A method and device for drone search and rescue deployment based on AIS

By linking AIS positioning and wind field modeling, and combining particle filtering algorithm and multi-source weighted fusion, the trajectory of UAVs is optimized, solving the problem of accuracy of UAV deployment in maritime search and rescue, and realizing precise deployment and intelligent decision-making under complex sea conditions.

CN121254645BActive Publication Date: 2026-01-30SHANGHAI FUKUN AVIATION TECH CO LTD
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202511832069.5
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2025-12-08
Publication Date
2026-01-30
Estimated Expiration
2045-12-08

AI Technical Summary

Technical Problem

Traditional manual search and rescue methods are slow to respond and have insufficient coverage in maritime search and rescue. Existing drone drop devices lack wind field correction mechanisms, making it impossible to achieve accurate drone drop guidance. Automatic identification-assisted search and rescue systems have not been integrated with drone trajectory analysis, making it difficult to meet the needs of rapid positioning and timely material support.

Method used

By linking AIS positioning, wind field modeling, and trajectory correction, the particle filter algorithm is used to screen and analyze AIS signals. A hierarchical wind disturbance model is constructed by combining multi-source weighted fusion, the trajectory information is optimized, the execution trajectory command set is generated, and the UAV search and rescue drop is carried out in combination with real-time control feedback.

Benefits of technology

It enables precise positioning of distressed targets and high-precision delivery of supplies in complex sea conditions, improves the intelligence and automation level of maritime rescue, and makes up for the shortcomings of the traditional method of relying on experience-based tables for correction.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN121254645B_ABST
    Figure CN121254645B_ABST
Patent Text Reader

Abstract

This invention discloses an AIS-based UAV search and rescue deployment method and apparatus, relating to the field of data processing technology. The method includes: decoding the AIS signal to obtain decoded information; performing ship information filtering and analysis on the decoded information using a particle filter algorithm to obtain target ship information; obtaining ballistic trajectory calculation information based on the target ship information through ballistic trajectory analysis and ballistic calculation; preprocessing real-time wind field data; constructing a layered wind disturbance model based on the preprocessed real-time wind field data; correcting and optimizing the ballistic trajectory information based on the ballistic calculation information and the layered wind disturbance model to obtain optimized ballistic trajectory information; and generating an execution trajectory command set based on B-spline curve smoothing using the optimized ballistic trajectory information to perform UAV search and rescue deployment in conjunction with real-time control feedback. This invention achieves ballistic trajectory correction control under wind field disturbances, thereby achieving precise UAV deployment of distressed targets under complex sea conditions.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to an AIS-based drone search and rescue delivery method and apparatus. Background Technology

[0002] In the field of maritime search and rescue, due to complex geographical environments, limited visibility, and drastic changes in sea conditions, traditional manual search and rescue methods often suffer from slow response times and insufficient coverage, making it difficult to meet the needs for rapid location and timely material support. Currently, the drone-based delivery systems used to replace manual search and rescue lack wind field correction mechanisms and heavily rely on empirical parameters, limiting their practicality. Furthermore, automatic identification-assisted search and rescue systems have failed to integrate well with drone trajectory analysis, hindering accurate drone delivery guidance. Therefore, there is an urgent need for a drone delivery method that integrates automatic identification system signal recognition, environmental factor correction, and precise trajectory control to achieve rapid target location and high-precision material delivery at sea, thereby improving the intelligence and automation level of maritime rescue. Summary of the Invention

[0003] The purpose of this invention is to overcome the shortcomings of the prior art. This invention provides an AIS-based UAV search and rescue delivery method and device, which realizes intelligent decision-making and rapid execution of delivery missions through the linkage of AIS positioning, wind field modeling and trajectory correction.

[0004] To address the aforementioned technical problems, this invention provides an AIS-based UAV search and rescue deployment method, the method comprising:

[0005] Receive the Automatic Identification System (AIS) signal and decode the AIS signal to obtain decoded information;

[0006] The decoded information is filtered and analyzed using a particle filter algorithm to obtain the target ship information.

[0007] Based on the target ship information, ballistic trajectory analysis and ballistic calculation are performed to obtain ballistic calculation information. The collected real-time wind field data is preprocessed based on a sliding window to obtain preprocessed real-time wind field data. A hierarchical wind disturbance model is constructed based on the preprocessed real-time wind field data and multi-source weighted fusion.

[0008] Based on the layered wind disturbance model, wind disturbance terms are determined, and the ballistic trajectory information is corrected and optimized based on the ballistic solution information and the wind disturbance terms combined with the disturbance influence coefficient matrix to obtain optimized ballistic trajectory information. Based on B-spline curve smoothing, the optimized ballistic trajectory information is used to generate an execution trajectory command set, and the UAV search and rescue deployment is performed based on the execution trajectory command set combined with real-time control feedback.

[0009] Optionally, the step of obtaining ballistic trajectory information based on the target ship information through ballistic trajectory analysis and ballistic calculation includes:

[0010] Based on the target ship information, the fourth-order Runge-Kutta RK4 algorithm is used to analyze the ballistic trajectory and obtain the ballistic trajectory information.

[0011] Ballistics calculation is performed based on ballistic trajectory information to obtain ballistic calculation information.

[0012] Optionally, decoding the AIS signal to obtain decoded information includes:

[0013] The structured data format of the AIS signal is obtained, and the AIS signal is decoded based on the structured data format to obtain decoded information.

[0014] Optionally, the step of performing ship information filtering and analysis on the decoded information based on the particle filter algorithm to obtain target ship information includes:

[0015] Candidate vessel information is obtained by matching the decoded information based on the maritime mobile service identification code;

[0016] The candidate vessel information is filtered by fields and the safety broadcast information is extracted to obtain the target safety broadcast information, and the distressed vessel information is determined based on the target safety broadcast information;

[0017] The drift trajectory is analyzed based on the particle filter algorithm combined with real-time ocean current data to obtain drift trajectory information, and the target vessel information is determined based on the distressed vessel information and drift trajectory information.

[0018] Optionally, the step of performing ballistic trajectory analysis using the fourth-order Runge-Kutta RK4 algorithm based on the target ship information to obtain ballistic trajectory information includes:

[0019] Based on the target ship information, initial parameters and time steps for UAV deployment are set. Using the RK4 algorithm, ballistic trajectory coordinate points are analyzed based on these parameters and time steps to obtain several ballistic trajectory coordinate points. The expression for the RK4 algorithm is as follows:

[0020] ,

[0021] ,

[0022] ,

[0023] ,

[0024] ,

[0025] in, The slope at the initial point. For data time points, These are the initial deployment parameters for the drone. For based on The midpoint slope, where h is the time step. For based on The slope at the midpoint, The key slope, To update the value;

[0026] Connect several ballistic trajectory coordinate points to obtain ballistic trajectory information.

[0027] Optionally, the step of preprocessing the collected real-time wind field data based on a sliding window to obtain preprocessed real-time wind field data, and constructing a hierarchical wind disturbance model based on the preprocessed real-time wind field data combined with multi-source weighted fusion, includes:

[0028] Anomaly removal is performed on the real-time wind field data based on a sliding window combined with median filtering to obtain preprocessed real-time wind field data.

[0029] The weighting coefficients of the preprocessed real-time wind field data are determined based on the inverse variance method, and the preprocessed real-time wind field data is then subjected to multi-source weighted fusion based on the weighting coefficients to obtain the target real-time wind field data.

[0030] Based on a preset height hierarchy, wind speed vectors are extracted from real-time target wind field data, and a three-dimensional wind field tensor is constructed based on the wind speed vectors.

[0031] A hierarchical wind disturbance model is constructed based on the three-dimensional wind field tensor.

[0032] Optionally, the step of performing ballistic calculation based on ballistic trajectory information to obtain ballistic calculation information includes:

[0033] The center-of-mass velocity of the projectile is extracted based on the ballistic trajectory information, and aerodynamic equations are constructed based on the center-of-mass velocity.

[0034] Based on the aerodynamic equations, a set of dynamic motion equations for the projectile is constructed, and the trajectory is calculated based on the set of dynamic motion equations to obtain the trajectory calculation information.

[0035] Optionally, the expression for the aerodynamic equation is:

[0036]

[0037] in, For relative velocity, The vector of the center of mass velocity. The wind speed vector at the spatiotemporal point where the projectile is located;

[0038] The expression for the set of dynamic equations is:

[0039] ,

[0040] Where x is the x-coordinate of the center of mass of the thrown object, y is the y-coordinate of the center of mass of the thrown object, and z is the z-coordinate of the center of mass of the thrown object. Let x be the x-component of the velocity of the center of mass of the thrown object in the coordinate system. Let y be the component of the velocity of the center of mass of the thrown object in the coordinate system. Let t be the z-axis component of the velocity of the center of mass of the thrown object in the coordinate system, and t be time. Let m be the air density, m be the mass of the projectile, CD be the drag coefficient, and A be the frontal area. For relative velocity, Let x be the x-axis component of the wind speed field. Let y be the y-axis component of the wind speed field. Let g be the z-axis component of the wind speed field, and g be the gravitational acceleration constant.

[0041] Optionally, the step of determining wind disturbance terms based on the layered wind disturbance model, and correcting and optimizing the ballistic trajectory information based on the ballistic solution information and the wind disturbance terms combined with the disturbance influence coefficient matrix to obtain optimized ballistic trajectory information, generating an execution trajectory command set based on B-spline curve smoothing using the optimized ballistic trajectory information, and performing UAV search and rescue drop based on the execution trajectory command set combined with real-time control feedback, includes:

[0042] The theoretical impact point information is determined based on the ballistic solution information, and the impact point offset and wind disturbance term are determined based on the layered wind disturbance model.

[0043] The error vector is calculated based on the theoretical landing point information and the landing point offset, and the velocity compensation amount is constructed by combining the error vector with the wind disturbance term and using the disturbance influence coefficient matrix.

[0044] Obtain the reverse correction component, and based on the velocity compensation amount combined with the reverse correction component, correct and optimize the ballistic trajectory information to obtain optimized ballistic trajectory information.

[0045] Based on B-spline curve smoothing and UAV attitude constraints, the optimized ballistic trajectory information is used to generate an execution trajectory instruction set.

[0046] The UAV enters a preset deployment window based on a three-dimensional coordinate sequence, detects the current environmental data and real-time attitude stability, and adjusts the execution trajectory instruction set based on the current environmental data and real-time attitude stability to obtain the adjusted execution trajectory instruction set.

[0047] Control the drone to perform search and rescue drops within a preset drop window according to the adjusted execution trajectory instruction set.

[0048] In addition, the present invention also provides a drone search and rescue drop device, the device comprising:

[0049] Signal decoding module: used to receive AIS signals from the Automatic Identification System and decode the AIS signals to obtain decoded information;

[0050] Ship information analysis module: used to perform ship information filtering and analysis on the decoded information based on the particle filter algorithm to obtain target ship information;

[0051] Ballistics calculation and model building module: Based on the target ship information, ballistic trajectory analysis and ballistic calculation processing are performed to obtain ballistic calculation information, and the collected real-time wind field data is preprocessed based on a sliding window to obtain preprocessed real-time wind field data. Based on the preprocessed real-time wind field data, a hierarchical wind disturbance model is constructed by combining multi-source weighted fusion.

[0052] The ballistic trajectory optimization module is used to determine the wind disturbance term based on the layered wind disturbance model, and to correct and optimize the ballistic trajectory information based on the ballistic solution information and the wind disturbance term combined with the disturbance influence coefficient matrix to obtain optimized ballistic trajectory information. Based on B-spline curve smoothing, the optimized ballistic trajectory information is used to generate an execution trajectory instruction set, and the execution trajectory instruction set is combined with real-time control feedback to perform UAV search and rescue drop.

[0053] In this invention, AIS signals are decoded to obtain decoded information. A particle filter algorithm is used to filter and analyze the decoded information to improve the positioning accuracy of distressed vessels and achieve precise analysis of drift information of distressed targets. Based on the target vessel information, the RK4 algorithm is used for ballistic trajectory analysis to improve the accuracy of the ballistic trajectory analysis. A hierarchical wind disturbance model is constructed based on the preprocessed real-time wind field data and multi-source weighted fusion. Ballistic trajectory calculation is performed based on the ballistic trajectory information to obtain ballistic calculation information. Based on the ballistic calculation information and the hierarchical wind disturbance model, the ballistic trajectory information is corrected and optimized to achieve ballistic trajectory correction control under wind field disturbances. This effectively compensates for the shortcomings of traditional empirical table-based correction methods in dynamic environments, thereby achieving precise UAV delivery of distressed targets under complex sea conditions. Simultaneously, through the linkage of AIS positioning, wind field modeling, and ballistic correction, intelligent decision-making and rapid execution of the delivery mission are achieved. Attached Figure Description

[0054] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0055] Figure 1 This is a flowchart illustrating the AIS-based drone search and rescue drop method in an embodiment of the present invention.

[0056] Figure 2 This is a flowchart illustrating an AIS-based drone search and rescue drop method according to another embodiment of the present invention.

[0057] Figure 3 This is a schematic diagram of the structural composition of the AIS-based drone search and rescue delivery device in an embodiment of the present invention. Detailed Implementation

[0058] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0059] Example 1

[0060] Please see Figure 1 , Figure 1 This is a flowchart illustrating the AIS-based drone search and rescue drop method in an embodiment of the present invention, the method comprising:

[0061] S11: Receive the Automatic Identification System (AIS) signal and decode the AIS signal to obtain decoded information;

[0062] In the specific implementation of this invention, an Automatic Identification System (AIS) receiver receives AIS signals from surrounding vessels. The AIS signal information includes the vessel's unique identification number, position coordinates (longitude and latitude), speed, heading, position reporting time, etc. The AIS signal is decoded according to the structured data format of the AIS signal. By decoding the AIS signal, important information such as the vessel's position coordinates and heading can be extracted, providing sufficient data support for subsequent vessel information filtering, ballistic analysis, etc. based on relevant information.

[0063] S12: Based on the particle filter algorithm, perform ship information filtering and analysis on the decoded information to obtain target ship information;

[0064] In the specific implementation of this invention, the decoded information is matched according to the maritime mobile service identification code. The decoded information of the matched vessels is used as candidate vessel information. The candidate vessel information is filtered by the vessel type field and safety broadcast information is extracted to obtain safety-related broadcast information. The distressed vessel information is located based on the broadcast information. At the same time, a particle filter algorithm is used to analyze the drift trajectory by combining real-time ocean current data and distressed vessel information. The drift trajectory and distressed vessel information are combined to form target vessel information. This effectively solves the problem of failure of traditional optical / radar under complex sea conditions, realizes accurate prediction of the drift path of distressed targets, ensures accurate positioning of distressed targets at sea, and avoids large errors in ballistic trajectory analysis.

[0065] S13: Based on the target ship information, ballistic trajectory analysis and ballistic calculation processing are performed to obtain ballistic calculation information. Based on the sliding window, the collected real-time wind field data is preprocessed to obtain preprocessed real-time wind field data. Based on the preprocessed real-time wind field data, a hierarchical wind disturbance model is constructed by combining multi-source weighted fusion.

[0066] In the specific implementation of this invention, the ballistic trajectory information is analyzed using the fourth-order Runge-Kutta (RK4) algorithm based on the target ship information. The ballistic trajectory is then calculated using the aerodynamic equations and the dynamic motion equations of the projectile, accurately determining the theoretical impact point. Weighting coefficients for the preprocessed real-time wind field data are determined based on the inverse variance method. Multi-source weighted fusion of the preprocessed real-time wind field data is then performed based on these weighting coefficients to obtain the target real-time wind field data. Wind speed vectors are extracted from the target real-time wind field data based on preset height hierarchies. A hierarchical wind disturbance model is constructed based on the three-dimensional wind field tensor built from the wind speed vectors, forming a detailed wind field profile. This achieves effective modeling of wind speed disturbances, providing more reliable wind disturbance data for subsequent analysis and correction of impact point deviations.

[0067] S14: Based on the layered wind disturbance model, determine the wind disturbance term, and based on the ballistic solution information and the wind disturbance term combined with the disturbance influence coefficient matrix, correct and optimize the ballistic trajectory information to obtain optimized ballistic trajectory information. Based on B-spline curve smoothing, use the optimized ballistic trajectory information to generate an execution trajectory instruction set, and based on the execution trajectory instruction set combined with real-time control feedback, perform UAV search and rescue drop.

[0068] In the specific implementation of this invention, the theoretical impact point is determined based on the ballistic calculation information, and the impact point offset and wind disturbance term are determined based on the layered wind disturbance model. The error vector is determined using the theoretical impact point and impact offset. Based on the error vector and the wind disturbance term, the velocity compensation amount is determined using the disturbance influence coefficient matrix to correct and optimize the initial velocity and elevation angle of the ballistic trajectory information. An execution trajectory command set is generated based on B-spline curve smoothing combined with UAV attitude constraints and optimized ballistic trajectory information. Based on the execution trajectory command set and real-time control feedback, UAV search and rescue deployment is performed, achieving ballistic trajectory correction control under wind field disturbances. This effectively compensates for the shortcomings of traditional empirical table-based correction methods in dynamic environments, thereby achieving precise UAV deployment of distressed targets under complex sea conditions.

[0069] In this embodiment of the invention, AIS signals are decoded to obtain decoded information. A particle filter algorithm is used to filter and analyze the ship information within the decoded information, improving the positioning accuracy of distressed vessels and enabling precise analysis of drift information of distressed targets. Based on the target ship information, the RK4 algorithm is used for ballistic trajectory analysis, improving the accuracy of the ballistic trajectory analysis. A hierarchical wind disturbance model is constructed based on the preprocessed real-time wind field data and multi-source weighted fusion. Ballistic trajectory calculation is performed based on the ballistic trajectory information to obtain ballistic calculation information. Based on the ballistic calculation information and the hierarchical wind disturbance model, the ballistic trajectory information is corrected and optimized, achieving ballistic trajectory correction control under wind field disturbances. This effectively compensates for the shortcomings of traditional empirical table-based correction methods in dynamic environments, thereby achieving precise UAV delivery of distressed targets under complex sea conditions. Simultaneously, through the linkage of AIS positioning, wind field modeling, and ballistic correction, intelligent decision-making and rapid execution of the delivery mission are achieved.

[0070] Example 2

[0071] Please see Figure 2 , Figure 2 This is a flowchart illustrating an AIS-based drone search and rescue drop method according to another embodiment of the present invention, the method comprising:

[0072] S201: Receive the Automatic Identification System (AIS) signal and decode the AIS signal to obtain decoded information;

[0073] In a specific implementation of the present invention, decoding the AIS signal to obtain decoding information includes: acquiring the structured data format of the AIS signal, and performing decoding processing on the AIS signal based on the structured data format to obtain decoding information.

[0074] Specifically, the Automatic Identification System (AIS) receiver receives AIS signals from surrounding vessels. The AIS signal information includes the vessel's unique identification number, position coordinates (longitude and latitude), speed, heading, and position reporting time. The system obtains structured data format information from the AIS signal. AIS signals are transmitted using a structured standard format. For example, an AIS signal might be !AIVDM,1,1,,B,15Muq?0P00PD;88MD5MTd?v@E:@h,0*5C, where !AIVDM represents AIS VHF Data. The message (received) contains 1,1,,B, indicating a total of 1 frame, the first frame, empty padding, and channel B; 15Muq?0P00PD;88MD5MTd?v@E:@h represents a six-bit ASCII encoded payload; 0 represents padding; and *5C represents the checksum. This part is encoded in **6-bit ASCII, where ** represents binary data, and each character represents 6 bits. The AIS signal is decoded based on the structured data format, and information such as ship position and speed is extracted from the AIS signal according to this structured data format, thus obtaining the decoded information.

[0075] S202: Based on the particle filter algorithm, perform ship information filtering and analysis on the decoded information to obtain target ship information;

[0076] In the specific implementation of this invention, the step of using a particle filter algorithm to filter and analyze the decoded information to obtain target vessel information includes: matching the decoded information based on the maritime mobile service identification code to obtain candidate vessel information; performing field filtering and safety broadcast information extraction on the candidate vessel information to obtain target safety broadcast information, and determining distressed vessel information based on the target safety broadcast information; and performing drift trajectory analysis based on the particle filter algorithm combined with real-time ocean current data to obtain drift trajectory information, and determining target vessel information based on the distressed vessel information and drift trajectory information.

[0077] Specifically, the decoded information is matched based on the Maritime Mobile Service Identity (MMSI). In search and rescue scenarios, the MMSI of the target device is known data, such as lifeboats and AIS beacons worn by personnel. In the obtained decoded information, the received MMSI is matched with the target database. The vessel corresponding to the matched MMSI is marked as a candidate target. The target database contains a whitelist and a blacklist. The blacklist contains irrelevant targets such as confirmed safe merchant ships, warships, and fishing boats. The whitelist contains the MMSIs of known distressed vessels, life rafts, and personal AIS signals. Obtaining the information of the candidate targets means obtaining the candidate vessel information. The candidate vessel information is filtered by field and safety broadcast information is extracted. The vessel type field in the candidate vessel information is filtered, such as 30-39 for fishing boats and small boats, and 98-99 for lifeboats. Priority is given to filtering categories belonging to life-saving equipment or small targets. After filtering out the types belonging to life-saving equipment and small targets according to the field, the safety-related broadcast information in the filtered candidate vessel information is determined. Safety-related broadcast information such as EMCY distress warning is obtained. The target safety broadcast information is obtained, and the distressed vessel information is determined based on the target safety broadcast information. The object with the distress signal is located according to the broadcast information, and the information of the distressed object is obtained, which is the distressed vessel information. Real-time ocean current data is collected, and drift trajectory analysis is performed based on the particle filtering algorithm combined with the real-time ocean current data. The current position of the distressed vessel is determined according to the distressed vessel information. A set of particles is initialized based on the real-time ocean current data and the current position of the distressed vessel. The position of each particle is updated according to the environmental drift, and the particle weight is updated according to the observation data. Particles with low weights are eliminated. The process of updating the position of particles, updating the particle weights, and eliminating particles is repeated for a preset number of iterations to obtain the final drift trajectory, i.e., the drift trajectory information. The particle filtering algorithm has strong robustness and can achieve centimeter-level prediction of the drift path of the distressed target. Based on the distressed vessel information and drift trajectory information, the target vessel information is determined, i.e., the vessel's position information, such as latitude and longitude, speed, and heading, is determined according to the distressed vessel information and drift trajectory information.

[0078] S203: Based on the target ship information, the fourth-order Runge-Kutta RK4 algorithm is used to analyze the ballistic trajectory and obtain the ballistic trajectory information;

[0079] In the specific implementation of this invention, the step of using the fourth-order Runge-Kutta RK4 algorithm to perform ballistic trajectory analysis based on the target ship information to obtain ballistic trajectory information includes: setting initial parameters and time steps for UAV deployment based on the target ship information; and using the RK4 algorithm to analyze ballistic trajectory coordinate points based on the initial parameters and time steps to obtain several ballistic trajectory coordinate points. The expression for the RK4 algorithm is as follows:

[0080] ,

[0081] ,

[0082] ,

[0083] ,

[0084]

[0085] in, The slope at the initial point. For data time points, These are the initial deployment parameters for the drone. For based on The midpoint slope, where h is the time step. For based on The slope at the midpoint, The key slope, To update the value, connect several ballistic trajectory coordinate points to obtain ballistic trajectory information.

[0086] Specifically, the initial parameters and time step for UAV deployment are set based on the target ship information. This involves setting the UAV's deployment altitude, speed, and angle according to the target ship's position information. A relatively small time step, such as 0.01s, can be chosen. Using the fourth-order Runge-Kutta (RK4) algorithm, the initial parameters and time step are used to analyze the ballistic trajectory coordinates. The gradient at different time points is estimated using the RK4 algorithm, and the velocity and position increments at each stage are calculated. The results from all stages are combined to calculate the final new state, i.e., the corresponding trajectory coordinates. The updated value is used as the starting point for the next time step. This process is repeated until the calculated z-axis coordinates are less than or equal to zero or cover a preset time range, resulting in several ballistic trajectory coordinates. The expression for the RK4 algorithm is:

[0087] ,

[0088] ,

[0089] ,

[0090] ,

[0091]

[0092] in, The slope at the initial point. For data time points, These are the initial deployment parameters for the drone. For based on The midpoint slope, where h is the time step. For based on The slope at the midpoint, The key slope, To update the values, connect several ballistic trajectory coordinate points to obtain ballistic trajectory information.

[0093] S204: Preprocess the collected real-time wind field data based on a sliding window to obtain preprocessed real-time wind field data, and construct a hierarchical wind disturbance model based on the preprocessed real-time wind field data combined with multi-source weighted fusion.

[0094] In the specific implementation of this invention, the step of preprocessing the collected real-time wind field data based on a sliding window to obtain preprocessed real-time wind field data, and constructing a hierarchical wind disturbance model based on the preprocessed real-time wind field data combined with multi-source weighted fusion, includes: performing anomaly removal processing on the real-time wind field data based on a sliding window combined with a median filtering method to obtain preprocessed real-time wind field data; determining the weight coefficients of the preprocessed real-time wind field data based on the inverse variance method, and performing multi-source weighted fusion on the preprocessed real-time wind field data based on the weight coefficients to obtain target real-time wind field data; extracting wind speed vectors from the target real-time wind field data based on preset height hierarchies, and constructing a three-dimensional wind field tensor based on the wind speed vectors; and constructing a hierarchical wind disturbance model based on the three-dimensional wind field tensor.

[0095] Specifically, real-time wind field data is collected using ultrasonic anemometers and laser wind radar. Ultrasonic anemometers, installed on drones or other platforms, can sense wind speed and direction in the flight environment in real time. Laser wind radar can provide a more accurate wind field profile over a larger area, typically used to acquire wind field information within a 0–50 meter range. It measures the movement of particles in the air by reflecting a laser beam, thereby calculating wind speed and direction. Anomaly removal is performed on the real-time wind field data using a sliding window combined with median filtering, identifying abrupt changes and drift values, such as abnormal wind speed peaks and data breaks. These abrupt changes and drift values ​​are then removed using a sliding window mean and median filtering method, ensuring the continuity of the wind field data and obtaining preprocessed real-time wind field data. The weighting coefficients of the preprocessed real-time wind field data are determined using the inverse variance method. The inverse variance method is used to determine the corresponding weighting coefficients for data from different sensors within the same height layer. The measurement variance of each sensor is calculated, and the weighting coefficients are calculated based on the measurement variance of each sensor combined with the formula of the inverse variance method. Multi-source weighted fusion is then performed on the preprocessed real-time wind field data based on these weighting coefficients, i.e., data from different sensors are weighted and fused with their respective weighting coefficients to obtain the target real-time wind field data. Wind speed vectors are extracted from the target real-time wind field data based on preset height layers, such as every 10 meters. A three-dimensional wind field tensor is constructed based on these wind speed vectors, representing the wind speed vector field that varies with space and time. A layered wind disturbance model is constructed based on the three-dimensional wind field tensor. The average wind speed vector for each height layer can be calculated from the three-dimensional wind field tensor, and a layered wind disturbance model is constructed based on the average wind speed vector. The layered wind disturbance model is a multi-level model of the wind field.

[0096] S205: Extract the center-of-mass velocity of the projectile based on ballistic trajectory information, and construct aerodynamic equations based on the center-of-mass velocity;

[0097] In the specific implementation of this invention, the center-of-mass velocity of the projectile is extracted based on the ballistic trajectory information, and an aerodynamic equation is constructed based on the center-of-mass velocity. Under the condition of wind field presence, an unsteady aerodynamic equation is established. This aerodynamic equation can describe the motion of the flying object under the influence of different airflows. The expression of the aerodynamic equation is as follows:

[0098]

[0099] in, For relative velocity, The vector of the center of mass velocity. The wind speed vector is the point in spacetime where the projectile is located.

[0100] S206: Construct a set of dynamic motion equations for the projectile based on aerodynamic equations, and perform ballistic calculations based on the set of dynamic motion equations to obtain ballistic calculation information;

[0101] In the specific implementation of this invention, a set of dynamic motion equations for the projectile is constructed based on aerodynamic equations, and the expression of the set of dynamic motion equations is as follows:

[0102] ,

[0103] Where x is the x-coordinate of the center of mass of the thrown object, y is the y-coordinate of the center of mass of the thrown object, and z is the z-coordinate of the center of mass of the thrown object. Let x be the x-component of the velocity of the center of mass of the thrown object in the coordinate system. Let y be the component of the velocity of the center of mass of the thrown object in the coordinate system. Let t be the z-axis component of the velocity of the center of mass of the thrown object in the coordinate system, and t be time. Let m be the air density, m be the mass of the projectile, CD be the drag coefficient, and A be the frontal area. For relative velocity, Let x be the x-axis component of the wind speed field. Let y be the y-axis component of the wind speed field. Let g be the z-axis component of the wind speed field, g be the gravitational acceleration constant, and the trajectory is calculated based on the dynamic equations of motion. The trajectory calculation adopts the dynamic equations of motion, taking into account factors such as air resistance, gravity, and wind field interference, to accurately calculate the flight path information of the object, thus obtaining the trajectory calculation information.

[0104] S207: Based on the layered wind disturbance model, determine the wind disturbance term, and based on the ballistic solution information and the wind disturbance term combined with the disturbance influence coefficient matrix, correct and optimize the ballistic trajectory information to obtain optimized ballistic trajectory information. Based on B-spline curve smoothing, use the optimized ballistic trajectory information to generate an execution trajectory instruction set, and based on the execution trajectory instruction set combined with real-time control feedback, perform UAV search and rescue drop.

[0105] In the specific implementation of this invention, the process of determining wind disturbance terms based on the layered wind disturbance model, and correcting and optimizing the ballistic trajectory information based on the ballistic solution information and the wind disturbance terms combined with the disturbance influence coefficient matrix to obtain optimized ballistic trajectory information, generating an execution trajectory command set based on B-spline curve smoothing using the optimized ballistic trajectory information, and performing UAV search and rescue drop based on the execution trajectory command set combined with real-time control feedback, includes: determining theoretical landing point information based on ballistic solution information, and determining landing point offset and wind disturbance terms based on the layered wind disturbance model; calculating an error vector based on the theoretical landing point information and landing point offset, and using the disturbance influence coefficient matrix based on the error vector combined with the wind disturbance terms. A velocity compensation amount is constructed using a matrix; a reverse correction component is obtained, and the ballistic trajectory information is corrected and optimized based on the velocity compensation amount and the reverse correction component to obtain optimized ballistic trajectory information; an execution trajectory command set is generated using the optimized ballistic trajectory information based on B-spline curve smoothing and UAV attitude constraints; the UAV enters a preset delivery window based on a three-dimensional coordinate sequence, detects the current environmental data and real-time attitude stability, and adjusts the execution trajectory command set based on the current environmental data and real-time attitude stability to obtain a feedback-adjusted execution trajectory command set; the UAV is controlled to perform search and rescue drops within the preset delivery window according to the feedback-adjusted execution trajectory command set.

[0106] Specifically, the theoretical impact point information is determined based on ballistic calculation information. The future impact point of the projectile can be estimated using the RK4 algorithm or particle filtering based on the ballistic calculation information; this is the theoretical impact point information. The impact point offset is then determined based on a layered wind disturbance model. The average wind speed vector of the main wind direction at the flight altitude layer is determined according to the layered wind disturbance model, and the flight prediction time of the projectile is obtained. Based on the average wind speed vector and the flight prediction time, the impact point offset caused by wind disturbance is estimated. The expression for the impact point offset can be:

[0107] ,

[0108] in, Let W be the landing point offset, W be the average wind speed vector, and T be the flight prediction time. Based on the layered wind disturbance model, the wind disturbance term is determined and then incorporated into a preset dynamic optimization equation to obtain the wind disturbance term.

[0109] The error vector is calculated based on the theoretical landing point information and the landing point offset. The expression for the error vector is as follows:

[0110] ,

[0111] Where e is the error vector and rth is the theoretical landing point. Here, rtarget is the target position, representing the location of the distressed object. A velocity compensation is constructed based on the error vector and the disturbance term. The corresponding wind disturbance term is then linearly superimposed onto the state equation to obtain the velocity correction. This velocity compensation is determined using the disturbance influence coefficient matrix based on the velocity correction and the error vector. This velocity compensation is used to adjust the direction and amplitude of the initial velocity vector to counteract lateral or longitudinal disturbances caused by wind. For cases where crosswinds are dominant, a reverse correction component is added to the initial velocity in the lateral direction. The trajectory information is then corrected and optimized based on the velocity compensation and the reverse correction component. Control parameters such as the initial launch velocity and elevation angle are adjusted according to the velocity compensation and the reverse correction component to obtain optimized trajectory information. When the wind speed is high at a certain level, the object may be affected by strong winds, causing trajectory deviation. By adjusting the initial launch velocity, elevation angle, and other control parameters in real time, this deviation can be effectively reduced to cope with wind shear and ensure that the launch point is as close to the target as possible. Furthermore, by comparing the theoretical landing point with the current target position, a displacement compensation vector caused by wind disturbance can be generated in real time. The expression for the displacement compensation vector is:

[0112] ,

[0113] in, Here, rpred is the displacement compensation vector, and rpred is the predicted theoretical landing point. The system determines the target's current position and adjusts the initial velocity, direction, and elevation angle parameters of the projectile accordingly. This correction mechanism can proactively optimize the delivery attitude and improve landing accuracy under conditions of rapid wind speed changes or sudden wind shear, ensuring meter-level accuracy even in strong winds or crosswinds.

[0114] Based on B-spline curve smoothing combined with UAV attitude constraints, the optimized ballistic trajectory information is used to generate an execution trajectory instruction set. The UAV attitude constraints include pitch angle, heading angle, and upper limit of release speed. The optimized ballistic trajectory information is smoothed by B-spline curves in combination with UAV attitude constraints to form an execution trajectory instruction set. A B-spline curve is a smooth curve composed of multiple control points. It determines the curve shape by interpolating between control points.

[0115] The UAV enters a preset delivery window based on a three-dimensional coordinate sequence. The UAV performs trajectory tracking control based on the three-dimensional coordinate sequence, correcting its heading and altitude in real time to enter the preset delivery window. It detects current environmental data and real-time attitude stability, which can be continuously detected by onboard sensors. Current environmental data may include wind speed and sea state changes. Simultaneously, corresponding sensors detect UAV attitude changes to determine real-time attitude stability. Based on the current environmental data and real-time attitude stability, the execution trajectory command set is adjusted. A compensation vector can be re-determined based on the current environmental data and real-time attitude stability to further adjust the execution trajectory command set. Simultaneously, the hierarchical wind disturbance model can be dynamically updated based on the current environmental data, correcting the velocity in each iteration to obtain the adjusted execution trajectory command set.

[0116] The system controls the UAV to perform search and rescue drops within a preset drop window according to the adjusted execution trajectory instruction set, ensuring a smooth and reliable drop process with high landing accuracy. It achieves real-time closed-loop correction, supports wind field-attitude-trajectory linkage control, and improves drop accuracy and system stability under complex sea conditions.

[0117] In this embodiment of the invention, AIS signals are decoded to obtain decoded information. A particle filter algorithm is used to filter and analyze the ship information within the decoded information, improving the positioning accuracy of distressed vessels and enabling precise analysis of drift information of distressed targets. Based on the target ship information, the RK4 algorithm is used for ballistic trajectory analysis, improving the accuracy of the ballistic trajectory analysis. A hierarchical wind disturbance model is constructed based on the preprocessed real-time wind field data and multi-source weighted fusion. Ballistic trajectory calculation is performed based on the ballistic trajectory information to obtain ballistic calculation information. Based on the ballistic calculation information and the hierarchical wind disturbance model, the ballistic trajectory information is corrected and optimized, achieving ballistic trajectory correction control under wind field disturbances. This effectively compensates for the shortcomings of traditional empirical table-based correction methods in dynamic environments, thereby achieving precise UAV delivery of distressed targets under complex sea conditions. Simultaneously, through the linkage of AIS positioning, wind field modeling, and ballistic correction, intelligent decision-making and rapid execution of the delivery mission are achieved.

[0118] Example 3

[0119] Please see Figure 3 , Figure 3 This is a schematic diagram of the structural composition of an AIS-based drone search and rescue delivery device according to an embodiment of the present invention. The device includes:

[0120] Signal decoding module 31: used to receive the AIS signal from the Automatic Identification System and decode the AIS signal to obtain decoding information;

[0121] Ship information analysis module 32: used to perform ship information filtering and analysis on the decoded information based on the particle filter algorithm to obtain target ship information;

[0122] Ballistics calculation and model building module 33: Based on the target ship information, ballistic trajectory analysis and ballistic calculation processing are performed to obtain ballistic calculation information, and the collected real-time wind field data is preprocessed based on a sliding window to obtain preprocessed real-time wind field data. Based on the preprocessed real-time wind field data, a hierarchical wind disturbance model is constructed by combining multi-source weighted fusion.

[0123] The ballistic trajectory optimization module 34 is used to determine the wind disturbance term based on the layered wind disturbance model, and to correct and optimize the ballistic trajectory information based on the ballistic solution information and the wind disturbance term combined with the disturbance influence coefficient matrix to obtain optimized ballistic trajectory information. Based on B-spline curve smoothing, the optimized ballistic trajectory information is used to generate an execution trajectory instruction set, and based on the execution trajectory instruction set combined with real-time control feedback, the UAV search and rescue drop is performed.

[0124] In the specific implementation of this invention, the specific implementation of the device item can be referred to the implementation of the method item above, and will not be repeated here.

[0125] In this embodiment of the invention, AIS signals are decoded to obtain decoded information. A particle filter algorithm is used to filter and analyze the ship information within the decoded information, improving the positioning accuracy of distressed vessels and enabling precise analysis of drift information of distressed targets. Based on the target ship information, the RK4 algorithm is used for ballistic trajectory analysis, improving the accuracy of the ballistic trajectory analysis. A hierarchical wind disturbance model is constructed based on the preprocessed real-time wind field data and multi-source weighted fusion. Ballistic trajectory calculation is performed based on the ballistic trajectory information to obtain ballistic calculation information. Based on the ballistic calculation information and the hierarchical wind disturbance model, the ballistic trajectory information is corrected and optimized, achieving ballistic trajectory correction control under wind field disturbances. This effectively compensates for the shortcomings of traditional empirical table-based correction methods in dynamic environments, thereby achieving precise UAV delivery of distressed targets under complex sea conditions. Simultaneously, through the linkage of AIS positioning, wind field modeling, and ballistic correction, intelligent decision-making and rapid execution of the delivery mission are achieved.

[0126] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, which may include: read-only memory (ROM), random access memory (RAM), magnetic disk or optical disk, etc.

[0127] Furthermore, the above provides a detailed description of an AIS-based UAV search and rescue dropping method and apparatus provided by the embodiments of the present invention. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. An AIS-based unmanned aerial vehicle search and rescue dropping method, characterized in that, The method comprises: receiving an automatic identification system (AIS) signal and decoding the AIS signal to obtain decoded information; performing ship information screening analysis on the decoded information based on a particle filtering algorithm to obtain target ship information; performing ballistic trajectory analysis and ballistic calculation based on the target ship information to obtain ballistic calculation information, pre-processing collected real-time wind field data based on a sliding window to obtain pre-processed real-time wind field data, and constructing a hierarchical wind disturbance model based on the pre-processed real-time wind field data combined with multi-source weighted fusion; determining a wind disturbance term based on the hierarchical wind disturbance model, and correcting and optimizing the ballistic trajectory information based on the ballistic calculation information and the wind disturbance term combined with a disturbance influence coefficient matrix to obtain optimized ballistic trajectory information, generating an execution trajectory instruction set based on the optimized ballistic trajectory information using B-spline curve smoothing, and performing unmanned aerial vehicle search and rescue throwing based on the execution trajectory instruction set combined with real-time control feedback; wherein the pre-processing of the collected real-time wind field data based on a sliding window to obtain pre-processed real-time wind field data, and the construction of a hierarchical wind disturbance model based on the pre-processed real-time wind field data combined with multi-source weighted fusion, comprises: performing abnormal elimination processing on the real-time wind field data based on a sliding window combined with a median filtering method to obtain pre-processed real-time wind field data; determining a weight coefficient of the pre-processed real-time wind field data based on a variance reciprocal method, and performing multi-source weighted fusion on the pre-processed real-time wind field data based on the weight coefficient to obtain target real-time wind field data; extracting a wind speed vector based on a pre-set height layer using the target real-time wind field data, and constructing a three-dimensional wind field tensor based on the wind speed vector; and constructing a hierarchical wind disturbance model based on the three-dimensional wind field tensor.

2. The AIS based UAV search and rescue drop method of claim 1, wherein, The ballistic trajectory analysis based on the target ship information using a fourth-order Runge-Kutta (RK4) algorithm to obtain ballistic trajectory information comprises: performing ballistic trajectory analysis based on the target ship information using a fourth-order Runge-Kutta (RK4) algorithm to obtain ballistic trajectory information; performing ballistic calculation based on the ballistic trajectory information to obtain ballistic calculation information.

3. The AIS based UAV search and rescue drop method of claim 1, wherein, The decoding of the AIS signal to obtain decoded information comprises: obtaining a structured data format of the AIS signal, and decoding the AIS signal based on the structured data format to obtain decoded information.

4. The AIS based UAV search and rescue drop method of claim 1, wherein, The ship information screening analysis based on the particle filtering algorithm on the decoded information to obtain target ship information comprises: matching the decoded information based on a waterborne mobile service identification code to obtain candidate ship information; performing field filtering and safety broadcast information extraction on the candidate ship information to obtain target safety broadcast information, and determining distress ship information based on the target safety broadcast information; performing drift trajectory analysis based on the particle filtering algorithm combined with real-time ocean current data to obtain drift trajectory information, and determining target ship information based on the distress ship information and the drift trajectory information.

5. The AIS based UAV search and rescue drop method of claim 2, wherein, The ballistic trajectory analysis based on the target ship information using a fourth-order Runge-Kutta (RK4) algorithm to obtain ballistic trajectory information comprises: The initial parameters and time steps of the unmanned aerial vehicle launching are set based on target ship information, and the coordinate points of the trajectory are analyzed based on the RK4 algorithm using the initial parameters and time steps of the unmanned aerial vehicle launching, so as to obtain a plurality of trajectory coordinate points, and the expression of the RK4 algorithm is: , , , , , wherein, is the initial point slope, is the data time point, is the initial UAV release parameter, is the midpoint slope based on h is the time step, is the midpoint slope based on h is the time step, is the focus slope, is the update value; The trajectory information is obtained by connecting the plurality of trajectory coordinate points.

6. The AIS based UAV search and rescue drop method of claim 2, wherein, The trajectory calculation is performed based on the trajectory information, and trajectory calculation information is obtained, including: The center of mass velocity of the projectile is extracted based on the trajectory information, and the aerodynamic equation is constructed based on the center of mass velocity; The dynamic motion equation set of the projectile is constructed based on the aerodynamic equation, and the trajectory calculation is performed based on the dynamic motion equation set, so as to obtain the trajectory calculation information.

7. The AIS based UAV search and rescue drop method of claim 6, wherein, The expression of the aerodynamic equation is: , wherein is the relative velocity, is the vector of the center of mass velocity, is the wind velocity vector at the space-time point where the projectile is located; The expression of the dynamic motion equation set is: , wherein x is the x-axis coordinate of the centroid position of the projectile, y is the y-axis coordinate of the centroid position of the projectile, z is the z-axis coordinate of the centroid position of the projectile, is the x-axis component of the centroid velocity of the projectile in the coordinate system, is the y-axis component of the centroid velocity of the projectile in the coordinate system, is the z-axis component of the centroid velocity of the projectile in the coordinate system, t is time, is the air density, m is the mass of the projectile, CD is the drag coefficient, A is the wind- facing area, is the relative velocity, is the x-axis component of the wind velocity field, is the y-axis component of the wind velocity field, is the z-axis component of the wind velocity field, g is the gravitational acceleration constant.

8. The AIS based UAV search and rescue drop-off method of claim 1, wherein, The wind disturbance term is determined based on the layered wind disturbance model, and the trajectory information is modified and optimized based on the trajectory calculation information and the wind disturbance term combined with the disturbance influence coefficient matrix, so as to obtain optimized trajectory information, and the execution trajectory instruction set is generated based on the optimized trajectory information using B-spline curve smoothing, and the unmanned aerial vehicle search and rescue launching is performed based on the execution trajectory instruction set combined with real-time control feedback, including: The theoretical landing point information is determined based on the trajectory calculation information, and the landing point offset and the wind disturbance term are determined based on the layered wind disturbance model; The error vector is calculated based on the theoretical landing point information and the landing point offset, and the velocity compensation is constructed based on the error vector combined with the wind disturbance term using the disturbance influence coefficient matrix; The reverse correction component is obtained, and the trajectory information is modified and optimized based on the velocity compensation combined with the reverse correction component, so as to obtain the optimized trajectory information; The execution trajectory instruction set is generated based on B-spline curve smoothing combined with the unmanned aerial vehicle attitude constraint using the optimized trajectory information; The unmanned aerial vehicle enters the preset launching window based on the three-dimensional coordinate sequence, detects the current environment data and real-time attitude stability, and adjusts the execution trajectory instruction set based on the current environment data and real-time attitude stability, so as to obtain the feedback adjusted execution trajectory instruction set; The unmanned aerial vehicle is controlled to perform search and rescue launching according to the feedback adjusted execution trajectory instruction set in the preset launching window.

9. A drone search and rescue throwing device, comprising: The device comprises: A signal decoding module is configured to receive an automatic identification system (AIS) signal and decode the AIS signal to obtain decoded information. A ship information analysis module is configured to perform ship information screening analysis on the decoded information based on a particle filtering algorithm to obtain target ship information. A trajectory calculation and model construction module is configured to perform trajectory analysis and trajectory calculation processing based on the target ship information to obtain trajectory calculation information, pre-process real-time wind field data based on a sliding window to obtain pre-processed real-time wind field data, and construct a layered wind disturbance model based on the pre-processed real-time wind field data combined with multi-source weighted fusion. The ballistic trajectory optimization module is configured to determine a wind disturbance term based on the hierarchical wind disturbance model, correct and optimize the ballistic trajectory information based on the ballistic solution information, the wind disturbance term, and a disturbance influence coefficient matrix, obtain optimized ballistic trajectory information, generate an execution trajectory instruction set based on the optimized ballistic trajectory information using a B-spline curve smoothing method, and perform a UAV search and rescue drop based on the execution trajectory instruction set combined with real-time control feedback. The method includes: performing abnormality elimination processing on the real-time wind field data based on a sliding window and a median filtering method to obtain preprocessed real-time wind field data; determining a weight coefficient of the preprocessed real-time wind field data based on a variance reciprocal method, and performing multi-source weighted fusion on the preprocessed real-time wind field data based on the weight coefficient to obtain target real-time wind field data; extracting a wind speed vector based on the target real-time wind field data and a preset height layering, and constructing a three-dimensional wind field tensor based on the wind speed vector; and constructing a hierarchical wind disturbance model based on the three-dimensional wind field tensor.

Citation Information

Patent Citations

  • Fusion algorithm-based unmanned aerial vehicle focus drop point prediction and adjustment method and system

    CN120029324A

  • Air-drop target prediction method, device and equipment for multi-dimensional parameter dynamic evolution under multi-model driving

    CN120450168A