-prone underwater acoustic communication environment. This invention reduces communication complexity through spatial clustering, decreases communication frequency through event triggering, and ensures negotiation reliability through multi-dimensional verification. The specific implementation includes the following four mutually coordinating sub-processes: (1) Layered negotiation architecture and subgroup construction. First, based on the one-dimensional position coordinates of UUVs on the control line. The cluster is divided into several geographically contiguous subgroups using a sliding window or K-nearest neighbor strategy. Each subgroup contains 3 to 5 adjacent UUVs, ensuring that the distance between members within a subgroup is much smaller than the distance between subgroups, forming a physical structure of strong local coupling and weak global coupling. The maximum subgroup size is 5 to avoid excessive internal negotiation overhead (the computational complexity of force field equilibrium increases quadratically with the number of members); the minimum subgroup size is 3 to ensure that even if a single UUV fails, the subgroup still has redundant negotiation capabilities.

[0036] (2) Subgroup representative election mechanism: each subgroup independently executes representative election driven by its own capabilities, and each UUV broadcasts its own equivalent patrol length. Compare all members within a subgroup. The UUV with the largest value is selected as the representative; the election result is confirmed via a short message, and all subsequent external communications of the subgroup are performed by this representative. This mechanism takes into account that the most capable UUV usually has more power, better sensors, and a more central location, making it naturally suitable as a coordination node.

[0037] (3) Layered negotiation and event triggering mechanism. The negotiation process is carried out in two layers: within the subgroup, all members independently solve the boundary with the left and right neighbors by constructing a virtual covering force field based on S3. and The endpoint coordinates (i.e., the leftmost / rightmost boundaries) of each subgroup are exchanged only by the representative UUVs of adjacent subgroups to verify the consistency of cross-subgroup connectivity.

[0038] The key optimization lies in the event triggering mechanism: each UUV continuously monitors the absolute change between the new computation boundary and the boundary of the previous cycle. Only when Only when the boundary adjustment instruction generation and verification process is initiated will the process be started; otherwise, the current allocation will be maintained, significantly reducing invalid communication.

[0039] When adjacent subgroups exchange endpoint coordinates, if a gap or overlap is detected between the two subgroups, a force field fine-tuning mechanism is triggered: the gap or overlap region is treated as a virtual transition segment; within this region, a weak force field node is temporarily introduced, with its equivalent patrol length set to a minimum value, such as 0.1. It participates in local force field equilibrium; by solving the equilibrium points represented by the node and the two end subgroups, the boundary is automatically adjusted to a seamless connection; after fine-tuning, the virtual node is removed and the original structure is restored.

[0040] (4) Multi-dimensional security verification process: After the boundary adjustment is triggered, the receiver UUV (or subgroup representative) performs triple verification on the instruction: Geometric verification: Determine new boundaries in the command. Whether it is within the physical reach of UUV, i.e., whether it meets the requirements. If the value exceeds the limit, it indicates that the allocation violates the energy constraint, and the verification fails.

[0041] Capability verification: Calculate the adjusted patrol section length Check if it meets the requirements. Slight overloading is permitted. To cope with dynamic disturbances, but to prevent severe overload.

[0042] Historical verification: Calculate the boundary variation rate ,in, For the allocation period, To preset the maximum boundary movement speed, if If the result is not found, it is considered an abnormal jump, which may be caused by a communication error or malicious attack, and the verification fails.

[0043] If any verification fails, the instruction is immediately discarded, a security log is recorded, and a "negotiation anomaly" event is reported to the subgroup representative, triggering local renegotiation or degraded operation.

[0044] (5) Subgroup dynamic repartitioning mechanism: In order to cope with task evolution, the system continuously monitors two types of abnormal states: subgroup size exceeds the limit, the number of members is <3 or >5 due to UUV joining / leaving; gaps continue to exceed the limit, there are uncovered gaps between adjacent subgroup endpoints that exceed the preset tolerance value and continue for more than 3 allocation cycles.

[0045] Once any of the above situations is detected, dynamic subgroup re-partitioning is triggered, and a "re-partitioning request" is broadcast globally. This request is initiated only by the neighbors of the failed UUV and is not broadcast to the entire network. All UUVs re-perform geographical clustering based on their latest locations. After the new subgroup completes representative election, hierarchical negotiation resumes.

[0046] S5. Control each UUV to perform patrol tasks according to the verified patrol boundaries, and dynamically correct the equivalent patrol length and patrol boundaries based on the actual coverage performance feedback.

[0047] Specifically, such as Figure 2 As shown, the deviation between the expected coverage performance and the actual coverage performance is transformed into an online correction signal for the capacity model, and stable convergence is achieved through a controlled update mechanism.

[0048] Actual coverage effectiveness Defined as: the number of valid detection events obtained by a UUVi per unit time within its assigned patrol area. A valid detection event refers to a detection record where the sonar echo signal-to-noise ratio is higher than a preset threshold and is confirmed by the target classification module as a potential crossing target. Actual coverage effectiveness. It mainly relies on the number of detection events, but when there is no target crossing for a long time, The variance may approach zero, causing dynamic correction to fail. Therefore, this invention introduces an alternative feedback signal: when there are no detection events for three consecutive rounds, the environmental noise stability index is activated to calculate the variance of the sonar echo signal. If the fluctuation is less than the threshold, the environment is considered stable and the current allocation is reasonable, and the correction is paused. At the same time, combined with the consistency of the bottom echo, if the sonar echo pattern matches the historical database with a degree of greater than 90%, it is determined to be a normal cruise state, and the current configuration is maintained.

[0049] Expected coverage effectiveness The theoretical detectability is estimated based on the virtual force field strength and sensor model during the allocation phase, representing the theoretical detectability under ideal conditions.

[0050] Once a complete cycle of performance data is completed, or every 30 minutes, the first data to arrive is recorded locally by the UUV and used for subsequent correction calculations.

[0051] The system calculates the relative deviation: ,in For actual coverage effectiveness, For expected coverage effectiveness; only when The correction process is initiated at a certain time. This threshold has been calibrated through extensive simulations. Values ​​below this threshold generally indicate random fluctuations and require no response; values ​​above this threshold indicate systematic modeling errors, such as sudden changes in ocean currents or insufficient compensation for terrain obstruction. If the correction amount is adjusted for three consecutive rounds... If all values ​​are less than 5%, dynamic correction will be paused and the system will enter steady-state mode to reduce the computational and communication overhead of the embedded platform.

[0052] When a correction is triggered, the updated formula is: .

[0053] The correction term symbol is ( The decision is made if the actual effectiveness exceeds expectations. This indicates that UUV capabilities have been underestimated and need to be increased. This grants it a larger patrol area; if the actual effectiveness is lower than expected, that is... This indicates an overestimation of capabilities or environmental degradation, leading to a reduction in [something unclear]. To prevent resource waste or coverage gaps. Introducing To ensure system stability, a single correction ratio is set to no more than 50%, preventing severe oscillations caused by sudden noise. The gain coefficient λ = 0.2 is the optimal compromise determined by Monte Carlo simulation; too large a value leads to oscillation, while too small a value results in slow convergence. This means that every 10% performance deviation causes approximately a 2% capability adjustment, achieving smooth and gradual optimization.

[0054] calculate Then, the UUV broadcasts it to neighboring members or subgroup representatives; all relevant UUVs reconstruct the virtual overlay force field based on the new capability value. Resolve for the force field equilibrium point and update the adjacent patrol boundaries. The new boundary takes effect after the security verification described in S4, and the UUV will perform the next cycle patrol according to the new instructions.

[0055] S6. When a UUV failure is detected, the regional takeover and local force field reconstruction process is triggered.

[0056] Specifically, this step ensures cluster task continuity through a lightweight fault-tolerant mechanism. When a UUV fails due to a heartbeat timeout or a subgroup representative's failure, the system immediately identifies its originally assigned patrol zone. Subsequently, this zone is divided according to the equivalent patrol length ratio of adjacent UUVs and assigned to the two UUVs on the left and right. Next, only the affected UUVs reconstruct their local virtual coverage force field based on the updated equivalent patrol length and solve for the new patrol boundary. The new boundary takes effect after the security check described in S4, while the remaining UUVs maintain their original tasks. This process involves only local communication and computation, avoiding global reallocation and ensuring rapid recovery of full coverage with minimal communication overhead under single-point failure.

[0057] This invention also provides an underwater unmanned vehicle (UUV) swarm patrol area allocation system, comprising: a data acquisition unit for acquiring ocean current data and status information of each UUV in a target water area; a force field construction unit for calculating the equivalent patrol length based on the ocean current data and the status information of each UUV, and constructing a virtual coverage force field associated with the equivalent patrol length and the ocean current data; a boundary negotiation unit for determining the patrol boundary between adjacent UUVs based on the virtual coverage force field, and performing multi-dimensional security verification on the boundary adjustment command during the distributed negotiation process; and a dynamic optimization unit for dynamically correcting the equivalent patrol length and patrol boundary based on the actual coverage performance feedback of each UUV.

[0058] Specifically, the data acquisition unit integrates pre-stored marine environment databases with real-time sampling data from UUVs equipped with ADCP (Advanced Digital Acoustic Processing Unit), generates a spatiotemporally consistent ocean current field after sliding window filtering, and combines this with the UUV's reported position, remaining battery power, heading, and sonar parameters to form a joint environment-platform state input. The force field construction unit calculates the equivalent patrol length of each UUV based on this input and substitutes it with a force field intensity function containing an ocean current attenuation term and a Gaussian diffusion kernel to generate a one-dimensional virtual coverage force field. The boundary negotiation unit exchanges endpoint coordinates among subgroup representatives, triggering a triple safety check (geometric, capability, and historical checks) only when boundary changes exceed a threshold. After successful checks, the non-overlapping patrol boundaries are updated. The dynamic optimization unit periodically collects the number of effective detection events per unit time as the actual coverage effectiveness. When the deviation from the expected value exceeds a threshold, the equivalent patrol length is corrected according to controlled nonlinear rules, driving local force field and boundary updates. All units operate collaboratively to achieve cluster patrol allocation that is independent of target priors, embeds underwater physical constraints, and possesses adaptive and survivable capabilities.

[0059] 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 method for allocating patrol areas in an underwater unmanned vehicle (UUV) swarm, the swarm comprising multiple heterogeneous UUVs, characterized in that, The method includes the following steps: S1. Acquire underwater environmental current data, digital elevation model (DEM), and status information and sensor performance parameters of each UUV in the target water area; S2. Based on the underwater environment ocean current data, DEM and the status information of each UUV and sensor performance parameters, calculate the equivalent patrol length of each UUV. The calculation model of the equivalent patrol length quantifies the asymmetric impact of ocean current on UUV navigation energy consumption and compensates for terrain obstruction and sensor performance differences. S3. Construct a virtual coverage force field based on the equivalent patrol length of each UUV, and determine the patrol boundary between adjacent UUVs according to the virtual coverage force field; S4. During the distributed negotiation process, multi-dimensional security checks are performed on the boundary adjustment command only when the boundary change exceeds the preset trigger threshold. S5. Control each UUV to perform patrol tasks according to the verified patrol boundaries, and dynamically correct the equivalent patrol length and patrol boundaries based on the actual coverage performance feedback. S6. When a UUV failure is detected, the regional takeover and local force field reconstruction process is triggered.

2. The method according to claim 1, characterized in that, The status information of each UUV includes its location, remaining battery power, speed, and heading; The equivalent patrol length of each UUV The calculation formula is: ； in, For dynamic safety margin, and satisfying , Based on the safety margin, This is the uncertainty scaling factor. Uncertainty in ocean current prediction; The sensor performance factor is given by the following condition: , Let be the effective sonar detection width of the i-th UUV. The base width; ρ is the remaining power, and c is the energy consumption coefficient. Let γ be the speed, and γ be the drag exponent with γ = 3.

0. The energy consumption coefficient of ocean currents. For the i-th UUV position, The ocean current velocity at the location of the UUV. Let the angle between the i-th UUV's heading and the ocean current be... For losses during round-trip voyages, It is a terrain shading penalty term, and satisfies... ,in, This is the terrain shadow penalty coefficient. For DEM and sonar heading Calculate the sonar beam obstruction area.

3. The method according to claim 2, characterized in that, The intensity function of the virtual covering force field for: in, Let x be the intensity of the coverage force field generated by the i-th UUV at position x. The ocean current attenuation coefficient is... The radius of force field diffusion and L is the total length of the control line, and N is the total number of UUVs. Let i be the position of the i-th UUV.

4. The method according to claim 3, characterized in that, The patrol boundary The solution formula is: ; in, , Let be the ocean current attenuation factors at the locations of the i-th and j-th UUVs, respectively. This is the ocean current offset coefficient. For prediction time windows.

5. The method according to claim 1, characterized in that, The multi-dimensional security verification includes: Geometric verification: Verify whether the boundary position in the verification command is within the physical reach of the UUV; Capability verification: Verify whether the adjusted patrol interval length exceeds the preset threshold of the UUV equivalent patrol length; Historical verification: Verify whether the variation of the boundary position relative to the historical boundary is within the allowable range; If any verification fails, discard the boundary adjustment instruction and trigger a security alert.

6. The method according to claim 2, characterized in that, The dynamic correction includes: Calculate relative deviation ,in For actual coverage effectiveness, To achieve the expected coverage effectiveness; when When the percentage is >15%, update the equivalent patrol length: ,λ=0.2; Based on equivalent patrol length Reconstruct the virtual overlay force field and update adjacent boundaries.

7. The method according to claim 1, characterized in that, The ocean current data was obtained through the following methods: Load ocean current maps from a pre-stored marine environment database; Local ocean current data is sampled in real time by the ADCP device on the UUV and updated after being filtered by a sliding window. When ocean current data is missing, the ocean current-related items are set to zero, degenerating into a mode without ocean current compensation.

8. The method according to claim 1, characterized in that, The distributed negotiation process employs a layered negotiation mechanism: UUVs are clustered into several subgroups based on geographical location, with each subgroup having a size of 3 to 5. Each seed group elects the UUV with the strongest election ability as its representative; Virtual coverage force field equalization is performed within the subgroup, and the multi-dimensional security verification is performed between subgroups by representing the UUV exchange endpoint coordinates. When the number of UUVs in a subgroup exceeds the threshold or the gap continues to exceed the limit, dynamic subgroup re-partitioning is triggered.

9. The method according to claim 6, characterized in that, The actual coverage effectiveness The number of effective UUV detection events per unit time is calculated. It is executed once after each round of patrol mission or at fixed intervals. When the feedback correction amount is less than 5% for three consecutive rounds, dynamic correction is paused to reduce computational overhead.

10. A system for allocating patrol areas for underwater unmanned vehicles (UAVs) in clusters, characterized in that, include: The data acquisition unit is used to acquire ocean current data and status information of each UUV in the target water area; The force field construction unit is used to calculate the equivalent patrol length based on the ocean current data and the state information of each UUV, and to construct a virtual coverage force field associated with the equivalent patrol length and the ocean current data. The boundary negotiation unit is used to determine the patrol boundary between adjacent UUVs based on the virtual coverage force field, and to perform multi-dimensional security verification on the boundary adjustment command during the distributed negotiation process. The dynamic optimization unit is used to dynamically adjust the equivalent patrol length and patrol boundary based on the actual coverage performance feedback of each UUV.

Images