Distributed unmanned ship cluster collaborative awareness and control integration method and system in high sea condition environment

Abstract

The invention relates to a distributed unmanned ship cluster collaborative awareness and control integration method and system in a high sea condition environment. The method comprises the steps of equipment installation and distributed communication network construction; task target and environment information is collected and subjected to preprocessing, space-time calibration and data fusion, fused data is transmitted and shared through a distributed communication network, and a distributed Kalman filtering algorithm is used for joint data processing; on the basis of an unmanned ship dynamics model containing interference factors, a model prediction control algorithm is adopted to carry out rolling optimization control input, hierarchical path planning is carried out through an improved A * algorithm and a dynamic window method, cooperation among unmanned ships is kept through a cooperative control protocol, and control parameters are automatically adjusted according to sea conditions; and constructing a collaborative optimization objective function, adjusting collaborative perception and collaborative control strategies according to the minimum value of the objective function, and performing closed-loop feedback optimization on the collaborative perception process. Compared with the prior art, the sensing accuracy, the control stability and the collaborative operation efficiency of the unmanned ship cluster in a high sea condition environment can be effectively improved, and the ability of the unmanned ship cluster to cope with a severe sea environment is remarkably enhanced.

Description

Technical Field

[0001] The present invention relates to marine engineering and intelligent unmanned system technology, and in particular to an integrated method and system for collaborative perception and control of a distributed unmanned boat cluster in a high sea condition environment. Background Art

[0002] With the increasing frequency of marine development and maritime rescue activities, unmanned aerial vehicles (UAVs) are increasingly being used in the marine sector due to their low cost, low risk, and long-term operation. However, in high sea conditions (Sea State Level 5 and above), the ocean surface presents a harsh and dynamic environment with strong waves, high winds, and complex ocean currents, posing significant challenges to UAV operations.

[0003] In high sea conditions, the primary challenges faced by unmanned boats are severe disturbances and unstable attitudes. The impact of large waves can cause the boats to sway, pitch, and heave significantly. High winds further exacerbate these instabilities, making it difficult for them to maintain a normal navigational attitude. For example, during an actual maritime search and rescue mission, the sea state reached level 6, and several unmanned boats experienced severe roll while performing their mission. Some even temporarily lost their ability to operate due to a loss of attitude control.

[0004] The risk of path deviation is significantly exacerbated in high sea conditions. Complex ocean currents can generate unpredictable forces, altering the planned route of the unmanned vehicle. The combined effects of strong winds and waves can also cause the vehicle to deviate from its target path. This severely impacts the accuracy of track keeping and the efficiency of target tracking for unmanned vehicles performing search and rescue missions. In some past search and rescue operations in high sea conditions, path deviations have significantly prolonged rescue efforts, potentially even missing the optimal opportunity for rescue.

[0005] The existing technology has many shortcomings when dealing with high sea conditions. Traditional unmanned boat control methods are mostly based on model design under stable sea conditions. Under the complex interference of high sea conditions, the adaptability and robustness of the control algorithm are poor, and it is unable to effectively deal with the posture changes and path deviations of the unmanned boat. In addition, the existing unmanned boat cluster collaborative operation methods mostly focus on collaboration in a stable environment, and lack targeted design for special situations in high sea conditions. CN108549369A discloses a system and method for the collaborative formation of multiple unmanned boats under complex sea conditions. The system includes a collaborative control center, a wireless transmission module, a single-boat sea condition monitoring module, a single-boat environmental perception module, a single-boat GPS / IMU module, a single-boat navigation and obstacle avoidance module, and a single-boat control module. The collaborative control center sends commands to each unmanned vehicle based on the collaborative formation mission requirements and the wave conditions at each vehicle's location. Wireless transmission modules exchange information between the unmanned vehicles and the control center. The GPS / IMU module and environmental perception module provide real-time information on the vehicle's position and attitude, as well as obstacle information. The navigation and obstacle avoidance module and control module control the unmanned vehicles to safely and efficiently complete the collaborative formation mission. The sea condition monitoring module calculates the actual wave level based on information detected by the marine radar. While this method considers the environmental disturbance caused by waves, it only addresses collaborative formation and fails to consider the impact of high sea conditions on perception. Regarding perception, the perception range of a single unmanned vehicle is limited, and the accuracy and reliability of sensors can be severely affected in harsh high sea conditions. For example, radar and sonar equipment can misjudge or miss targets under interference from high waves and strong winds, making it difficult to accurately acquire target information and surrounding environmental conditions. Therefore, existing technologies cannot achieve efficient integrated collaborative perception and control, making it difficult to meet the requirements of unmanned vehicle swarms performing complex missions in high sea conditions.

[0006] Therefore, there is an urgent need for an integrated method of collaborative perception and control of distributed unmanned boat swarms that can adapt to high sea conditions, so as to improve the operational capability and reliability of unmanned boats in harsh marine environments. Summary of the Invention

[0007] The purpose of the present invention is to provide an integrated method and system for collaborative perception and control of distributed unmanned boat clusters in high sea conditions. By optimizing the collaborative mechanism and control strategy of the unmanned boat cluster, it can effectively solve the problems faced by the unmanned boats in high sea conditions, such as posture instability, path deviation and low collaborative operation efficiency, and improve the reliability and success rate of the unmanned boat cluster in performing tasks in high sea conditions.

[0008] The purpose of the present invention can be achieved by the following technical solutions:

[0009] A method for integrating collaborative perception and control of a distributed unmanned boat swarm in a high sea state environment includes the following steps:

[0010] System construction: Install multi-source heterogeneous sensors, wireless communication equipment, and sea condition sensors on each unmanned boat; formulate dedicated communication protocols and data formats, and use error correction coding technology to build a distributed communication network between unmanned boats;

[0011] Collaborative perception: Utilize multi-source heterogeneous sensors to collect and pre-process mission objectives and environmental information in real time. Use spatiotemporal calibration technology to unify the data of each UAV to the UAV's coordinate system and the same time reference, and perform data fusion. Each UAV transmits and shares the fused data through a distributed communication network, uses a distributed Kalman filter algorithm for joint data processing, and integrates blockchain technology to ensure data security and traceability.

[0012] Collaborative control: Based on the UAV dynamics model including interference factors, a model predictive control algorithm is used to rollingly optimize the control input and adjust the UAV's attitude. Before the mission begins, an improved A* algorithm is used to perform global path planning based on the mission objectives and environmental information. During navigation, a dynamic window method is used to adjust the local path based on real-time perception information and the UAV's own status. The UAVs maintain their spacing and collaborative relationship through a collaborative control protocol. The comprehensive sea condition assessment index is calculated based on the sea condition parameters collected by the sea condition sensors, and the relevant parameters of the collaborative control process are automatically adjusted.

[0013] Integrated fusion of collaborative perception and control: Determine the information interaction content of collaborative perception and collaborative control, and use error correction coding technology to encode and decode the transmitted information; construct a collaborative optimization objective function, and adjust the collaborative perception and collaborative control strategies according to the minimum value of the objective function; feed back the actual state of the controlled unmanned boat to the collaborative perception process, calculate the state error, and optimize the collaborative perception process to form a closed-loop feedback.

[0014] The multi-source heterogeneous sensors include radar, sonar, inertial navigation system, global positioning system and visual sensor, and the sea condition sensor includes wave height sensor, wind speed sensor and ocean current speed sensor.

[0015] In the collaborative perception, the spatiotemporal calibration technology is used to unify the data of each unmanned vehicle to the coordinate system of the unmanned vehicle body and the same time reference, and the data fusion is performed as follows:

[0016] The data collected by the radar, sonar and visual sensors in any unmanned vehicle are unified to the same time base through time interpolation method. And according to the conversion relationship between each sensor and the coordinate system of the unmanned vehicle body, the data under the coordinate of each sensor is converted to the coordinate system of the unmanned vehicle body to complete the time and space calibration.

[0017] The DS evidence theory is used to fuse the data of each sensor of any unmanned vehicle after time and space calibration to obtain the fused data of the unmanned vehicle.

[0018] The unmanned boat dynamics model including interference factors is specifically:

[0019] Assume that the position vector of the unmanned boat in the fixed coordinate system is η=[x,y,z] T , the attitude vector is θ=[φ,θ,ψ] T , the velocity vector is v = [u,v,w,p,q,r] T , where x, y, z are three-dimensional position coordinates, φ is the roll angle, θ is the pitch angle, ψ is the yaw angle, u, v, w are the linear velocities along the x, y, z axes of the fixed coordinate system, respectively, and p, q, r are the angular velocities around the x, y, z axes of the fixed coordinate system, respectively;

[0020] The dynamic model of the unmanned boat is expressed as: Where M is the inertia matrix, which contains the mass and moment of inertia of the unmanned boat; C(v) is the Coriolis force and centripetal force matrix; D(v) is the hydrodynamic damping matrix; g(η) is the resultant force vector of gravity and buoyancy; τ = [τ u ,τ v ,τ w ,τ p ,τ q ,τ r ] T is the control input vector, which is composed of the force and torque generated by the thruster and the servo, τ u ,τ v ,τ w They correspond to the forces generated by the propeller and steering gear along the x, y, and z axes of the fixed coordinate system, respectively, and are used to control the movement of the unmanned boat in three linear directions and change its position. p ,τ q ,τ r They correspond to the torques generated by the propeller and servo along the x, y, and z axes of the fixed coordinate system, respectively, and are used to control the roll, pitch, and yaw attitude of the unmanned boat and change its direction and angle; τ d =[τ du ,τ dv ,τ dw ,τ dp ,τ dq ,τ dr ] T is the disturbance force and torque vector, τ du ,τ dv ,τ dw They represent the linear forces generated by the interference factors of waves, wind and ocean currents in the x, y and z axes of the fixed coordinate system. These forces will interfere with the linear motion of the unmanned boat and cause it to deviate from the predetermined position trajectory. dp ,τ dq ,τ drThey represent the moments generated by the interference factors of waves, wind and ocean currents around the x, y and z axes of the fixed coordinate system, which will interfere with the posture of the unmanned boat.

[0021] stability, which results in changes in roll, pitch, and yaw angles; is the acceleration vector.

[0022] In the collaborative control, the model predictive control algorithm is used to optimize the control input in a rolling manner, and the posture of the unmanned boat is adjusted as follows:

[0023] At each sampling time k, the model predictive control algorithm is used to predict the future N according to the current state of the unmanned boat η(k), θ(k), and v(k). p The state at each moment, where N p To predict the time domain, the unmanned boat dynamics model is discretized to obtain the prediction model: Where i = 0, 1, ..., N p -1, ΔT is the sampling period, R(θ) is the attitude transformation matrix, and the angular velocity v in the body coordinate system is converted to r =[p,q,r] T Transformed to a fixed coordinate system, f is the discretized dynamic function;

[0024] Construct the objective function J that includes posture error and control input changes:

[0025]

[0026] Among them, θ ref (k+i) is the desired attitude, Q is the attitude error weight matrix, which is used to weigh the importance of different attitude angle errors; P is the control input change weight matrix, which avoids excessive fluctuations in the control input;

[0027] By solving the minimum value of the objective function J, we can get the future N c The optimal control input sequence τ at each moment * (k),τ * (k+1),…,τ * (k+N c -1), where N c To control the time domain, N c ≤N p , only the first control input τ * (k) Applied to the unmanned boat, the above process is repeated at the next sampling moment to achieve rolling optimization control of the unmanned boat's attitude.

[0028] In the collaborative control path planning, the global path planning adjusts the first evaluation function considering the impact of high sea conditions on navigation, and the local path is adjusted by generating a combination of speed and angular velocity within the dynamic window and selecting the optimal combination after evaluation by the second evaluation function. Among them,

[0029] The first evaluation function is expressed as: f(n) = g(n) + h(n), where f(n) is the first evaluation function, g(n) is the actual cost from the node to the starting point, h(n) is the estimated cost from the node to the target point, and h(n) = α·d(n) / (v rated - β·(|v sea | + |v wind |)), α and β are adjustment coefficients, d(n) is the straight-line distance from node n to the target point, v sea is the sea wave speed at the current position, v wind is the wind speed, v rated is the rated speed of the unmanned boat;

[0030] The second evaluation function is expressed as: J local = γ1·d obstacle (s(v, ω)) + γ2·d goal (s(v, ω)) + γ3·|ω|, where J local is the second evaluation function, γ1, γ2, and γ3 are weight coefficients, d obstacle (s(v, ω)) is the minimum distance between the trajectory and the obstacle, d goal (s(v, ω)) is the distance between the end point of the trajectory and the target point, (v, ω) is the combination of speed and angular velocity, v is the linear velocity, ω is the angular velocity, and s(v, ω) is the predicted movement trajectory of the unmanned boat in the future for a period of time.

[0031] The specific way for the unmanned boats to maintain the spacing and collaborative relationship through the collaborative control protocol is as follows:

[0032] Let the distance between the i-th unmanned boat and the j-th unmanned boat be The expected spacing is d0. When d ij < d0, the i-th unmanned boat adjusts its speed and direction according to the repulsive force model: Among them, a repulsion is the acceleration generated by the repulsive force, and k repulsion is the repulsive force coefficient; when d ij > d0, the i-th unmanned boat approaches the j-th unmanned boat according to the attractive force model to ensure the safe and collaborative operation of the entire cluster under high sea conditions.

[0033] The specific way to calculate the comprehensive sea condition evaluation index based on the sea condition parameters collected by the sea condition sensor is as follows:

[0034] The sea condition parameters including the wave height H are acquired in real time by installing sea condition sensors on the unmanned boat. s , wind speed V w , ocean current speed V c ,

[0035] Calculate the comprehensive evaluation index S of sea conditions: Among them, H s_max 、V w_max 、V c_max are the historical maximum values of wave height, wind speed and ocean current speed, respectively; μ1, μ2 and μ3 are weight coefficients.

[0036] The collaborative optimization objective function is: integration =λ1·J perception +λ2·J control ,

[0037] Among them, J integration To collaboratively optimize the objective function, J perception is the perception error correlation function, which is used to measure the deviation between the perception information and the actual situation; control is a control performance-related function that reflects the control effect of the control strategy on the attitude and path of the unmanned boat; λ1 and λ2 are weight coefficients used to adjust the importance of perception and control in the optimization process;

[0038] For target position perception, the perception error correlation function J perception Expressed as: Where N is the number of sampling times, r true is the target’s true position, r est The target position estimate output by the collaborative sensing process;

[0039] The control performance related function J control The goal of combining attitude control and path planning is expressed as: Among them, N p is the prediction time domain, θ ref (k+i) is the desired posture, Q is the posture error weight matrix, P is the control input change weight matrix, τ is the control input vector, M is the number of speed and angular velocity combinations considered in local path planning, γ1, γ2, γ3 are weight coefficients, d obstacle (s(v,ω)) is the minimum distance between the trajectory and the obstacle, d goal (s(v,ω)) is the distance between the end point of the trajectory and the target point, (v,ω) is the combination of velocity and angular velocity, v is the linear velocity, ω is the angular velocity, and s(v,ω) is the motion trajectory of the unmanned boat in the future period predicted based on (v,ω).

[0040] An integrated system for collaborative perception and control of distributed unmanned boat swarms in high sea conditions, including:

[0041] System construction module: Install multi-source heterogeneous sensors, wireless communication equipment and sea condition sensors on each unmanned boat; formulate dedicated communication protocols and data formats, adopt error correction coding technology, and build a distributed communication network between unmanned boats;

[0042] Collaborative Perception Module: Utilizes multi-source heterogeneous sensors to collect and pre-process mission objectives and environmental information in real time. Uses spatiotemporal calibration technology to unify the data of each UAV into the UAV's coordinate system and the same time reference, and then performs data fusion. Each UAV transmits and shares the fused data through a distributed communication network, using a distributed Kalman filter algorithm for joint data processing, and integrating blockchain technology to ensure data security and traceability.

[0043] Collaborative control module: Based on the UAV dynamics model including interference factors, the model predictive control algorithm is used to rollingly optimize the control input and adjust the UAV's attitude. Before the mission begins, the improved A* algorithm is used to perform global path planning based on the mission objectives and environmental information. During navigation, the dynamic window method is used to adjust the local path based on real-time perception information and the UAV's own status. The UAVs maintain their spacing and collaborative relationship through a collaborative control protocol. The comprehensive sea condition assessment index is calculated based on the sea condition parameters collected by the sea condition sensor, and the relevant parameters of the collaborative control process are automatically adjusted.

[0044] Collaborative perception and control integrated fusion module: Determine the information interaction content between the collaborative perception module and the collaborative control module, and use error correction coding technology to encode and decode the transmitted information; construct a collaborative optimization objective function, and adjust the strategies of the collaborative perception module and the collaborative control module according to the minimum value of the objective function; feed back the actual state of the controlled unmanned boat to the collaborative perception module, calculate the state error and optimize the collaborative perception module to form a closed-loop feedback.

[0045] Compared with the prior art, the present invention has the following beneficial effects:

[0046] At the collaborative perception level, the present invention's multi-source heterogeneous sensor fusion technology plays a key role. In harsh sea conditions, data collected by various sensors, such as radar, sonar, and vision, is susceptible to interference. However, the present invention eliminates temporal and spatial differences between sensors through spatiotemporal calibration and deeply integrates the data using algorithms such as DS evidence theory. This fusion approach no longer relies on information from a single sensor, but instead integrates the strengths of different sensors, effectively filtering out interfering data and providing the unmanned vehicle swarm with accurate and comprehensive target and environmental information. In actual search and rescue operations, targets that were previously difficult to locate due to sensor errors can now be accurately identified, providing a reliable information foundation for subsequent rescue operations. The distributed collaborative perception network further expands the boundaries of perception capabilities. Through the distributed Kalman filtering algorithm, real-time sharing and joint processing of perception data between unmanned vehicles is achieved, significantly expanding the perception range of the entire swarm. Even targets that are far away and difficult for a single unmanned vehicle to detect can be detected promptly by the swarm. Furthermore, the application of blockchain technology ensures the security and integrity of data during transmission and storage, eliminating the risk of data tampering and ensuring that the unmanned vehicle swarm always makes decisions based on authentic and reliable information.

[0047] The innovative design of the collaborative control process of the present invention significantly improves the maneuverability and safety of unmanned boats in high sea conditions. The attitude control based on model predictive control (MPC) fully considers interference factors such as waves, wind and ocean currents, and uses the MPC algorithm to rolling optimize the control input by constructing accurate dynamic and kinematic models. When encountering large waves and strong winds, the unmanned boat can quickly adjust its attitude and maintain a stable navigation state to avoid interruption of operation due to attitude loss of control. In the past, in high sea conditions, unmanned boats were often unable to perform tasks normally due to unstable attitudes, but the present invention can ensure that the unmanned boats can continue to operate stably, greatly improving the continuity of task execution. The hierarchical path planning and collaborative control strategy take into account both the global and local aspects, and plan a reasonable and safe navigation path for the unmanned boat. The improved A* algorithm fully considers the impact of high sea conditions on navigation when planning the global path and avoids dangerous areas; the dynamic window method can respond to sudden obstacles and severe sea conditions in real time when adjusting the local path. Furthermore, the UAVs maintain a reasonable distance between each other through a collaborative control protocol, effectively avoiding collisions and ensuring the safety of the entire swarm's collaborative operations in complex sea conditions. The adaptive control strategy further endows the UAVs with powerful environmental adaptability. It automatically and rapidly adjusts control parameters and strategies based on real-time sea conditions, ensuring the UAVs maintain optimal operating conditions and can confidently cope with the complex and ever-changing sea conditions.

[0048] The integrated fusion of collaborative perception and control of the present invention uses efficient information interaction channels and error correction coding technology to ensure the accurate transmission of perception information and control instructions even in unstable communication environments with high sea conditions, thus avoiding decision-making errors due to information loss or errors. In addition, the present invention comprehensively considers perception errors and control performance, dynamically adjusts perception and control strategies according to actual conditions, and achieves the best coordination between the two. The closed-loop feedback loop forms a self-improving system, which optimizes the perception model according to the actual state after control, further improves the accuracy of perception, and thus provides more reliable information for the control module. This deeply collaborative integration mechanism enables the unmanned boat cluster to quickly respond to changes in high sea conditions and demonstrate higher efficiency and reliability when performing tasks. In complex tasks such as search and rescue in high sea conditions, the unmanned boat cluster can reach the target area more quickly and accurately to carry out rescue work, significantly improving the success rate of mission execution.

[0049] To sum up, the integrated method of collaborative perception and control of distributed unmanned boat clusters in high sea conditions of the present invention effectively solves the difficulties faced by unmanned boat operations in high sea conditions through a series of innovative technologies, greatly improves the operation capability and reliability of unmanned boat clusters, and has important practical application value and broad market prospects. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 Flow chart of the method of the present invention. DETAILED DESCRIPTION

[0051] The present invention is described in detail below with reference to the accompanying drawings and specific embodiments. This embodiment is implemented based on the technical solution of the present invention, and provides a detailed implementation method and specific operation process, but the protection scope of the present invention is not limited to the following embodiments.

[0052] This embodiment provides an integrated method for collaborative perception and control of distributed unmanned boat clusters in high sea conditions. For harsh dynamic environments such as sea conditions level 5 and above, large waves, high wind speeds, and complex ocean currents, this embodiment studies key technologies such as multi-source heterogeneous sensor fusion, distributed collaborative perception network construction, attitude control based on model predictive control, hierarchical path planning and collaborative control, and integrated fusion of collaborative perception and control of unmanned boat clusters in this environment. This aims to improve the target perception accuracy, attitude control stability, path planning rationality, and collaborative operation efficiency of unmanned boat clusters in high sea conditions, and provide reliable technical support and solutions for complex tasks such as maritime search and rescue, marine environmental monitoring, and marine resource exploration. Figure 1 As shown, the method includes the following steps:

[0053] Step 1) System Construction: Each UAV is equipped with multi-source heterogeneous sensors, wireless communication equipment, and sea condition sensors. A dedicated communication protocol and data format are developed, and error-correcting coding technology is employed to build a distributed communication network between the UAVs. Furthermore, a transformation relationship between the UAV coordinate system and the sensor coordinate system is established to facilitate spatiotemporal calibration of the sensors.

[0054] Step 11) Sensor Installation and Configuration

[0055] In a high sea state environment, it is difficult for a single sensor to fully and accurately obtain target and environmental information. To this end, in this embodiment, each unmanned boat is equipped with a multi-source heterogeneous sensor including radar, sonar, inertial navigation system (INS), global positioning system, and visual sensor (GPS). Among them, the radar is installed in an open position on the top of the unmanned boat to ensure that the detection range is unobstructed, and its pitch angle and azimuth angle are adjusted so that its detection range covers the area in front of and on both sides of the unmanned boat; the sonar is installed on the bottom of the unmanned boat to ensure that it can effectively detect underwater targets and terrain information; the INS is fixed near the center of mass of the unmanned boat to reduce the impact of attitude changes on the measurement results; the GPS antenna is installed at the highest and unobstructed point of the unmanned boat to obtain accurate positioning information; the visual sensor is installed in a suitable position according to the mission requirements, such as the bow and stern, to ensure that key visual information can be captured.

[0056] Initial parameter settings are performed for each sensor, including sampling frequency and measurement range. Using specialized calibration equipment and methods, the sensors are time- and spatially calibrated. For time calibration, a high-precision clock source is used to synchronize the clocks of each sensor to ensure the time consistency of the collected data. Spatial calibration involves establishing a transformation relationship between the sensor coordinate system and the UAV's coordinate system. The position and attitude parameters of each sensor relative to the UAV's coordinate system are measured and recorded, and the rotation matrix and translation vector are calculated to complete the spatial calibration.

[0057] Step 12) Communication network construction

[0058] Wireless communication equipment installation: Install high-performance wireless communication equipment on the unmanned boat, such as a communication module based on 5G or ad hoc networking technology. Ensure the antenna of the communication equipment is properly installed to avoid signal obstruction and interference. Configure the communication equipment's frequency band, power, and other parameters to meet the requirements of long-distance, stable communication in high sea conditions.

[0059] Communication Protocol and Data Format Setting: Develop a dedicated communication protocol that specifies the rules and procedures for data transmission between UAVs, including mechanisms for sending, receiving, and confirming data. Define a unified data format to ensure that information such as sensor data and control commands can be accurately transmitted and interpreted between different UAVs. Use error-correcting coding techniques, such as low-density parity-check (LDPC), to encode transmitted data, improving data transmission reliability in unstable communication environments with high sea conditions.

[0060] Step 13) Deployment of sea condition monitoring system

[0061] Install sea condition sensors on the unmanned boat, such as wave height sensors, wind speed sensors, and ocean current velocity sensors. The wave height sensor uses non-contact radar wave measurement and is installed on the side of the unmanned boat to ensure accurate measurement of wave height. The wind speed sensor is installed on the mast atop the unmanned boat to monitor wind speed and direction in real time. The ocean current velocity sensor uses an acoustic Doppler velocimeter (ADCP) and is installed on the bottom of the unmanned boat to measure ocean current speed and direction. Connect the data transmission interface collected by the sea condition sensor to the unmanned boat's main control system to ensure real-time data transmission to the main control system for processing.

[0062] Step 2) Collaborative Perception: Utilize multi-source heterogeneous sensors to collect and pre-process mission objectives and environmental information in real time. Use spatiotemporal calibration technology to unify the data of each UAV into the UAV body coordinate system and the same time reference, and perform data fusion. Each UAV transmits and shares the fused data through a distributed communication network, uses a distributed Kalman filter algorithm for joint data processing, and combines blockchain technology to ensure data security and traceability.

[0063] Step 21) Use multi-source heterogeneous sensors to collect mission objectives and environmental information in real time and pre-process them.

[0064] Each sensor collects target and environmental information in real time at a set sampling frequency. Radar collects information such as the target's distance, direction, and speed; sonar obtains data such as the position and depth of underwater targets; INS collects the UAV's attitude, angular velocity, and acceleration; GPS provides positioning data such as the UAV's latitude, longitude, and altitude; and visual sensors collect images or video information.

[0065] After data acquisition is complete, preliminary processing is performed on the data, including noise removal and data smoothing. For radar and sonar data, filtering algorithms are used to remove clutter interference. For image data collected by visual sensors, image enhancement and deblurring are performed to improve image quality. The data is also timestamped to facilitate subsequent time alignment and data fusion.

[0066] Step 22) Use spatiotemporal calibration technology to unify the data of each UAV to the UAV body coordinate system and the same time reference, and perform data fusion, specifically including:

[0067] Step 221) The data collected by the radar, sonar and visual sensors in any unmanned boat are unified to the same time base through time interpolation method, and the data under the coordinates of each sensor are converted to the coordinate system of the unmanned boat body according to the conversion relationship between each sensor and the coordinate system of the unmanned boat body to complete the time and space calibration.

[0068] Assume that the target position data collected by the radar sensor is r radar =(x r ,y r ,z r ), sonar data collection is r sonar =(x s ,y s ,z s ), the data collected by the visual sensor is r vision =(x v ,y v ,z v ).

[0069] In terms of time calibration, it is assumed that the sampling period of each sensor is T radar 、T sonar 、T vision , through time interpolation, the data collected at different times are unified to the same time base. For time t, the radar data is linearly interpolated: Among them, t i ≤t <t i+1 , t i and t i+1 is the adjacent sampling time of radar.

[0070] In spatial calibration, the transformation relationship between each sensor and the UAV body coordinate system is established. Let the rotation matrix from the radar coordinate system to the UAV body coordinate system be R r , the translation vector is T r , then the radar data is converted to the body coordinate system as follows: It is the radar data in the coordinate system of the unmanned boat.

[0071] Similarly, similar spatiotemporal calibration processing is performed on sonar and visual sensor data.

[0072] Step 222) DS evidence theory is used to fuse the time-space calibrated data of each sensor of any unmanned vehicle to obtain the fused data of the unmanned vehicle.

[0073] First, determine the identification framework, that is, all possible target states and environmental conditions. Then, assign a basic probability distribution function to each proposition based on the characteristics and historical data of each sensor. Let Θ be the identification framework, which includes all possible conditions such as target presence and absence. The basic probability distribution functions of each sensor for different propositions are m radar 、m sonar 、m vision . Fusion is performed using Dempster's composition rule:

[0074]

[0075] Among them, A is a subset of Θ, n is the number of sensors (here n = 3), and more reliable target and environment information judgment is obtained through fusion.

[0076] Step 23) Each unmanned boat transmits and shares the fused data through a distributed communication network, and performs joint data processing using a distributed Kalman filter algorithm.

[0077] A distributed collaborative sensing network based on wireless communication is built between the UAVs. Each UAV transmits integrated sensing data to neighboring UAVs via the wireless communication network, while also receiving data from other UAVs. Encryption technology is used during data transmission to ensure data security and prevent theft or tampering.

[0078] Subsequently, the distributed Kalman filter algorithm is used to jointly process the perception data of the entire cluster. Each unmanned boat predicts and updates its state based on its own state and the data received from other unmanned boats. Through iterative calculations, the estimation of target state and environmental parameters is continuously optimized to expand the perception range of the entire cluster. Suppose there are N unmanned boats in the unmanned boat cluster, and the state vector of the i-th unmanned boat is Where (x i ,y i ,z i ) is the position, is the speed, θ i is the heading angle, is the heading angular velocity.

[0079] For the i-th unmanned boat, its state prediction equation is: in, is the predicted state, F i (k-1) is the state transfer matrix, is the estimated state at the previous moment, B i (k-1) is the control input matrix, u i (k-1) is the control input.

[0080] The forecast covariance is: in, is the prediction covariance, P i (k-1|k-1) is the covariance of the previous moment, Q i (k-1) is the process noise covariance.

[0081] When the i-th unmanned boat receives the information of the j-th unmanned boat, information fusion is performed:

[0082]

[0083] Among them, K i (k) is the Kalman gain, H i (k) is the observation matrix, R i (k) is the observation noise covariance, z i (k) and z j (k) are the observation values of the i-th and j-th unmanned boats respectively, and I is the unit matrix.

[0084] Step 24) Introduce blockchain technology to ensure data security and traceability.

[0085] To ensure data security, blockchain technology is introduced to store and manage sensory data, ensuring immutability and traceability. The sensory data is hashed to generate a data fingerprint, which is then packaged into data blocks along with information such as timestamps. After consensus is reached among all unmanned boat nodes through a consensus mechanism (such as the Practical Byzantine Fault Tolerance (PBFT) algorithm), the data is linked to the blockchain, ensuring immutability and traceability. In harsh communication environments like high seas, the distributed storage and redundancy of blockchain technology improve data transmission reliability.

[0086] Step 3) Collaborative control: Based on the UAV dynamics model including interference factors, a model predictive control algorithm is used to rollingly optimize the control input and adjust the UAV's posture. Before the mission begins, the improved A* algorithm is used to perform global path planning in combination with the mission objectives and environmental information. During navigation, the dynamic window method is used to adjust the local path based on real-time perception information and the UAV's own status. The UAVs maintain their spacing and collaborative relationship through a collaborative control protocol. The comprehensive sea condition assessment index is calculated based on the sea condition parameters collected by the sea condition sensor, and the relevant parameters of the collaborative control process are automatically adjusted.

[0087] Step 31) Establish a dynamic model of the unmanned vehicle including interference factors.

[0088] In high sea conditions, unmanned boats are susceptible to disturbances such as waves, strong winds, and complex ocean currents, making their attitude highly unstable. To precisely control the attitude of the unmanned boat, a dynamics model for the unmanned boat was first constructed that took into account various disturbance factors.

[0089] Assume that the position vector of the unmanned boat in the fixed coordinate system is η=[x,y,z] T, the attitude vector is θ=[φ,θ,ψ] T , the velocity vector is v = [u,v,w,p,q,r] T , where x, y, z are three-dimensional position coordinates, φ is the roll angle, θ is the pitch angle, ψ is the yaw angle, u, v, w are the linear velocities along the x, y, z axes of the fixed coordinate system, respectively, and p, q, r are the angular velocities around the x, y, z axes of the fixed coordinate system, respectively;

[0090] The dynamic model of the unmanned boat is expressed as: Where M is the inertia matrix, which contains the mass and moment of inertia of the unmanned boat; C(v) is the Coriolis force and centripetal force matrix; D(v) is the hydrodynamic damping matrix; g(η) is the resultant force vector of gravity and buoyancy; τ = [τ u ,τ v ,τ w ,τ p ,τ q ,τ r ] T is the control input vector, which is composed of the force and torque generated by the thruster and the servo, τ u ,τ v ,τ w They correspond to the forces generated by the propeller and steering gear along the x, y, and z axes of the fixed coordinate system, respectively, and are used to control the movement of the unmanned boat in three linear directions and change its position. p ,τ q ,τ r They correspond to the torques generated by the propeller and servo along the x, y, and z axes of the fixed coordinate system, respectively, and are used to control the roll, pitch, and yaw attitude of the unmanned boat and change its direction and angle; τ d =[τ du ,τ dv ,τ dw ,τ dp ,τ dq ,τ dr ] T is the disturbance force and torque vector, τ du ,τ dv ,τ dw They represent the linear forces generated by the interference factors of waves, wind and ocean currents in the x, y and z axes of the fixed coordinate system. These forces will interfere with the linear motion of the unmanned boat and cause it to deviate from the predetermined position trajectory. dp ,τ dq ,τ dr They represent the moments generated by the interference factors of waves, wind and ocean currents around the x, y and z axes of the fixed coordinate system. These moments will interfere with the attitude stability of the unmanned boat, causing changes in the roll, pitch and yaw angles; is the acceleration vector.

[0091] After the model is established, the parameters in the model, such as the inertia matrix and hydrodynamic damping matrix, are determined. Accurate parameter values can be obtained through theoretical calculations, experimental measurements, or simulation optimization.

[0092] Step 32) Use the model predictive control algorithm to optimize the control input and adjust the attitude of the unmanned boat.

[0093] At each sampling moment, a model predictive control algorithm is used to predict the attitude change trend of the unmanned vehicle over a period of time in the future (the prediction horizon) based on the current state of the unmanned vehicle (position, attitude, speed, etc.) and environmental information (wave height, wind speed, ocean current speed, etc.). An objective function is constructed that includes terms such as attitude error and control input changes. The minimum value of the objective function is solved using a rolling optimization algorithm to obtain the optimal control input sequence for the future period of time (the control horizon).

[0094] The first control command in the calculated optimal control input sequence is sent to the UAV's propellers and steering gear to adjust the UAV's attitude. At the next sampling moment, the above prediction, optimization, and control command execution process is repeated to achieve real-time and precise control of the UAV's attitude.

[0095] At each sampling time k, the model predictive control algorithm is used to predict the future N according to the current state of the unmanned boat η(k), θ(k), and v(k). p The state at each moment, where N p In order to predict the time domain, the unmanned boat dynamics model is discretized to obtain the prediction model: Where i = 0, 1, ..., N p -1, ΔT is the sampling period, R(θ) is the attitude transformation matrix, and the angular velocity v in the body coordinate system is converted to r =[p,q,r] T Converted to a fixed coordinate system, f is the discretized dynamic function.

[0096] Construct the objective function J that includes posture error and control input changes:

[0097]

[0098] Among them, θ ref (k+i) is the desired attitude, Q is the attitude error weight matrix, which is used to weigh the importance of different attitude angle errors; P is the control input change weight matrix, which avoids excessive fluctuations in the control input;

[0099] By solving the minimum value of the objective function J, we can get the future N c The optimal control input sequence τ at each moment * (k),τ * (k+1),…,τ* (k+N c -1), where N c To control the time domain, N c ≤N p , only the first control input τ * (k) Applied to the unmanned boat, the above process is repeated at the next sampling moment to achieve rolling optimization control of the unmanned boat's attitude.

[0100] Step 33) Path Planning

[0101] In high sea conditions, path planning for unmanned vehicles (UVs) must balance global objectives with local dynamics. A hierarchical path planning strategy is employed, consisting of two levels: global and local. Global path planning considers the impact of high sea conditions on navigation by adjusting the first evaluation function. Local path adjustment involves generating velocity and angular velocity combinations within a dynamic window and selecting the optimal combination after evaluation using the second evaluation function.

[0102] Step 331) Global path planning

[0103] Before the mission begins, a modified A* algorithm is used to plan a rough global path for each unmanned vehicle based on mission objectives (such as search and rescue target location and operational area boundaries) and known ocean environmental information (seabed topography, no-fly zones, etc.). During the algorithm's execution, the impact of high sea conditions on navigation is fully considered, and the evaluation function is adjusted to ensure that the planned path avoids areas with adverse sea conditions and dangerous zones.

[0104] Specifically, an improved A algorithm is used according to the mission objectives and known marine environmental information (such as seabed topography, no-fly zones, etc.). The traditional A algorithm uses the sum of the actual cost g(n) from the node to the starting point and the estimated cost h(n) from the node to the target point, f(n) = g(n) + h(n), as the evaluation function to select the path node. In high sea conditions, considering the impact of factors such as waves and wind speed on navigation time and energy consumption, this embodiment improves the estimated cost h(n) to obtain the first evaluation function f(n): h(n) = α·d(n) / (v rated -β·(|v sea |+|v wind |)), α and β are adjustment coefficients, d(n) is the straight-line distance from node n to the target point, v sea is the wave speed at the current location, v wind is the wind speed, v rated This improvement makes the planned path more adaptable to high sea conditions and generates a rough global path for each unmanned vehicle.

[0105] Step 332) Local path adjustment

[0106] During the navigation of each unmanned boat, according to the real-time perceived environmental information (such as sudden large wave areas, obstacles, etc.) and its own state, the global path is locally adjusted using the dynamic window method. A series of possible combinations of linear velocity and angular velocity are generated within the dynamic window, and each combination is evaluated through an evaluation function. The optimal combination of linear velocity and angular velocity is selected as the control input for the unmanned boat to achieve real-time path adjustment, avoiding obstacles and adverse sea condition areas.

[0107] Specifically, during the navigation of each unmanned boat, according to the real-time perceived environmental information (such as sudden large wave areas, other obstacles, etc.) and its own state, the local path is adjusted using the dynamic window method. Based on the kinematic and dynamic constraints of the unmanned boat, a series of possible combinations of linear velocity and angular velocity are generated at the current position, forming a dynamic window. For each combination of linear velocity and angular velocity (v, ω), where v is the linear velocity and ω is the angular velocity, (v, ω) predicts the motion trajectory s(v, ω) of the unmanned boat in the future for a period of time, and the second evaluation function is calculated: J local = γ1·d obstacle (s(v, ω)) + γ2·d goal (s(v, ω)) + γ3·|ω|, where J local is the second evaluation function, γ1, γ2, γ3 are weight coefficients, d obstacle (s(v, ω)) is the minimum distance between the trajectory and the obstacle, d goal (s(v, ω)) is the distance between the end point of the trajectory and the target point. The combination of linear velocity and angular velocity that minimizes the second evaluation function is selected as the control input for local path adjustment.

[0108] Step 333) Cooperative control

[0109] The distance between unmanned boats is monitored in real time. When the distance is too close, the speed and direction are adjusted according to the repulsive force model to avoid collisions; when the distance is too far, other unmanned boats are approached according to the attractive force model to ensure the cooperative operation of the entire cluster. Information such as position, speed, and heading is exchanged in real time between unmanned boats so as to adjust their own states in a timely manner and maintain the cooperation of the cluster.

[0110] Specifically, during the path planning process, unmanned boats maintain a reasonable spacing and cooperative relationship through a cooperative control protocol. Let the distance between the i-th unmanned boat and the j-th unmanned boat be The expected spacing is d0. When d ij < d0, the i-th unmanned boat adjusts its speed and direction according to the repulsive force model: where a repulsion is the acceleration generated by the repulsive force, k repulsion is the repulsive force coefficient; when d ijWhen d0 > d0, the i-th unmanned boat approaches the j-th unmanned boat according to the gravity model to ensure the safety and collaborative operation of the entire cluster in high sea conditions.

[0111] Introducing a gravity model into the collaborative perception module further enhances the UAV swarm's accurate perception of the surrounding environment and their relative positions. Based on Newton's law of universal gravitation, the gravity model describes the interactions between UAVs and with surrounding targets, such as buoys and other operating vessels.

[0112] Assume the mass of unmanned boat i is m i , the mass of unmanned boat j is m j , the distance vector between the two is r ij =(x ij ,y ij ,z ij ), its module length According to Newton's law of universal gravitation, the gravitational vector F generated by unmanned boat j on unmanned boat i is ij for:

[0113]

[0114] Here, G is the gravitational constant. In practical applications, the concept of "quality" can be analogized to factors such as the importance of an unmanned vehicle or its mission priority, and can be quantified by appropriately assigning weights. For example, an unmanned vehicle undertaking a critical monitoring mission could be assigned a higher "quality" value, giving it a more prominent position in swarm collaborative perception and action, and exerting a stronger "attraction" on other unmanned vehicles.

[0115] This gravity model can be used to assist in the construction of a distributed collaborative perception network for UAV clusters. When processing data jointly, the gravity is used to adjust the weight of information interaction between UAVs. When UAV i receives data from multiple UAVs, for the gravity F ij Data transmitted by larger UAVs is given a higher weight for fusion, so that the final perception result favors information from UAVs that are closely related to the UAV and have high mission importance. This helps the UAV swarm prioritize information that has a greater impact on overall mission execution in high sea conditions, improving the relevance and effectiveness of collaborative perception. Through the gravity model, the UAV swarm can more effectively integrate resources at the perception level, highlight key information, further optimize collaborative perception performance, and better cope with the complex and changing environmental challenges of high sea conditions.

[0116] Step 34) Adaptive control strategy

[0117] The high sea state environment is complex and changeable. The design of adaptive control strategy enables the unmanned boat to adjust control parameters and strategies in real time according to the sea conditions.

[0118] Step 341) Sea state parameter monitoring and assessment

[0119] Sea condition sensors collect real-time sea condition parameters such as wave height, wind speed, and current velocity, and transmit this data to the unmanned boat's main control system. The system then calculates a comprehensive sea condition assessment index based on a pre-set formula to assess the severity of the current sea conditions.

[0120] Specifically, the sea condition parameters are obtained in real time by installing a sea condition sensor on the unmanned boat. In this embodiment, the sea condition parameters include the wave height H s , wind speed V w , ocean current speed V c .

[0121] Calculate the comprehensive evaluation index S of sea conditions: Among them, H s_max 、V w_max 、V c_max are the historical maximum values of wave height, wind speed and ocean current speed, respectively; μ1, μ2 and μ3 are weight coefficients.

[0122] Step 342) Adjust the control parameters according to the comprehensive evaluation index S of the sea condition.

[0123] Based on comprehensive sea condition assessment indicators, the adaptive control algorithm in the main control system automatically adjusts relevant parameters in the collaborative control module. In MPC-based attitude control, parameters such as the prediction time domain, control time domain, and attitude error weight matrix are adjusted. In hierarchical path planning, parameters related to the improved A* algorithm and dynamic window method are adjusted, such as the evaluation function weights for global path planning and local path adjustment, enabling the unmanned vehicle to better adapt to different sea conditions.

[0124] For example, in MPC-based attitude control, when S increases, the prediction horizon N is increased. p and control time domain N c , giving the controller a longer timeframe to predict and adjust the UAV's attitude; simultaneously, increasing the attitude error weight matrix Q allows for tighter control of the UAV's attitude. In hierarchical path planning, when S is high, increasing the value of β in the improved A* algorithm allows global path planning to prioritize avoiding areas with adverse sea conditions. In the dynamic window method, increasing the value of γ1 increases the priority for obstacle avoidance, ensuring safe and efficient operation of the UAV in challenging sea conditions.

[0125] Step 4) Integrated fusion of collaborative perception and control: Determine the information interaction content of collaborative perception and collaborative control, and use error correction coding technology to encode and decode the transmitted information; construct a collaborative optimization objective function, and adjust the collaborative perception and collaborative control strategies according to the minimum value of the objective function; feed back the actual state of the controlled unmanned boat into the collaborative perception process, calculate the state error, and optimize the collaborative perception process to form a closed-loop feedback.

[0126] In order for unmanned aerial vehicle swarms to quickly and accurately respond to complex and changing external conditions and efficiently complete their missions in high-sea environments, it is necessary to establish a feedback mechanism that integrates collaborative perception and control, achieving a deep fusion of the two. This mechanism primarily consists of three key components: information exchange, collaborative optimization, and closed-loop feedback.

[0127] Step 41) Information Interaction

[0128] An efficient and accurate information exchange channel must be established between collaborative perception and collaborative control. The information collected during collaborative perception includes target location, environmental parameters, and obstacle information, while the information collected during control includes the UAV's attitude, speed, and control instructions. To achieve standardized information transmission and processing, an information exchange vector is defined.

[0129] Assume that the information vector output by collaborative perception is I perception =[r target ,E env ,O obs ] T , where r target is the target position vector, E env is the environmental parameter vector (including wave height, wind speed, ocean current speed, etc.), O obs is the obstacle information vector (such as obstacle position, size, etc.); the information vector output by the collaborative control module is I control =[η,v,τ] T , namely the position, speed and control input vector of the unmanned boat.

[0130] The coded information is transmitted from the collaborative sensing module to the collaborative control module and vice versa via a wireless communication network. The receiving end decodes the received information and uses a parity check matrix and decoding algorithm to detect and correct errors to ensure accurate information.

[0131] During the information exchange process, considering the instability of the communication link under high sea conditions, error correction coding technology, such as low-density parity check code (LDPC), is used to encode the transmitted data. Let the original data be D and the encoded data be D encoded , the encoding process can be expressed as: D encoded =G·D, where G is the generator matrix. The receiving end receives data D received Then, the check matrix H is used for verification:

[0132] If the check result S is all 0, the data is correct; otherwise, a decoding algorithm (such as a belief propagation algorithm) is used to correct errors to ensure that the information is accurately transmitted to the corresponding module.

[0133] Step 42) Collaborative Optimization

[0134] Based on the interactive information, a collaborative optimization model is constructed to achieve coordinated adjustment of perception and control strategies. In high sea conditions, interference such as waves and wind can affect both perception accuracy and control effectiveness, requiring collaborative optimization to balance the two.

[0135] Step 421) Objective Function Construction and Calculation: Based on the goals of collaborative perception and collaborative control, a collaborative optimization objective function is constructed. At each sampling moment, the values of the perception error-related function and the control performance-related function are calculated and substituted into the collaborative optimization objective function to obtain the objective function value at the current moment.

[0136] Define the collaborative optimization objective function as: J integration =λ1·J perception +λ2·J control ,

[0137] Among them, J integration To collaboratively optimize the objective function, J perception is the perception error correlation function, which is used to measure the deviation between the perception information and the actual situation; control is a control performance-related function that reflects the control effect of the control strategy on the attitude and path of the unmanned boat; λ1 and λ2 are weight coefficients used to adjust the importance of perception and control in the optimization process.

[0138] For target position perception, the perception error correlation function J perception Expressed as: Where N is the number of sampling times, r true is the target’s true position, r est The target position estimate is the output of the collaborative sensing process.

[0139] Control performance related function J control The goal of combining attitude control and path planning is expressed as: Where M is the number of speed and angular velocity combinations considered in local path planning.

[0140] Step 422) Solve the collaborative optimization objective function J through the optimization algorithm integration The minimum value of is obtained to obtain the collaboratively optimized perception strategy and control parameters. Based on the optimization results, the sensor fusion weights and data processing algorithms of the collaborative perception module, as well as the posture control parameters and path planning strategy of the collaborative control module, are adjusted to achieve collaborative optimization of the perception and control strategies.

[0141] Step 43) Closed-loop feedback

[0142] A closed-loop feedback loop is established to feed back the actual state of the unmanned boat after the collaborative control process executes the control action to the collaborative perception process, which is used to update the algorithm parameters of the collaborative perception process and optimize the perception strategy.

[0143] Assume that the actual state of the unmanned boat after executing the control command is η actual (k), v actual (k), and the desired state η of the collaborative control ref (k), v ref (k) There is an error e η (k) = η actual (k)-η ref (k), e v (k)=v actual (k)-v ref (k).

[0144] The collaborative perception process adjusts its own parameters according to the error information. For example, in the fusion of multi-source heterogeneous sensors, the weight of each sensor data fusion is dynamically adjusted to improve the accuracy of perception. Let the original fusion weight of the mth sensor be w m , the adjusted weight is w m′ , then: w m′ =w m +Δw m ·f(e η (k),e v (k)), where Δw m is the weight adjustment step size, and f(·) is the function that calculates the weight adjustment amount based on the error. This method optimizes the perception model and improves the perception accuracy, thereby providing more reliable information for the control module and forming a closed-loop operation mode that integrates collaborative perception and control.

[0145] The above is an introduction to the method embodiment. The following further illustrates the solution of the present invention through a system embodiment.

[0146] An integrated system for collaborative perception and control of distributed unmanned boat swarms in high sea conditions, including:

[0147] System construction module: Install multi-source heterogeneous sensors, wireless communication equipment and sea condition sensors on each unmanned boat; formulate dedicated communication protocols and data formats, adopt error correction coding technology, and build a distributed communication network between unmanned boats;

[0148] Collaborative Perception Module: Utilizes multi-source heterogeneous sensors to collect and pre-process mission objectives and environmental information in real time. Uses spatiotemporal calibration technology to unify the data of each UAV into the UAV's coordinate system and the same time reference, and then performs data fusion. Each UAV transmits and shares the fused data through a distributed communication network, using a distributed Kalman filter algorithm for joint data processing, and integrating blockchain technology to ensure data security and traceability.

[0149] Collaborative control module: Based on the UAV dynamics model including interference factors, the model predictive control algorithm is used to rollingly optimize the control input and adjust the UAV's attitude. Before the mission begins, the improved A* algorithm is used to perform global path planning based on the mission objectives and environmental information. During navigation, the dynamic window method is used to adjust the local path based on real-time perception information and the UAV's own status. The UAVs maintain their spacing and collaborative relationship through a collaborative control protocol. The comprehensive sea condition assessment index is calculated based on the sea condition parameters collected by the sea condition sensor, and the relevant parameters of the collaborative control process are automatically adjusted.

[0150] Collaborative perception and control integrated fusion module: Determine the information interaction content between the collaborative perception module and the collaborative control module, and use error correction coding technology to encode and decode the transmitted information; construct a collaborative optimization objective function, and adjust the strategies of the collaborative perception module and the collaborative control module according to the minimum value of the objective function; feed back the actual state of the controlled unmanned boat to the collaborative perception module, calculate the state error and optimize the collaborative perception module to form a closed-loop feedback.

[0151] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the described module can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here.

[0152] The integrated method and system for collaborative perception and control of distributed unmanned boat swarms in high sea conditions, proposed in this paper, addresses the numerous challenges posed to unmanned boat operations in high sea conditions through systematic technological innovation. This system has demonstrated significant and tangible benefits in many practical applications. By innovatively integrating multi-source heterogeneous sensor data processing, distributed collaborative perception networks, and hierarchical path planning, the system effectively improves the perception accuracy, control stability, and collaborative operation efficiency of unmanned boat swarms in high sea conditions. This significantly enhances the swarm's ability to cope with harsh marine environments, providing reliable technical support for unmanned boats to perform complex tasks in high sea conditions. This system has significant practical application value and broad market prospects.

[0153] The above describes in detail the preferred embodiments of the present invention. It should be understood that those skilled in the art can make numerous modifications and variations based on the concepts of the present invention without inventive effort. Therefore, any technical solutions that can be derived by those skilled in the art through logical analysis, reasoning, or limited experimentation based on the concepts of the present invention and the prior art should be within the scope of protection defined by the claims.

Claims

1. A method for integrated collaborative perception and control of distributed unmanned boat swarms in high sea conditions, characterized by: The following steps are involved: System construction: Install multi-source heterogeneous sensors, wireless communication equipment, and sea condition sensors on each unmanned boat; formulate dedicated communication protocols and data formats, and use error correction coding technology to build a distributed communication network between unmanned boats; Collaborative perception: Utilize multi-source heterogeneous sensors to collect and pre-process mission objectives and environmental information in real time. Use spatiotemporal calibration technology to unify the data of each UAV to the UAV's coordinate system and the same time reference, and perform data fusion. Each UAV transmits and shares the fused data through a distributed communication network, uses a distributed Kalman filter algorithm for joint data processing, and integrates blockchain technology to ensure data security and traceability. Collaborative control: Based on the UAV dynamics model including interference factors, a model predictive control algorithm is used to rollingly optimize the control input and adjust the UAV's attitude. Before the mission begins, an improved A* algorithm is used to perform global path planning based on the mission objectives and environmental information. During navigation, a dynamic window method is used to adjust the local path based on real-time perception information and the UAV's own status. The UAVs maintain their spacing and collaborative relationship through a collaborative control protocol. The comprehensive sea condition assessment index is calculated based on the sea condition parameters collected by the sea condition sensors, and the relevant parameters of the collaborative control process are automatically adjusted. Integrated fusion of collaborative sensing and control: Determine the information exchange content of collaborative sensing and collaborative control, and use error correction coding technology to encode and decode the transmitted information; Construct a collaborative optimization objective function, and adjust the collaborative perception and collaborative control strategies according to the minimum value of the objective function; feed back the actual state of the controlled unmanned boat into the collaborative perception process, calculate the state error and optimize the collaborative perception process to form a closed-loop feedback.

2. The method for integrated collaborative perception and control of distributed unmanned boat swarms in high sea conditions according to claim 1 is characterized in that: The multi-source heterogeneous sensors include radar, sonar, inertial navigation system, global positioning system and visual sensor, and the sea condition sensor includes wave height sensor, wind speed sensor and ocean current speed sensor.

3. The method for integrated collaborative perception and control of distributed unmanned boat swarms in high sea conditions according to claim 1 is characterized in that: In the collaborative perception, the spatiotemporal calibration technology is used to unify the data of each unmanned vehicle to the coordinate system of the unmanned vehicle body and the same time reference, and the data fusion is performed as follows: The data collected by the radar, sonar and visual sensors in any unmanned vehicle are unified to the same time base through time interpolation method. And according to the conversion relationship between each sensor and the coordinate system of the unmanned vehicle body, the data under the coordinate of each sensor is converted to the coordinate system of the unmanned vehicle body to complete the time and space calibration. The DS evidence theory is used to fuse the data of each sensor of any unmanned vehicle after time and space calibration to obtain the fused data of the unmanned vehicle.

4. The method for integrated collaborative perception and control of distributed unmanned boat swarms in high sea conditions according to claim 1, characterized in that: The unmanned boat dynamics model including interference factors is specifically: Assume that the position vector of the unmanned boat in the fixed coordinate system is η=[x,y,z] T , the attitude vector is θ=[φ,θ,ψ] T , the velocity vector is v = [u,v,w,p,q,r] T , where x, y, z are three-dimensional position coordinates, φ is the roll angle, θ is the pitch angle, ψ is the yaw angle, u, v, w are the linear velocities along the x, y, z axes of the fixed coordinate system, respectively, and p, q, r are the angular velocities around the x, y, z axes of the fixed coordinate system, respectively; The dynamic model of the unmanned boat is expressed as: Where M is the inertia matrix, which contains the mass and moment of inertia of the unmanned boat; C(v) is the Coriolis force and centripetal force matrix; D(v) is the hydrodynamic damping matrix; g(η) is the resultant force vector of gravity and buoyancy; τ = [τ u ,τ v ,τ w ,τ p ,τ q ,τ r ] T is the control input vector, which is composed of the force and torque generated by the thruster and the servo, τ u ,τ v ,τ w They correspond to the forces generated by the propeller and steering gear along the x, y, and z axes of the fixed coordinate system, respectively, and are used to control the movement of the unmanned boat in three linear directions and change its position. p ,τ q ,τ r They correspond to the torques generated by the propeller and servo along the x, y, and z axes of the fixed coordinate system, respectively, and are used to control the roll, pitch, and yaw attitude of the unmanned boat and change its direction and angle; τ d =[τ du ,τ dv ,τ dw ,τ dp ,τ dq ,τ dr ] T is the disturbance force and torque vector, τ du ,τ dv ,τ dw They represent the linear forces generated by the interference factors of waves, wind and ocean currents in the x, y and z axis directions of the fixed coordinate system, τ dp ,τ dq ,τ dr They represent the moments generated by the interference factors of waves, wind and ocean currents around the x, y and z axes of the fixed coordinate system respectively; is the acceleration vector.

5. The method for integrated collaborative perception and control of distributed unmanned boat swarms in high sea conditions according to claim 4 is characterized in that: In the collaborative control, the model predictive control algorithm is used to optimize the control input in a rolling manner, and the posture of the unmanned boat is adjusted as follows: At each sampling time k, the model predictive control algorithm is used to predict the future N according to the current state of the unmanned boat η(k), θ(k), and v(k). p The state at each moment, where N p To predict the time domain, the unmanned boat dynamics model is discretized to obtain the prediction model: Where i = 0, 1, ..., N p -1, ΔT is the sampling period, R(θ) is the attitude transformation matrix, and the angular velocity v in the body coordinate system is converted to r =[p,q,r] T Transformed to a fixed coordinate system, f is the discretized dynamic function; Construct the objective function J that includes posture error and control input changes: Among them, θ ref (k+i) is the desired attitude, Q is the attitude error weight matrix, which is used to weigh the importance of different attitude angle errors; P is the control input change weight matrix, which avoids excessive fluctuations in the control input; By solving the minimum value of the objective function J, we can get the future N c The optimal control input sequence τ at each moment * (k),τ * (k+1),…,τ * (k+N c -1), where N c To control the time domain, N c ≤N p , only the first control input τ * (k) Applied to the unmanned boat, the above process is repeated at the next sampling moment to achieve rolling optimization control of the unmanned boat's attitude.

6. The method for integrated collaborative perception and control of distributed unmanned boat swarms in high sea conditions according to claim 1 is characterized in that: In the path planning of cooperative control, the global path planning considers the impact of high sea conditions on navigation and adjusts the first evaluation function. The local path adjustment generates speed and angular velocity combinations within the dynamic window and selects the optimal combination after evaluation by the second evaluation function. The first evaluation function is expressed as: f(n) = g(n) + h(n), where f(n) is the first evaluation function, g(n) is the actual cost from the node to the starting point, h(n) is the estimated cost from the node to the target point, and h(n) = α·d(n) / (v rated -β·(|v sea |+|v wind |)), α and β are adjustment coefficients, d(n) is the straight-line distance from node n to the target point, v sea is the wave speed at the current location, v wind is the wind speed, v rated is the rated speed of the unmanned boat; The second evaluation function is expressed as: local =γ1·d obstacle (s(v,ω))+γ2·d goal (s(v,ω))+γ3·|ω|, where, J local is the second evaluation function, γ1, γ2, γ3 are weight coefficients, d obstacle (s(v,ω)) is the minimum distance between the trajectory and the obstacle, d goal (s(v,ω)) is the distance between the end point of the trajectory and the target point, (v,ω) is the combination of velocity and angular velocity, v is the linear velocity, ω is the angular velocity, and s(v,ω) is the motion trajectory of the unmanned boat in the future period predicted based on (v,ω).

7. The method for integrated collaborative perception and control of distributed unmanned boat swarms in high sea conditions according to claim 1 is characterized in that: The specific method for maintaining the distance and cooperative relationship between the unmanned boats through the cooperative control protocol is as follows: Let the distance between the $i$-th unmanned boat and the $j$-th unmanned boat be The expected spacing is $d_0$. When $d$ ij $< d_0$, the $i$-th unmanned boat adjusts its speed and direction according to the repulsive force model: where $a$ repulsion is the acceleration generated by the repulsive force, and $k$ repulsion is the repulsive force coefficient; when $d$ ij $> d_0$, the $i$-th unmanned boat approaches the $j$-th unmanned boat according to the gravitational force model to ensure the safe and collaborative operation of the entire cluster under high sea conditions.

8. The method for integrated collaborative perception and control of a distributed unmanned boat swarm in a high sea state environment according to claim 1 is characterized in that: The comprehensive sea condition evaluation index calculated based on the sea condition parameters collected by the sea condition sensor is specifically: The sea condition parameters including the wave height H are acquired in real time by installing sea condition sensors on the unmanned boat. s , wind speed V w , ocean current speed V c , Calculate the comprehensive evaluation index S of sea conditions: Among them, H s_max 、V w_max 、V c_max are the historical maximum values of wave height, wind speed and ocean current speed, respectively; μ1, μ2 and μ3 are weight coefficients.

9. The method for integrated collaborative perception and control of distributed unmanned boat swarms in high sea conditions according to claim 1, characterized in that: The collaborative optimization objective function is: integration =λ1·J perception +λ2·J control , Among them, J integration To collaboratively optimize the objective function, J perception is the perception error correlation function, which is used to measure the deviation between the perception information and the actual situation; control is a control performance-related function that reflects the control effect of the control strategy on the attitude and path of the unmanned boat; λ1 and λ2 are weight coefficients used to adjust the importance of perception and control in the optimization process; For target position perception, the perception error correlation function J perception Expressed as: Where N is the number of sampling times, r true is the target’s true position, r est The target position estimate output by the collaborative sensing process; The control performance related function J control The goal of combining attitude control and path planning is expressed as: Among them, N p is the prediction time domain, θ ref (k+i) is the desired posture, Q is the posture error weight matrix, P is the control input change weight matrix, τ is the control input vector, M is the number of speed and angular velocity combinations considered in local path planning, γ1, γ2, γ3 are weight coefficients, d obstacle (s(v,ω)) is the minimum distance between the trajectory and the obstacle, d goal (s(v,ω)) is the distance between the end point of the trajectory and the target point, (v,ω) is the combination of velocity and angular velocity, v is the linear velocity, ω is the angular velocity, and s(v,ω) is the motion trajectory of the unmanned boat in the future period predicted based on (v,ω).

10. A distributed unmanned boat swarm collaborative perception and control integrated system in a high sea state environment, characterized by: include: System construction module: Install multi-source heterogeneous sensors, wireless communication equipment and sea condition sensors on each unmanned boat; formulate dedicated communication protocols and data formats, adopt error correction coding technology, and build a distributed communication network between unmanned boats; Collaborative perception module: uses multi-source heterogeneous sensors to collect and pre-process mission objectives and environmental information in real time; Using time-space calibration technology, the data of each unmanned vehicle is unified into the coordinate system of the unmanned vehicle itself and the same time reference, and data fusion is performed. Each unmanned vehicle transmits and shares the fused data through a distributed communication network, and uses a distributed Kalman filter algorithm for joint data processing, combined with blockchain technology to ensure data security and traceability. Collaborative control module: Based on the UAV dynamics model including interference factors, the model predictive control algorithm is used to rollingly optimize the control input and adjust the UAV's attitude. Before the mission begins, the improved A* algorithm is used to perform global path planning based on the mission objectives and environmental information. During navigation, the dynamic window method is used to adjust the local path based on real-time perception information and the UAV's own status. The UAVs maintain their spacing and collaborative relationship through a collaborative control protocol. The comprehensive sea condition assessment index is calculated based on the sea condition parameters collected by the sea condition sensor, and the relevant parameters of the collaborative control process are automatically adjusted. Collaborative perception and control integrated fusion module: Determine the information interaction content between the collaborative perception module and the collaborative control module, and use error correction coding technology to encode and decode the transmitted information; construct a collaborative optimization objective function, and adjust the strategies of the collaborative perception module and the collaborative control module according to the minimum value of the objective function; feed back the actual state of the controlled unmanned boat to the collaborative perception module, calculate the state error and optimize the collaborative perception module to form a closed-loop feedback.

Images