Unmanned aerial vehicle multi-field coupling dynamic safety control method and system for complex sea conditions

Abstract

The application discloses a kind of unmanned plane multi-field coupling dynamic safety control method and system for complex sea conditions, method includes: obtaining environment multi-source data, and carry out synchronous benchmark calibration processing, obtain environment background field matrix;Solving the field intensity distribution of the preset route of unmanned plane, and the dynamic comprehensive risk index of each point in route is calculated;Dynamic comprehensive risk index is compared with dangerous threshold value, when dynamic comprehensive risk index is less than dangerous threshold value, maintain current route until landing;When dynamic comprehensive risk index is greater than dangerous threshold value, according to route stage, trajectory is adjusted until landing.The application can accurately capture the nonlinear fluid mechanics distortion in complex space by real-time calculation of the dynamic comprehensive risk index of the preset route of unmanned plane, and comparison and adjustment with dangerous threshold value, when dynamic comprehensive risk index is greater than dangerous threshold value, trajectory is adjusted until landing, improve the safety and stability of unmanned plane flight landing.

Description

Technical Field

[0001] This invention relates to the field of unmanned aerial vehicle (UAV) control, and in particular to a method and system for multi-field coupled dynamic safety control of UAVs in complex sea conditions. Background Technology

[0002] With the deepening of the marine development strategy and the rapid development of ship automation, shipborne / unmanned aerial vehicles (UAVs) are playing an increasingly irreplaceable role in complex marine engineering tasks such as ocean search and rescue, maritime inspection, equipment status monitoring, and pre-deployment of supplies. However, unlike the stable operating environment on land, the weather conditions in the open ocean are constantly changing. During navigation and fixed-point operations, mother ships often experience severe six-degree-of-freedom motions, including rolling, pitching, and heaving. Furthermore, their superstructure and power exhaust systems create extremely complex fluid-structure interaction fields above the flight deck. When performing emergency recovery or high-sea-state take-off and landing missions, traditional UAVs mainly rely on manual remote control intervention or simple vertical fixed-point landing logic, making them highly susceptible to serious crashes due to sudden gusts of wind, visual blindness, or deck lifting. Therefore, the marine engineering community urgently needs an intelligent control system for UAVs that can adapt to harsh sea conditions around the clock, possessing feedforward environmental active avoidance and flexible anti-disturbance landing capabilities, to fundamentally ensure the safe collaborative operation of high-value aviation equipment and maritime mobile platforms.

[0003] Multi-field coupled environment perception and model predictive control (MPC) is a cutting-edge technology for autonomous navigation and dynamic risk avoidance of unmanned aerial vehicles (UAVs). Multi-field coupled feature extraction is applied to the control and monitoring of shipborne UAVs. By real-time mining of data output from the mother ship's weather radar and the underlying PLC system, it can accurately predict the complex distortion field in the sea-air boundary layer caused by strong winds, large waves, dense salt spray, and high-temperature thermal plumes generated by the mother ship's smokestack. This system-level information fusion design eliminates the need for redundant and easily interfered heavy weather sensors on the UAV, significantly reducing the UAV's payload and hardware modification costs. It fundamentally eliminates the risk of blind flight caused by the failure of a single airborne sensor in harsh marine environments, improving the system's engineering reliability and practicality.

[0004] Spatial phase matching landing is a flexible, collaborative recovery mechanism that combines the mother ship's swaying behavior with the UAV's dynamic characteristics. Applying spatial phase matching to UAV landing protection allows for high-frequency prediction of the mother ship's six-degree-of-freedom motion minimum time window (…). The system guides the drone to land during the phase of the smoothest movement on the deck. Combined with a three-dimensional predictive avoidance path generated by the MPC multi-objective optimization algorithm, it can proactively bypass the wake vortex and high-temperature exhaust gases generated by the superstructure of the mother ship. Its multi-level active defense design effectively breaks the rigid situation of traditional drones "forcibly penetrating the thermal field and crashing upon landing," further improving the fault tolerance and all-weather operation capability of shipborne auxiliary drone systems, reducing damage to drone motors caused by thermal stall or salt spray corrosion, and extending the service life of core electromechanical components.

[0005] Currently, existing technologies disclose methods for shipboard UAV landing control. However, existing UAV landing systems rely excessively on external visual cameras or lidar sensors, making them prone to target loss or blinding in humid, high-salt-spray, and exhaust-filled environments such as the tail section of the aircraft cabin. Furthermore, they lack a feedforward active environmental intervention module, making it impossible to prevent early minor airflow distortions along the flight path in real time. Existing research also discloses a traditional protection method based on error feedback, but its control logic lacks a multi-level predictive intervention strategy, making it unable to effectively cope with the initial latent disturbances of flexible sea breezes or thermal plumes. It also cannot effectively absorb and transform the nonlinear wind shear damage caused by multi-field coupling before it deteriorates, resulting in poor crash prevention and safe landing performance. Summary of the Invention

[0006] To address the problems of over-reliance on airborne sensors in existing perception architectures, which can lead to blindness, and the slow response and potential loss of control of traditional control strategies, this invention provides a multi-field coupled dynamic safety control method and system for UAVs in complex sea conditions. This method aims to achieve early warning and timely avoidance of crashes, thereby improving the reliability of UAV equipment in harsh marine environments.

[0007] The technical solution adopted in this invention is: A method for multi-field coupled dynamic safety control of unmanned aerial vehicles (UAVs) in complex sea conditions is provided, the method comprising: Acquire multi-source environmental data and perform synchronous benchmark calibration on the multi-source data to obtain the calibrated environmental background field matrix; The field strength distribution of the UAV's preset flight path is solved based on the environmental background field matrix, and the dynamic comprehensive risk index of "wind-wave-humidity-heat" at each point in the flight path is calculated based on the field strength distribution. The dynamic comprehensive risk index is compared with the risk threshold in real time. When the dynamic comprehensive risk index is less than the risk threshold, the current route is maintained until landing. When the dynamic comprehensive risk index is greater than the risk threshold, the trajectory is adjusted according to the route stage until landing.

[0008] According to the above scheme, the environmental multi-source data includes the six-degree-of-freedom attitude data of the UAV mothership, the real-time power of the main engine, the exhaust temperature, and the wind direction and speed sequence data.

[0009] According to the above scheme, the dynamic comprehensive risk index is specifically calculated by weighted fusion of wind field environmental parameters, wave environmental parameters, hot rain flow field parameters, and salt spray environmental parameters; the weight of each parameter is specifically allocated by the flight altitude of the UAV.

[0010] According to the above scheme, the wind field environmental parameters specifically include the distorted wind speed and maximum wind resistance threshold of the current coordinates of the UAV; the wave environmental parameters specifically include the real-time transient wave elevation and the safe wave height benchmark for the mother ship and UAV operations; the thermal rain flow field parameters specifically include the attenuation function of the high-temperature exhaust gas thermal plume flow field, the real-time power of the main engine and the exhaust temperature; and the salt spray environmental parameters specifically include the local space salt spray concentration index.

[0011] According to the above plan, trajectory adjustments based on flight path phases specifically include: When the horizontal distance between the drone and the mother ship's deck is less than a certain distance, the drone is determined to enter the landing preparation phase. At this time, high-frequency roll and pitch angular velocity data of the mother ship are collected and the minimum time window of the mother ship's sway is calculated. When it is calculated that the drone will be in the minimum time window within a certain period of time in the future, the drone is instructed to land at a certain vertical speed. When the UAV has not entered the landing preparation phase, the optimal control increment of the UAV trajectory data is solved by minimizing the multi-objective cost function within a certain time domain, and the UAV trajectory is adjusted based on the optimal control increment.

[0012] According to the above scheme, the multi-objective cost function specifically includes trajectory tracking deviation cost, multi-field coupling dynamic hazard penalty cost, and control increment smoothing cost.

[0013] According to the above scheme, the method also includes collecting the IMU attitude oscillation variance and the actual motor temperature rise curve of the UAV during the entire landing process after the UAV completes landing, and calculating and updating the danger threshold benchmark for the next UAV sortie based on the IMU attitude oscillation variance and the actual motor temperature rise curve.

[0014] A UAV multi-field coupled dynamic safety control system for complex sea conditions is also provided. The system is used to implement the UAV multi-field coupled dynamic safety control method for complex sea conditions described above. The system includes a data side, a solution side, and an execution side. The data side includes a meteorological instrument and wave measuring radar, a ship motion reference unit, and a ship's bottom-level PLC gateway, used to acquire multi-source environmental data; The solution side includes a spatiotemporal evolution model predictor, a dynamic comprehensive risk assessment module, an MPC multi-objective optimization solver, and a safety threshold database. It is used to solve the field strength distribution of the UAV's preset flight path based on the environmental background field matrix, and to calculate the "wind-wave-humidity-heat" dynamic comprehensive risk index of each point in the flight path based on the field strength distribution. The execution side includes an airborne status monitoring unit, a flight control motherboard calculation center, and a power drive and electronic speed control module. It is used to maintain the current flight path until landing when the dynamic comprehensive risk index is less than the risk threshold, and to adjust the trajectory according to the flight path stage until landing when the dynamic comprehensive risk index is greater than the risk threshold.

[0015] According to the above scheme, the ship's bottom-level PLC gateway is bidirectionally connected to the mother ship's main propulsion remote control system via an industrial communication bus; the spatiotemporal evolution model predictor is electrically connected to the data side; the dynamic comprehensive hazard assessment module is electrically connected to the spatiotemporal evolution model predictor; the MPC multi-objective optimization solver is bidirectionally connected to the dynamic comprehensive hazard assessment module and the airborne status monitoring unit; and the flight control motherboard calculation center is electrically connected to the MPC multi-objective optimization solver and the power drive and electronic control module.

[0016] A computer storage medium is also provided, which stores a computer program that can be executed by a processor. The computer program executes the above-described method for multi-field coupled dynamic safety control of UAVs in complex sea conditions.

[0017] The beneficial effects of this invention are as follows: By calculating the dynamic comprehensive risk index of "wind-wave-humidity-heat" for the preset flight path of the UAV in real time and comparing and adjusting it with the risk threshold, this invention can comprehensively consider the impact of complex sea conditions, high humidity, and high salt spray on the UAV flight path. It can accurately capture nonlinear hydrodynamic distortions in complex spaces, and when the dynamic comprehensive risk index is greater than the risk threshold, it can adjust the trajectory according to the flight path stage until landing. This realizes early warning and timely avoidance of UAV crashes in complex sea conditions, and improves the safety and stability of UAV flight and landing.

[0018] Furthermore, this invention achieves UAV landing by calculating the minimum time window of the mother ship's sway during the landing preparation phase, controlling the UAV to hover and wait in the dangerous phase, and smoothly enter the landing phase, thus realizing a flexible and coordinated landing with zero relative velocity and solving the problem of hard landing and destruction in high sea states. In addition, before entering the landing preparation phase, the optimal control increment of the UAV trajectory data is solved by minimizing the multi-objective cost function, forcing the UAV to perform transient three-dimensional flight around and lateral avoidance, effectively avoiding the power components being in a state of thermal overload for a long time, and further ensuring the safety and stability of UAV flight and landing.

[0019] Furthermore, this invention, by collecting the IMU attitude oscillation variance and the actual temperature rise curve of the motor throughout the entire landing process after the UAV completes its landing, and updating the danger threshold, ensures that the control system can adapt to the long-term changes in water quality (salinity, humidity) of mother ships of different tonnages and specific sea areas. This highly meets the current development needs of intelligent and unmanned marine equipment and enhances the application breadth of UAV equipment. Attached Figure Description

[0020] Figure 1 This is a schematic diagram of the method flow of the UAV multi-field coupling dynamic safety control method for complex sea conditions according to an embodiment of the present invention; Figure 2 This is an architecture diagram of a UAV multi-field coupled dynamic safety control system for complex sea conditions according to an embodiment of the present invention. Figure 3 This is a schematic diagram of the distribution of the high-temperature exhaust gas thermal plume flow field and the ship's wake distortion wind field in an embodiment of the present invention. Figure 4 This is a schematic diagram of the MPC multi-objective optimization solver architecture according to an embodiment of the present invention; Figure 5 This is a schematic diagram of the spatial phase matching landing mechanism according to an embodiment of the present invention; Figure 6 This is an overall block diagram of the control method according to an embodiment of the present invention. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0022] Currently, existing ship landing guidance systems rely excessively on external visual cameras or lidar on UAVs. In high-humidity, high-salt-fog marine environments, especially when traversing areas obscured by high-temperature smoke and dust from mother ship exhaust chimneys, these optical and wading sensors are highly susceptible to lens blurring, corrosion, or even "blinding." Due to the lack of advanced prediction of large-scale distorted flow fields, if a single sensor fails due to disturbance, the UAV is highly likely to be trapped in a blind flight situation, resulting in extremely poor engineering reliability. Furthermore, existing flight control systems (such as variable-parameter PID controllers) generally employ a passive protection logic of "error feedback - attitude compensation." When the system detects attitude or altitude deviations, the UAV is often already deep in the core region of a dangerous high-temperature, low-pressure flow field. Executing delayed thrust compensation at this point can easily lead to "thermal stall" or overload tripping of the drive motor, failing to effectively absorb and convert the nonlinear damage caused by multi-field coupling. Additionally, current mainstream ship landing strategies often employ a single vertical constant-speed descent, or degenerate into manual remote control fine-tuning in adverse sea conditions. This rigid mechanism completely ignores the violent six-degree-of-freedom (6-DOF) random swaying of the mother ship deck under high sea states. Forced blind landing is very likely to cause the drone to collide rigidly with the wave crests or troughs of the waves that are hitting the deck upwards; while manual remote control cannot exceed the limits of physiological reaction, resulting in an extremely high equipment damage rate.

[0023] To address the aforementioned problems, embodiments of the present invention provide a method for multi-field coupled dynamic safety control of unmanned aerial vehicles (UAVs) in complex sea conditions, such as... Figure 1 As shown, the method includes: S1. Acquire multi-source environmental data and perform synchronous benchmark calibration on the multi-source data to obtain the calibrated environmental background field matrix.

[0024] S2. Solve the field strength distribution of the UAV's preset flight path based on the environmental background field matrix, and calculate the dynamic comprehensive risk index of "wind-wave-humidity-heat" at each point in the flight path based on the field strength distribution.

[0025] S3. Compare the dynamic comprehensive risk index with the risk threshold in real time. When the dynamic comprehensive risk index is less than the risk threshold, maintain the current route until landing. When the dynamic comprehensive risk index is greater than the risk threshold, adjust the trajectory according to the route stage until landing.

[0026] Preferably, to achieve the data acquisition and processing of this embodiment, this embodiment provides a UAV multi-field coupled dynamic safety control system for complex sea conditions, such as... Figure 2 As shown, it includes a data side 100, a solution side 200, and an execution side 300. The data side 100 includes a meteorological instrument and wave measuring radar 101, a ship motion reference unit 102, and a ship bottom-level PLC gateway 103. The solution side 200 includes a spatiotemporal evolution model predictor 201, a dynamic comprehensive risk assessment module 202, an MPC multi-objective optimization solver 203, and a safety threshold database 204. The execution side 300 includes an airborne status monitoring unit 301, a flight control motherboard solution center 302, and a power drive and electronic speed control module 303.

[0027] Specifically, the ship's bottom-level PLC gateway 103 is bidirectionally connected to the mother ship's main propulsion remote control system via an industrial communication bus, and is used to extract the current operating speed and load percentage of the main engine in real time and send them to the calculation side; The spatiotemporal evolution model predictor 201 is electrically connected to the data side 100 and is used to fuse meteorological data and underlying load parameters to reconstruct the spatiotemporal boundary of multi-field coupling. The dynamic integrated hazard assessment module 202 is electrically connected to the spatiotemporal evolution model predictor 201 and is used to calculate the spatial hazard index ahead of the flight path during the flight cycle. The MPC multi-objective optimization solver 203 is bidirectionally connected to the dynamic integrated risk assessment module 202 and the airborne status monitoring unit 301, and is used to solve the optimal control law including risk penalties. The flight control motherboard calculation center 302 is electrically connected to the MPC multi-objective optimization solver 203 and the power drive and ESC module 303, and is used to analyze the control law and allocate the motor thrust vector.

[0028] Furthermore, in this embodiment, the spatiotemporal evolution model predictor 201 uses the load parameters extracted by the ship's bottom-level PLC gateway 103 as the heat source feedforward input. For example... Figure 3 As shown, the integrated design reconstructs the spatial distribution of the high-temperature exhaust plume flow field 403 above the ship's deck and the distorted wind field 404 of the ship's wake, transforming the lagging passive temperature alarm into an advanced flow field predictive distribution map, effectively reducing the risk of UAVs encountering transient wind shear and thermal stall during the approach phase.

[0029] The airborne condition monitoring unit 301 extracts IMU attitude drift features at high frequency. The design of two-dimensional feature extraction in both the time and frequency domains facilitates the system to simultaneously capture low-frequency position drift caused by macroscopic wind and waves, as well as high-frequency temperature abrupt changes and vibrations caused by the erosion of the motor in the high-concentration salt spray field 402.

[0030] like Figure 4 As shown, the MPC multi-objective optimization solver 203 adopts a cost function architecture that introduces dynamic penalty weights. This design enables the controller to respond to environmental degradation within milliseconds, forcibly inserting control increments for sideslip or spiking into the normal trajectory tracking logic, ensuring a smooth flight trajectory under normal conditions, and quickly establishing an evasion path in the event of severe threats.

[0031] like Figure 5 As shown, the sequence of actions for the spatial phase-matching landing mechanism is as follows: In the mother ship's severe rolling danger zone (roll angle absolute value > 5°), the descent channel is forcibly locked, and the UAV hovers outside the deck to wait; when the optimal landing window is predicted to open (roll angle > 5°), the descent channel is forcibly locked, and the UAV hovers outside the deck to wait; angular velocity At that moment, the landing command is issued instantly. This phase-matching design reconstructs the landing timing direction to prevent the UAV from having a rigid collision with the deck, and the synchronously activated recoil thrust logic constitutes a complete flexible contact-and-ship dissipation path.

[0032] Specifically, the environmental multi-source data includes the six-degree-of-freedom attitude data of the unmanned aerial vehicle (UAV) mothership, real-time main engine power, exhaust temperature, and wind direction and speed sequence data. Data side 100 is configured according to a preset... The sampling frequency is used to collect the mother ship's six-degree-of-freedom attitude data and the wind direction and speed sequences from the weather instrument in real time. Simultaneously, the real-time power of the main engine is read from the ship's bottom-level PLC gateway 103. With exhaust temperature The ship's power grid frequency interference is filtered out by a digital low-pass filter, and the basic environmental background field matrix for the current voyage is calibrated and generated.

[0033] Specifically, the field strength distribution along the pre-set flight path of the UAV is constructed by the spatiotemporal evolution model predictor 201. The dynamic comprehensive hazard index is obtained through a sliding window of spatial observation and real-time calculation; it is specifically calculated based on a weighted fusion of wind field environmental parameters, wave environmental parameters, thermal rain flow field parameters, and salt spray environmental parameters; the weights of each parameter are specifically allocated by the UAV's flight altitude; and the current spatial coordinates are used for calculation. The dynamic comprehensive risk index of "wind-wave-humidity-heat" at the location (Dimensionless) can be expressed as:

[0034] In the formula, The distorted wind speed at the current coordinates. Transient wave elevation; This is the thermal plume temperature field decay function derived from the host load; Based on the current flight altitude of the drone Dynamically adaptively assigned weighting coefficients, for example, when height... At that time, wave weight Magnified many times; when the height At that time, thermal plume weight Dominant.

[0035] Specifically, This represents a multi-field coupled dynamic comprehensive risk index. It is a dimensionless comprehensive score; the higher the value, the higher the risk of crash or damage to the UAV at its current three-dimensional spatial location. This method compares it with a preset safety threshold (…). The comparison is performed to determine whether to trigger active avoidance.

[0036] in, These represent the dynamic weighting coefficients for each physical field. They correspond to the penalty percentages for wind, waves, heat, and salt spray, respectively. In actual control, these coefficients are not fixed but are adaptively adjusted according to the drone's flight altitude. For example, when the drone flies low over the sea surface, the wave weighting is adjusted accordingly. and salt spray weight This will be significantly amplified; when the drone climbs to the vicinity of the mothership's smokestack, the thermal plume weight... It will take the lead.

[0037] Among them, the wind field environmental parameters are aerodynamic characteristics, including the current spatial coordinates. This represents the actual wind speed at the drone's current location after it is obstructed by the superstructure of the mother ship, resulting in a wake vortex; and the maximum wind resistance threshold designed for the drone. The value placed in the denominator is used to normalize the wind speed data (i.e., convert it to a percentage or proportion between 0 and 1), measuring how close the current wind speed is to the drone's limit.

[0038] Among them, wave environment parameters are hydrodynamic characteristics, including real-time transient wave elevation. The current actual wave height obtained through shipborne wave-measuring radar; and the safe wave height benchmark for mother ship and UAV operations. It is also used as the denominator for normalization to assess the threat level of wave impact or splashing water to low-altitude flight paths.

[0039] Among them, the thermal plume flow field parameters are thermodynamic characteristics, including the attenuation function of the high-temperature exhaust gas thermal plume flow field. This is a nonlinear fluid dynamics function used to extrapolate the thermal intensity of the high-temperature exhaust gas ejected from the exhaust pipe as it diffuses through the air and reaches the drone's location; and the real-time operating power / load of the ship's main engine obtained through the underlying PLC. The greater its power, the wider its exhaust volume and disturbance range; and the real-time exhaust temperature of the ship's main engine. The higher the temperature, the stronger the destructive power of the thermal plume core.

[0040] Among them, the salt spray environmental parameters are corrosion and humidity characteristics, including the local space salt spray concentration index. This reflects the corrosive threat to the motor insulation and the machine body posed by seawater spray or high-concentration salt spray kicked up by the propeller under severe sea conditions.

[0041] Specifically, the index is evaluated in real time based on the calculation results. With preset safety threshold The specific criteria and execution logic are as follows: like This indicates that the airspace ahead is in good condition, and the system will maintain the current baseline approach route. like Furthermore, since the drone has not yet entered the final landing preparation phase, it is determined to face a "moderate to severe multi-field coupling threat" (such as directly facing a hot exhaust plume). The MPC multi-target optimization solver 203 triggers a level one defense command in the prediction time domain. Internal rolling solution of multi-objective cost function ; like And the drone's horizontal distance from the deck Having passed the danger zone and entered the final landing preparation phase, the secondary defense phase matching command is triggered. The flight control motherboard's calculation center 302 extracts the mother ship's high-frequency roll and pitch angular velocity data to estimate the minimum time window of the mother ship's sway. .

[0042] Specifically, multi-objective cost function Including trajectory tracking deviation cost, multi-field coupled dynamic hazard penalty cost, and control increment smoothing cost, it can be expressed as:

[0043] In the formula, For trajectory tracking weight matrix, For extremely dangerous penalty scalars (such as set to) level), To control the incremental smoothing weight matrix, the optimal control increment is solved using quadratic programming (QP). Forced drones It performs transient lateral translation or lifting movements to bypass the core region of the thermal plume.

[0044] Specifically, among them, Representing the total cost function value, the goal of the MPC solver is to find a sequence of actions, within acceptable constraints, that satisfies this... The value is the smallest; Indicates the prediction time domain. It refers to how far into the future the algorithm "looks," for example... This means that the controller will predict the situation within the next 10 time steps at the current moment; Represents a discrete-time step sequence variable, indicating the first step in the prediction time domain. One step in predicting the future; This refers to the control time domain, indicating how many time steps the controller will change its action in the future (usually...). ), passed After that step, assume that the control commands remain constant.

[0045] This represents the trajectory tracking deviation cost, used to penalize the drone for deviating from the planned approach route, thus encouraging the drone to fly along a safe landing path as much as possible; among which... Indicates the future number Predicted state variables of the unmanned aerial vehicle (UAV). These include its three-dimensional spatial coordinates or attitude angles; Indicates the future number The reference / expected state quantity for the step. That is, the originally planned standard approach route to the deck; This represents the trajectory tracking weight matrix, a diagonal matrix that determines the system's tolerance for "deviation from the flight path." The higher the value, the more rigidly the drone will fly along the original planned route.

[0046] The penalty cost for multi-field coupled dynamic hazards is the core active hazard avoidance term of this method. When a thermal plume or strong wind shear appears in front of the UAV, due to the existence of this term, the system will be willing to sacrifice some flight path tracking accuracy (the first term increases) to force a deviation to avoid the hazard and protect the equipment. Indicates the future number The multi-field coupled dynamic hazard index represents the combined lethality of wind, waves, heat, and salt spray at that location in the future; A scalar / weight representing a very high danger penalty (e.g., set to...) (Level), granting hazard avoidance a very high "veto power," once the future path is predicted. A dramatic increase This will lead to total cost The instantaneous explosion forced the solver to immediately abandon its original course and generate a yaw avoidance maneuver.

[0047] This represents the cost of controlling incremental smoothing, used to penalize overly aggressive rudder movements or sudden thrust changes, preventing the UAV from experiencing propulsion system failure or attitude loss due to excessively abrupt actions in rough sea conditions; among which... Indicates the future number The increment of the control command. That is, the change in thrust or rotor speed given to the motor in this second compared to the previous second; This represents the control increment smoothing weight matrix. The larger the value, the more the system tends to execute gentle and smooth control actions, which can effectively avoid extreme overshoot when avoiding danger (such as overturning and crashing into the sea by directly applying full rudder to avoid thermal airflow).

[0048] Specifically, when calculating the minimum time window of the mother ship's sway, the absolute value of the transient roll angle of the mother ship is... Or roll rate At that time, the program locks the descent channel, and the drone hovers on the side of the deck, waiting; only when the future is predicted... Internal Window (swing angle) and angular velocity At that time, a step descent command is issued to the UAV, which, in conjunction with the recoil thrust of the power drive and electronic speed control module, enables the UAV to achieve a vertical speed relative to the mother ship deck. The ship was able to make a soft touchdown in time to avoid a hard landing and destruction.

[0049] Specifically, the overall flowchart of the UAV multi-field coupled dynamic safety control method described above is as follows: Figure 6 As shown.

[0050] Preferably, the method further includes: S4, after the UAV completes landing, collecting the IMU attitude oscillation variance and the actual motor temperature rise curve of the UAV during the entire landing process, and calculating and updating the danger threshold benchmark for the next UAV sortie based on the IMU attitude oscillation variance and the actual motor temperature rise curve.

[0051] Specifically, after a successful landing, the system enters the state return phase. The airborne state monitoring unit 301 extracts the IMU attitude oscillation variance and the actual motor temperature rise curves throughout the landing process, and automatically updates the safety threshold benchmark for the next UAV sortie using the built-in edge computing self-learning module. This enables adaptive safety protection capabilities tailored to specific mother ships and sea weather conditions.

[0052] In addition, embodiments of the present invention also provide a computer storage medium storing a computer program that can be executed by a processor, the computer program executing the above-described method for multi-field coupled dynamic safety control of unmanned aerial vehicles in complex sea conditions.

[0053] In summary, the UAV multi-field coupled dynamic safety control method and system for complex sea conditions provided in this embodiment have the following technical effects: This invention achieves a perception topology upgrade based on multi-source heterogeneous information fusion, significantly improving the system's engineering reliability in harsh sea conditions and exhaust environments. Existing shipborne UAV landing guidance largely relies on sophisticated external sensors such as airborne visual cameras or lidar. These components are highly susceptible to target loss or "blinding" failure under conditions of high humidity, high salt spray, and high-temperature exhaust gases from the main and auxiliary engines at the stern of the ship. This invention avoids the single physical dependence on heavy airborne sensors and innovatively extracts the host load parameters of the mother ship's underlying PLC gateway directly through a communication protocol. This solution reduces the payload cost of UAVs while fundamentally eliminating the risk of blind flight and crashes caused by the failure of water / smoke sensors, significantly enhancing the system's adaptability and all-weather operation capability in extreme weather conditions.

[0054] A multi-field coupled dynamic scoring algorithm based on height-adaptive weights was constructed, achieving high-precision early warning under complex sea-air boundary layers. Addressing the wake vortex and thermal plume disturbances that easily arise above the flight deck when mother ships navigate against wind, this invention overcomes the limitations of traditional single weather parameter (such as wind speed alone) judgments and constructs a multi-field coupled dynamic hazard assessment model of "wind-wave-humidity-heat". This algorithm can dynamically allocate weights based on the UAV's real-time flight altitude (e.g., amplifying wave elevation weights at low altitudes and amplifying thermal plume weights at high altitudes) to accurately capture nonlinear hydrodynamic distortions in complex spaces. While ensuring accurate early warning, this gridded model effectively avoids overfitting and is suitable for real-time deployment on the edge computing platform of the UAV flight control motherboard.

[0055] A control strategy shifting from "passive error feedback" to "feedforward predictive avoidance" was constructed, effectively preventing thermal stall and attitude loss of the power system. Traditional UAV PID control mechanisms suffer from severe time delay and hysteresis, typically only implementing passive attitude compensation after the UAV has already entered a high-temperature thermal plume, experienced a sharp drop in altitude, or suffered motor overload. At this point, the sudden drop in lift can easily lead to an irreversible crash. This invention proactively intervenes during the danger latency period using an MPC multi-objective optimization solver, predicting the time domain... Internal rolling predicts future trajectory. By issuing transient avoidance control laws in advance, the rotor motor is prevented from being in a state of excessive heat load due to high temperature and low air density for a long time, which significantly reduces the irreversible physical damage to core electromechanical components caused by salt spray and high temperature.

[0056] A three-dimensional flexible avoidance mechanism based on dynamic penalty using a cost function was created to maximize the continuous operational efficiency of shipborne UAVs. Addressing the complex transient characteristics of sea surface airflow distortion, this invention breaks away from the traditional rigid protection logic of "hovering or returning to base upon encountering danger." When a moderate to severe threat is detected in the airspace ahead, the system, based on a multi-objective cost function... Significantly increase the penalty for dangerous situations The optimal control increment is obtained by forcibly solving a quadratic programming (QP) algorithm. The mechanism allows UAVs to bypass the core of the hot plume from the side with minimal cost without interrupting their approach and landing missions, thus achieving dynamic route self-repair and significantly improving the continuous availability of search and rescue or inspection missions.

[0057] A ship-machine cooperative spatial phase-matching landing strategy was designed, fundamentally solving the problem of deck swaying and subsequent crashes under high sea states. Facing the landing conditions under high sea states, this invention overcomes the control limitations of UAVs blindly performing vertical constant-velocity descent and proposes a cooperative landing strategy that predicts the six degrees of freedom motion of the mother ship. The system extracts the high-frequency roll and pitch angular velocities of the ship's MRU (Main Unit) to predict and lock the minimum time window of deck swaying. (roll angle and angular velocity) By forcibly locking the descent channel during the danger period and issuing a step-in command during the window of opportunity, the relative vertical velocity of the UAV at the moment of contact with the ship is ensured. It fundamentally replaces the traditional, extremely high-risk "manual remote-controlled blind landing" mode, eliminating the structural fractures caused by hard landing.

[0058] It possesses threshold self-learning and state regression evolution capabilities, significantly improving the system's adaptive and collaborative level for specific motherships and sea areas. Due to the differences in the aerodynamic characteristics of the superstructures of different motherships, fixed landing safety thresholds are often difficult to apply universally. This invention, after a successful UAV landing, extracts the IMU attitude oscillation variance and motor temperature rise curves throughout the entire process through the onboard state monitoring unit, and uses an algorithm model to establish environmental background field benchmarks and hazard thresholds. It performs self-learning and automatic updates. This closed-loop state regression mechanism ensures that the control system can adapt to long-term changes in water quality (salinity, humidity) of different tonnage mother ships and specific sea areas, which is highly in line with the current development needs of intelligent and unmanned marine equipment.

[0059] It should be noted that, depending on the implementation needs, the various steps / components described in this application can be broken down into more steps / components, or two or more steps / components or parts of the operation of steps / components can be combined into new steps / components to achieve the purpose of this invention.

[0060] The order of the steps in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0061] It should be understood that those skilled in the art can make improvements or modifications based on the above description, and all such improvements and modifications should fall within the protection scope of the appended claims.

Claims

1. A multi-field coupled dynamic safety control method for unmanned aerial vehicles (UAVs) in complex sea conditions, characterized in that, The method includes: Acquire multi-source environmental data and perform synchronous benchmark calibration on the multi-source data to obtain the calibrated environmental background field matrix; The field strength distribution of the UAV's preset flight path is solved based on the environmental background field matrix, and the dynamic comprehensive risk index of "wind-wave-humidity-heat" at each point in the flight path is calculated based on the field strength distribution. The dynamic comprehensive risk index is compared with the risk threshold in real time. When the dynamic comprehensive risk index is less than the risk threshold, the current route is maintained until landing. When the dynamic comprehensive risk index is greater than the risk threshold, the trajectory is adjusted according to the route stage until landing.

2. The UAV multi-field coupled dynamic safety control method for complex sea conditions according to claim 1, characterized in that, The environmental multi-source data includes the six-degree-of-freedom attitude data of the unmanned aerial vehicle (UAV) mothership, real-time power of the main engine, exhaust temperature, and wind direction and speed sequence data.

3. The UAV multi-field coupled dynamic safety control method for complex sea conditions according to claim 1, characterized in that, The dynamic comprehensive hazard index is specifically calculated based on the weighted fusion of wind field environmental parameters, wave environmental parameters, thermal rain flow field parameters, and salt spray environmental parameters; the weight of each parameter is specifically allocated by the flight altitude of the UAV.

4. The UAV multi-field coupled dynamic safety control method for complex sea conditions according to claim 3, characterized in that, The wind field environmental parameters specifically include the distorted wind speed and maximum wind resistance threshold at the current coordinates of the UAV; the wave environmental parameters specifically include the real-time transient wave elevation and the safe wave height benchmark for the mother ship and UAV operations; the thermal rain flow field parameters specifically include the attenuation function of the high-temperature exhaust gas thermal plume flow field, the real-time power of the main engine, and the exhaust temperature; and the salt spray environmental parameters specifically include the local space salt spray concentration index.

5. The UAV multi-field coupled dynamic safety control method for complex sea conditions according to claim 1, characterized in that, Adjusting the trajectory based on the flight path phase specifically includes: When the horizontal distance between the drone and the mother ship's deck is less than a certain distance, the drone is determined to enter the landing preparation phase. At this time, high-frequency roll and pitch angular velocity data of the mother ship are collected and the minimum time window of the mother ship's sway is calculated. When it is calculated that the drone will be in the minimum time window within a certain period of time in the future, the drone is instructed to land at a certain vertical speed. When the UAV has not entered the landing preparation phase, the optimal control increment of the UAV trajectory data is solved by minimizing the multi-objective cost function within a certain time domain, and the UAV trajectory is adjusted based on the optimal control increment.

6. The UAV multi-field coupled dynamic safety control method for complex sea conditions according to claim 5, characterized in that, The multi-objective cost function specifically includes trajectory tracking deviation cost, multi-field coupled dynamic hazard penalty cost, and control increment smoothing cost.

7. The UAV multi-field coupled dynamic safety control method for complex sea conditions according to claim 1, characterized in that, The method also includes collecting the IMU attitude oscillation variance and the actual motor temperature rise curve during the entire landing process of the UAV after it has landed, and calculating and updating the danger threshold benchmark for the next UAV sortie based on the IMU attitude oscillation variance and the actual motor temperature rise curve.

8. A multi-field coupled dynamic safety control system for unmanned aerial vehicles (UAVs) accommodating complex sea conditions, characterized in that: The system is used to implement the UAV multi-field coupled dynamic safety control method for complex sea conditions as described in any one of claims 1-7, and the system includes a data side, a solution side, and an execution side; The data side includes a meteorological instrument and wave measuring radar, a ship motion reference unit, and a ship's bottom-level PLC gateway, used to acquire multi-source environmental data; The solution side includes a spatiotemporal evolution model predictor, a dynamic comprehensive risk assessment module, an MPC multi-objective optimization solver, and a safety threshold database. It is used to solve the field strength distribution of the UAV's preset flight path based on the environmental background field matrix, and to calculate the "wind-wave-humidity-heat" dynamic comprehensive risk index of each point in the flight path based on the field strength distribution. The execution side includes an airborne status monitoring unit, a flight control motherboard calculation center, and a power drive and electronic speed control module. It is used to maintain the current flight path until landing when the dynamic comprehensive risk index is less than the risk threshold, and to adjust the trajectory according to the flight path stage until landing when the dynamic comprehensive risk index is greater than the risk threshold.

9. The UAV multi-field coupled dynamic safety control system for complex sea conditions according to claim 8, characterized in that, The ship's bottom-level PLC gateway is bidirectionally connected to the mother ship's main propulsion remote control system via an industrial communication bus; the spatiotemporal evolution model predictor is electrically connected to the data side; the dynamic comprehensive hazard assessment module is electrically connected to the spatiotemporal evolution model predictor; the MPC multi-objective optimization solver is bidirectionally connected to the dynamic comprehensive hazard assessment module and the airborne status monitoring unit; the flight control motherboard calculation center is electrically connected to the MPC multi-objective optimization solver and the power drive and electronic control module.

10. A computer storage medium, characterized in that, It contains a computer program that can be executed by a processor, which performs the UAV multi-field coupled dynamic safety control method for complex sea conditions as described in any one of claims 1-7.

Images