Dynamic Event Triggering and Adaptive Collision Avoidance Methods for Unmanned Surface Vessel Swarms under Communication Constraints

Abstract

This invention discloses a dynamic event triggering and adaptive collision avoidance method for unmanned surface vessel (USV) swarms under communication constraints, belonging to the field of USV cooperative control technology. The method includes: each USV monitoring the communication link status in real time and counting the number of packet loss steps of neighboring USVs; constructing an adaptive control obstacle function based on the number of packet loss steps, dynamically updating the adaptive safe collision avoidance distance; calculating the geometric deviation between the actual position and the historical predicted position of the USV, constructing a multi-criteria dynamic event triggering criterion based on dynamically evolving internal variables; executing differentiated control branches according to the triggering criterion, broadcasting the USV's status and resolving the optimal control sequence when triggered, and using the historical control sequence when not triggered; and iteratively achieving swarm formation maintenance and adaptive safe collision avoidance. This invention can still ensure zero-collision safety of the swarm under harsh operating conditions with a communication packet loss rate as high as 30%, while significantly reducing the consumption of communication and computing resources.

Description

Technical Field

[0001] This invention relates to the field of unmanned surface vessel (USV) cooperative control technology, specifically to a dynamic event triggering and adaptive collision avoidance method for USV swarms under communication constraints. This method is applicable to USV swarm formation maintenance and safe collision avoidance control in complex communication environments such as high packet loss rate and limited bandwidth. Background Technology

[0002] Unmanned surface vessel (USV) swarm collaborative control technology has always been a hot topic and frontier in the research of marine intelligent unmanned equipment, and is also one of the key technologies for achieving autonomous operations on complex water surfaces. Although significant progress has been made in research on swarm formation and collision avoidance control, the complex and ever-changing marine environment presents considerable challenges in translating these technologies into real-world applications, resulting in relatively slow progress. For example, the problem of swarm collision avoidance decision-making in communication-constrained environments has not yet been effectively solved. In particular, ensuring the absolute physical safety of the swarm remains a bottleneck that urgently needs to be overcome in complex communication conditions such as high packet loss rates and limited bandwidth.

[0003] Cooperative collision avoidance in unmanned surface vessel (USV) swarms refers to the ability of USVs to autonomously identify and safely avoid static obstacles or neighboring USVs while navigating in formation on the water surface. Currently, those skilled in the art often employ time-triggered distributed model predictive control methods to solve swarm formation and collision avoidance problems, injecting a standard control obstacle function with fixed parameters as a hard safety constraint. This method can indeed achieve good cooperative and collision avoidance results under ideal communication conditions, laying a solid theoretical foundation for swarm cooperation.

[0004] However, the effectiveness of such methods heavily relies on the assumption of an ideal and real-time communication network, ignoring the inherent vulnerability of data transmission at sea. In real-world marine environments, however, communication between unmanned surface vessels (USVs) is often accompanied by severe packet loss and time delays due to factors such as wave obstruction and network congestion. Therefore, traditional control methods based on ideal communication environments have significant limitations in practical applications. Collision avoidance in USV swarms should not be limited to collision avoidance schemes based on real-time information exchange, but should also focus on addressing the positional uncertainty and information lag caused by communication limitations.

[0005] The existing technology has the following drawbacks: Collision avoidance safety fails due to information lag in environments with limited communication. Existing technologies typically assume that unmanned surface vessels (USVs) can continuously and in real-time acquire the positions of neighboring vessels and establish a fixed physical safety radius accordingly. However, in actual marine operations, packet loss rates as high as 30% can lead to frequent communication link failures. In such situations, the USV's acquisition of neighboring vessel position information becomes severely delayed, and the continued movement of neighboring vessels during communication interruptions causes cumulative deviations between their actual and cached positions. Since the fixed safety radius cannot adaptively compensate for this uncertainty in position prediction caused by time lag, collisions within the swarm can occur.

[0006] The problem of wasted system resources caused by fixed-period communication mechanisms. Existing time-triggered mechanisms mostly use constant communication frequencies and computational step sizes, requiring the system to perform mandatory data broadcasting and matrix recalculation throughout the entire voyage. However, during stable navigation phases with no collision risk or low environmental threats, this high-frequency interaction generates a large amount of redundant data streams, which not only exacerbates the congestion of limited communication bandwidth at sea, but also causes ineffective waste of computing resources and energy. Summary of the Invention

[0007] To address the aforementioned issues, this invention proposes a dynamic event triggering and adaptive collision avoidance method for unmanned surface vessel (USV) swarms under communication constraints. This method aims to achieve zero-collision safety under extreme sea conditions and maximize the saving of communication network resources through on-demand scheduling.

[0008] This invention aims to solve the following two technical problems existing in the prior art: First, this addresses the issue of collision avoidance safety failure caused by information lag in communication-constrained environments. In complex marine environments with high packet loss rates, unmanned surface vessels (USVs) cannot obtain the position and status of neighboring vessels in real time and accurately. Traditional obstacle avoidance control functions based on fixed safety radii are insufficient to cover prediction errors caused by information lag, resulting in an extremely high risk of physical collisions when the swarm performs formation obstacle avoidance tasks.

[0009] Second, regarding the problem of wasted system resources caused by existing technologies under fixed-period communication mechanisms. Traditional time-triggered control methods perform mandatory data broadcasting and high-frequency prediction recalculation throughout the entire voyage, generating a large amount of redundant interactive data during the stable navigation phase, exacerbating the load on the limited communication bandwidth at sea, and significantly consuming computing resources.

[0010] To achieve the above objectives, this invention provides a method for dynamic event triggering and adaptive collision avoidance in unmanned surface vessel (USV) swarms under communication constraints, comprising the following steps: (1) Each unmanned surface vessel monitors the status of the inter-vessel communication link in real time, caches the status data of neighboring vessels locally, and counts the number of packet loss steps of neighboring vessels; (2) Construct an adaptive control obstacle function based on the number of packet loss steps of neighboring vessels, dynamically update the adaptive safe collision avoidance distance between unmanned vessels, and establish cluster collision avoidance safety constraints in combination with static obstacle boundary constraints; (3) Calculate the geometric deviation between the actual position of the vessel and the historical predicted position in real time, and at the same time sense the distance between the vessel and the surrounding static obstacles, and construct a multi-criteria dynamic event triggering criterion based on dynamic evolution internal variables; (4) Execute differentiated control branches based on dynamic event triggering criteria. If the triggering conditions are met, broadcast the current state of the vessel, re-solve the optimal control sequence of the distributed model predictive control, and execute it. If the triggering conditions are not met, block the communication broadcast and use the historical optimal control sequence for motion update. (5) Iterate through steps (1) to (4) to achieve unmanned surface vessel swarm formation maintenance and adaptive safe collision avoidance in communication-constrained scenarios with high packet loss rate.

[0011] Furthermore, the method for constructing the adaptive control obstacle function based on the number of packet loss steps of neighboring vessels is as follows: Dynamic safety boundary compensation is performed based on communication packet loss steps; adaptive collision avoidance constraints between unmanned surface vessels (USVs) and fixed obstacle avoidance constraints relative to static obstacles are constructed; a basic physical safety distance for USVs is set. An adaptive safe distance formula is constructed that increases linearly with the number of packet loss steps, taking the radius of the unmanned surface vessel as twice its radius. Where γ is the safety expansion factor, and its value is not less than the maximum relative motion speed of the neighboring vessel; L is the cumulative number of packet loss steps of the neighboring vessel; Δt is the discrete sampling time step of the system; An adaptive safe collision avoidance distance is introduced into the obstacle control function. In the prediction time domain of the distributed model predictive control, a safety hard constraint is constructed in the form of a distance square inequality to limit the distance between the future moving position of the unmanned surface vessel and the predicted position of the neighboring vessel, as well as the nearest point of the static obstacle, to not be less than the corresponding safe distance.

[0012] Furthermore, the specific method for incorporating the adaptive safe collision avoidance distance into the obstacle control function is as follows: Within the finite prediction time domain of distributed model predictive control, collision avoidance constraints and obstacle avoidance constraints are established between boats; the predicted future position of this boat is defined as... The predicted future position of the neighboring vessel, considering packet loss step compensation, is: The closest point of the static obstacle outline is Construct the control barrier function inequality constraint relationship: in, The fixed safe distance between the unmanned surface vessel (USV) and static obstacles is defined by the above inequality constraints, ensuring that the motion state of the USV is always confined within the safe and feasible region throughout the entire prediction time domain.

[0013] Furthermore, the method for constructing a multi-criteria dynamic event triggering criterion based on dynamically evolving internal variables is as follows: First, establish the iterative update equation for the dynamic evolution internal variables: In the formula, ρ∈(0,1) is the set discount factor; μ is the update rate of the dynamic variable; σ is the system preset basic static deviation threshold; E(t) is the geometric deviation between the actual position of the vessel and the historical predicted position; Further define dynamic event triggering conditions based on multiple criteria, such that satisfying any inequality condition triggers a recalculation of communication and control. Dynamic deviation criterion: ; Obstacle avoidance warning criteria: ; Where θ is a positive coefficient used to adjust the weight of the influence of dynamic variables; This represents the geometric distance between the vessel and the nearest static obstacle; ALERT_RADIUS is the preset static obstacle avoidance warning radius.

[0014] Furthermore, the specific method for counting the number of packet loss steps of neighboring vessels in step (1) is as follows: each unmanned vessel monitors the network communication status in real time during navigation and maintains a communication buffer locally; if a data packet from a neighboring vessel is successfully received, the buffer status is updated and the packet loss counter of the neighboring vessel is cleared to zero; if packet loss occurs, the packet loss counter of the neighboring vessel is incremented by 1, and the previous historical status of the neighboring vessel is kept unchanged in the buffer.

[0015] Furthermore, in step (4), when resolving the optimal control sequence of the distributed model predictive control, the adaptive control barrier function constructed in step (2) is used as the safety hard constraint to solve the comprehensive cost function that includes the approach to the target, maintaining formation and control input penalty.

[0016] Furthermore, the distributed model predictive control is optimized using a sequential stepwise quadratic programming algorithm.

[0017] This invention also proposes an unmanned surface vessel (USV) swarm control system employing the above-described method, comprising, on each USV: a communication status monitoring module, a packet loss step counting module, an adaptive control obstacle function construction module, a dynamic event triggering criterion module, a distributed model predictive control module, and a motion control execution module. Beneficial effects The method for dynamic event triggering and adaptive collision avoidance of unmanned surface vessel (USV) swarms under communication constraints provided by this invention has the following beneficial effects: 1. Ensuring safe navigation of clusters under extreme communication conditions: By constructing an adaptive control obstacle function based on the number of packet loss steps, the adaptive safety boundary automatically expands outward as the number of packet loss steps increases, compensating for the position uncertainty caused by information lag, and mathematically ensuring zero-collision safety under extreme conditions of high packet loss rate (e.g., 30%).

[0018] 2. Significantly reduce communication and computing resource consumption: By introducing event-triggered criteria that dynamically evolve with system deviations, the communication threshold is automatically relaxed during stable navigation phases and automatically tightened during dangerous obstacle avoidance phases. This maximizes the suppression of redundant data broadcasting while ensuring control accuracy, effectively resolving the conflict between high-frequency communication needs and limited maritime bandwidth.

[0019] 3. Ensure the recursive feasibility of the control system: Deeply couple the adaptive safety boundary with the dynamic triggering logic to ensure that during the silent period when no communication is triggered, the unmanned surface vessel can safely use the historical prediction sequence for motion updates, maintaining the stability of the swarm formation topology and the continuity of navigation.

[0020] 4. Excellent simulation results: Simulation results show that under complex working conditions with a 30% packet loss rate, this method can keep the minimum distance between unmanned surface vessels above the safety threshold and significantly reduce the communication load. It is superior to the traditional time-triggered and standard control barrier function methods in terms of communication resource consumption and maintaining safe distance. Attached Figure Description

[0021] Figure 1 This is a flowchart of the algorithm of the present invention; Figure 2 This is a schematic diagram of the adaptive control obstacle function collision avoidance principle of the present invention; Figure 3 This is a schematic diagram illustrating the dynamic event triggering principle of the present invention. Figure 4 A comparison chart of communication counts between traditional time-triggered and standard CBF, and the event-triggered and adaptive CBF and dynamic event-triggered and adaptive CBF proposed in this invention; Figure 5 The image shows a comparison of the minimum distance between unmanned surface vessels at each moment in simulations with traditional time-triggered and standard CBF, event-triggered and adaptive CBF, and dynamic event-triggered and adaptive CBF proposed in this invention, under a 30% packet loss rate. Figure 6 A comparison chart of the minimum distance between the unmanned surface vessel and obstacles at each moment in simulations with traditional time-triggered and standard CBF, event-triggered and adaptive CBF, and dynamic event-triggered and adaptive CBF proposed in this invention, under a 30% packet loss rate. Figure 7 This is a diagram showing the route scenario results of the dynamic event triggering and adaptive CBF of the present invention under a 30% packet loss rate. Detailed Implementation

[0022] The following specific embodiments illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification.

[0023] Example 1

[0024] This embodiment provides a collaborative control method for unmanned surface vessel (USV) swarms that deeply integrates dynamic event triggering mechanisms with adaptive control obstacle functions. It is specifically applied in communication-constrained environments, such as... Figure 1 The diagram shown is a flowchart of the algorithm of this invention.

[0025] 1. Adaptive control barrier function like Figure 7 As shown, in the global coordinate system middle, own boat Located at point Its speed is ; neighboring boat Considered a moving obstacle, located at point Its speed is To accurately describe security in communication-constrained environments, this invention simplifies the unmanned surface vessel and its motion envelope into a circle for analysis. First, the unmanned surface vessel... The model is represented as a point mass, and then based on the radius of the submarine... Communication status with neighboring vessels Perform dynamic safety collision avoidance.

[0026] Define basic physical safety distance The region defined by this radius is called the basic physical boundary. In communication-constrained environments, this invention introduces the cumulative packet removal steps of neighboring vessels. With safety inflation factor Construct an adaptive safe distance that expands dynamically over time. : in, Let be the discrete sampling time step of the system. The dynamically generated circular region is called the "vessel". Adaptive positional obstacles.

[0027] Define this vessel with neighboring boats Predicted relative positional relationships: in, To account for the predicted future positions of neighboring vessels after packet loss compensation, the predicted positions of our vessel and neighboring vessels within the prediction time domain of the distributed model predictive control are... The condition for determining a physical collision is that the geometric distance between the two objects falls below the adaptive safety boundary. To ensure the absolute safety of the cluster, the set of control inputs in the velocity space that satisfy the above collision conditions is defined as the relative collision danger zone. This invention transforms the above safety requirements into hard constraints of a distributed optimization problem by constructing an adaptive control barrier function: That is, Figure 2 As shown, when the number of communication packet losses When the limit is increased, the adaptive safety boundary It automatically expands outward to compensate for positional uncertainties caused by information lag. It can be seen that as long as the optimal control sequence of this vessel... By consistently satisfying this constraint, it can be ensured that the end point of the unmanned surface vessel remains outside the adaptive position obstacle throughout the prediction time domain. This model mathematically ensures that even under extreme conditions with a high packet loss rate, the unmanned surface vessel can still achieve zero collisions by spontaneously reserving a safety margin.

[0028] 2. Multi-criteria dynamic event triggering criteria During the control solution phase, the actual position of the vessel is calculated. Compared with historical predicted locations deviation (t), and simultaneously use onboard sensors to obtain the geometric distance between the vessel and the nearest static obstacle. The iterative update equation for the dynamic evolution of internal variables is as follows: In the formula, ρ∈(0,1) is the set discount factor; μ is the update rate of the dynamic variable; σ is the system preset basic static deviation threshold; E(t) is the geometric deviation between the actual position of the vessel and the historical predicted position; Further define dynamic event triggering conditions based on multiple criteria, such that satisfying any inequality condition triggers a recalculation of communication and control. Dynamic deviation criterion: ; Obstacle avoidance warning criteria: ; Where θ is a positive coefficient used to adjust the weight of the influence of dynamic variables; This represents the geometric distance between the vessel and the nearest static obstacle; ALERT_RADIUS is the preset static obstacle avoidance warning radius.

[0029] like Figure 3As shown, if the triggering condition is not met, the predictive control sequence from the previous step is used without further calculation or broadcasting. If the triggering condition is met, the distributed model predictive control recalculates a new optimal predictive control sequence and broadcasts the data.

[0030] 3. The specific plan includes the following steps: Step 1: Communication Link Monitoring and Data Caching: Each unmanned surface vessel (USV) monitors its network communication status in real time during navigation and maintains a local communication buffer. If a data packet from the USV is successfully received, the buffer status is updated, and the packet loss counter for that USV is reset to zero. If network congestion or obstruction causes packet loss and the latest data is not received, the packet loss counter is reset. Increment by 1, and force the previous historical state of the vessel to remain unchanged in the buffer area.

[0031] This data caching mechanism with a counter ensures that even in extreme situations where the network is unstable, the underlying control system remains traceable, maintaining the continuity of the unmanned surface vessel's data processing flow.

[0032] Step 2: Construct Adaptive CBF Collision Avoidance Constraints: In the prediction time domain of DMPC, the adaptive safety distance obtained in Step 1 is... Substituting into the constraint equations, we require that the square of the predicted distance over multiple future steps is greater than the square of the safe distance, i.e.: This step transforms the positional uncertainty caused by communication delays into a continuously expanding physical safety perimeter. This adaptive mechanism ensures that the model always provides sufficient avoidance margin for unknown vessel dynamics during the solution process, thus achieving collision avoidance.

[0033] Step 3, Multi-criteria event trigger determination: The system calculates in real time the geometric deviation between the vessel's current actual position and its historical predicted position. Simultaneously, the onboard sensors are used to obtain the geometric distance between the vessel and the nearest static obstacle. When the location prediction deviation exceeds a preset threshold, i.e. or When this occurs, it is determined that communication has been triggered.

[0034] This step combines excessive prediction deviation within the system with the distance to obstacles as a criterion, decoupling control calculation from the communication network. This allows the cluster to skip meaningless broadcasts and recalculations when navigating in calm waters, significantly saving limited bandwidth resources at sea.

[0035] Step 4: DMPC Optimization and Control Execution Based on the trigger determination result, the system executes a differentiated control branch. When the trigger condition is met, the vessel broadcasts its latest data and, combined with the adaptive CBF constraints constructed in step 2, re-solves the comprehensive cost function, which includes approaching the target, maintaining formation, and control input penalties, to obtain the optimal predictive control sequence. Then, it issues and executes the first step of the control input. When the trigger condition is not met, the system directly blocks data broadcasting and matrix recalculation, and directly issues and executes the next step of the control input from the historical optimal control sequence generated at the previous trigger moment.

[0036] This step executes the planned historical actions directly when not triggered, which ensures the smoothness of the unmanned surface vessel's trajectory and avoids wasting computing power.

[0037] Simulation Experiments and Beneficial Effects To verify the effectiveness of this invention, a simulation experiment was conducted in a communication-constrained environment with a 30% packet loss rate. The present invention (dynamic event triggering + adaptive CBF) was compared with traditional time-triggered + standard CBF and event-triggered + standard CBF methods.

[0038] Communication resource consumption: such as Figure 4 As shown, compared with the traditional time-triggered method, the present invention significantly reduces the number of communications and realizes on-demand scheduling of communication resources.

[0039] Security: such as Figure 5 As shown, this invention ensures that the minimum distance between unmanned surface vessels (USVs) is always maintained above a safe threshold, while traditional methods result in distances falling below the safe threshold. Figure 6 As shown, the present invention also exhibits stability when handling collision avoidance with static obstacles.

[0040] Overall effect: Figure 7 The invention demonstrates smooth navigation and safe collision avoidance trajectory under a 30% packet loss rate.

[0041] In summary, the dynamic event triggering and adaptive collision avoidance method provided by this invention can ensure the absolute safety of the cluster with zero collisions through adaptive safety boundaries in communication-constrained environments, and can also greatly save communication and computing resources through the dynamic event triggering mechanism, thus having significant technical advantages.

[0042] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Those skilled in the art can make various improvements and modifications without departing from the spirit and principles of the invention, such as using a Kalman filter error ellipse instead of a linear dilation factor, or replacing the optimization solver. These improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for dynamic event triggering and adaptive collision avoidance in unmanned surface vessel (USV) swarms under communication constraints, characterized in that, Includes the following steps: (1) Each unmanned surface vessel monitors the status of the inter-vessel communication link in real time, caches the status data of neighboring vessels locally, and counts the number of packet loss steps of neighboring vessels; (2) Construct an adaptive control obstacle function based on the number of packet loss steps of neighboring vessels, dynamically update the adaptive safe collision avoidance distance between unmanned vessels, and establish cluster collision avoidance safety constraints in combination with static obstacle boundary constraints; (3) Calculate the geometric deviation between the actual position of the vessel and the historical predicted position in real time, and at the same time sense the distance between the vessel and the surrounding static obstacles, and construct a multi-criteria dynamic event triggering criterion based on dynamic evolution internal variables; (4) Execute differentiated control branches based on dynamic event triggering criteria. If the triggering conditions are met, broadcast the current state of the vessel, re-solve the optimal control sequence of the distributed model predictive control, and execute it. If the triggering conditions are not met, block the communication broadcast and use the historical optimal control sequence for motion update. (5) Iterate through steps (1) to (4) to achieve unmanned surface vessel swarm formation maintenance and adaptive safe collision avoidance in communication-constrained scenarios with high packet loss rate.

2. The method according to claim 1, characterized in that, The method for constructing the adaptive control obstacle function based on the number of packet loss steps of neighboring vessels is as follows: Dynamic safety boundary compensation is performed based on communication packet loss steps; adaptive collision avoidance constraints between unmanned surface vessels (USVs) and fixed obstacle avoidance constraints relative to static obstacles are constructed; a basic physical safety distance for USVs is set. An adaptive safe distance formula is constructed that increases linearly with the number of packet loss steps, taking the radius of the unmanned surface vessel as twice its radius. Where γ is the safety expansion factor, and its value is not less than the maximum relative motion speed of the neighboring vessel; L is the cumulative number of packet loss steps of the neighboring vessel; Δt is the discrete sampling time step of the system; An adaptive safe collision avoidance distance is introduced into the obstacle control function. In the prediction time domain of the distributed model predictive control, a safety hard constraint is constructed in the form of a distance square inequality to limit the distance between the future moving position of the unmanned surface vessel and the predicted position of the neighboring vessel, as well as the nearest point of the static obstacle, to not be less than the corresponding safe distance.

3. The method according to claim 2, characterized in that, The specific method for incorporating adaptive safe collision avoidance distance into the obstacle control function is as follows: Within the finite prediction time domain of distributed model predictive control, collision avoidance constraints and obstacle avoidance constraints are established between boats; the predicted future position of this boat is defined as... The predicted future position of the neighboring vessel, considering packet loss step compensation, is: The closest point of the static obstacle outline is Construct the control barrier function inequality constraint relationship: in, The fixed safe distance between the unmanned surface vessel (USV) and static obstacles is defined by the above inequality constraints, ensuring that the motion state of the USV is always confined within the safe and feasible region throughout the entire prediction time domain.

4. The method according to claim 1, characterized in that, The method for constructing a multi-criteria dynamic event triggering criterion based on dynamically evolving internal variables is as follows: First, establish the iterative update equation for the dynamic evolution internal variables: In the formula, ρ∈(0,1) is the set discount factor; μ is the update rate of the dynamic variable; σ is the system preset basic static deviation threshold; E(t) is the geometric deviation between the actual position of the vessel and the historical predicted position; Further define dynamic event triggering conditions based on multiple criteria, such that communication and control recalculation is triggered when any inequality condition is met: Dynamic deviation criterion: ; Obstacle avoidance warning criteria: ; Where θ is a positive coefficient used to adjust the weight of the influence of dynamic variables; This represents the geometric distance between the vessel and the nearest static obstacle; ALERT_RADIUS is the preset static obstacle avoidance warning radius.

5. The method according to claim 1, characterized in that, The specific method for counting the number of packet loss steps of neighboring vessels in step (1) is as follows: each unmanned vessel monitors the network communication status in real time during navigation and maintains a communication buffer locally; if a data packet from a neighboring vessel is successfully received, the buffer status is updated and the packet loss counter of the neighboring vessel is cleared to zero; if packet loss occurs, the packet loss counter of the neighboring vessel is incremented by 1 and the previous historical status of the neighboring vessel is kept unchanged in the buffer.

6. The method according to claim 1, characterized in that, When resolving the optimal control sequence of the distributed model predictive control in step (4), the adaptive control barrier function constructed in step (2) is used as the safety hard constraint to solve the comprehensive cost function that includes the approach to the target, maintaining formation and control input penalty.

7. The method according to claim 1, characterized in that, The distributed model predictive control uses a sequential stepwise quadratic programming algorithm for optimization.

8. An unmanned surface vessel swarm control system employing the method described in any one of claims 1-7, characterized in that, The system includes the following modules installed on each unmanned surface vessel: a communication status monitoring module, a packet loss step count module, an adaptive control obstacle function construction module, a dynamic event triggering criterion module, a distributed model prediction control module, and a motion control execution module.

Images