Ship-based aircraft autonomous take-off and landing control method and related equipment

Abstract

The invention relates to the technical field of ship-based aircraft autonomous control, and discloses a ship-based aircraft autonomous take-off and landing control method, which performs parallel calculation based on a kinematics and coordinate conversion method and a deep learning model to form dual redundancy of a sensing level. And the optimal state estimation of the aircraft relative to the ship landing point is selected through respective reliability indexes for the calculation result based on kinematics and coordinate conversion and the calculation result based on the deep learning model, so that the fault-tolerant capability and reliability of the system are greatly improved. Moreover, higher precision can be realized by relying on the stability based on kinematics and coordinate transformation calculation and the optimization based on a deep learning model. In addition, relative pose observation values and distance measurement values of the sensors are fused, and navigation continuity is facilitated. The invention further provides an aircraft, a control system, equipment, a storage medium and a program product using the method.

Description

Technical Field

[0001] This invention relates to the field of autonomous control technology for shipborne aircraft, and in particular to an autonomous take-off and landing control method and related equipment for shipborne aircraft. Background Technology

[0002] With the widespread application of drone technology in maritime surveillance, marine exploration, and emergency rescue, achieving autonomous take-off and landing of drones from moving vessels has become a critical requirement. However, the complex marine environment and the real-time changes in position and attitude of ships as six-degree-of-freedom motion platforms pose a severe challenge to the precise take-off and landing control of drones.

[0003] Existing technologies primarily rely on high-precision GNSS (Global Navigation Satellite System), such as RTK (Real-Time Kinematic), to calculate the relative position by having both the ship and the UAV independently calculate their absolute positions using high-precision GNSS data. This method suffers from uncorrelated residual errors in the two independent calculation systems, such as multipath effects and receiver noise. Subtracting these errors after calculation amplifies the errors, leading to reduced relative positioning accuracy. Furthermore, GNSS signals are easily blocked when the UAV approaches the ship's side, causing positioning failure. When the main GNSS sensor experiences systematic errors or malfunctions, it cannot be resolved or function properly. Traditional kinematic and coordinate transformation-based methods are insufficient to compensate for errors such as time-varying system delays, resulting in poor UAV landing accuracy on ships. Summary of the Invention

[0004] The purpose of this invention is to provide a highly accurate and reliable autonomous take-off and landing control method and related equipment for shipborne aircraft.

[0005] To achieve the above objectives, the present invention provides the following technical solution: In a first aspect, the present invention provides an autonomous take-off and landing control method for shipborne aircraft, comprising: The first relative state estimate and the second relative state estimate of the aircraft relative to the ship landing point are obtained. The first relative state estimate is the relative position and velocity estimate of the aircraft and the ship landing point calculated by kinematics and coordinate transformation. The second relative state estimate is the relative position and velocity estimate of the aircraft and the ship landing point inferred by a deep learning model. Based on the first relative state estimate and the second relative state estimate, the aircraft-ship relative state estimate is output. Control commands for driving the aircraft are generated based on the aircraft-ship relative state estimation, and the aircraft is driven according to the control commands.

[0006] In a second aspect, the present invention provides an aircraft, comprising: Airborne data link communication module, used to receive data from the ship; Airborne flight control computer, the airborne flight control computer comprising: The relative state acquisition module is used to acquire a first relative state estimate and a second relative state estimate of the aircraft relative to the ship landing point. The first relative state estimate is the relative position and velocity estimate of the aircraft and the ship landing point calculated by kinematics and coordinate transformation. The second relative state estimate is the relative position and velocity estimate of the aircraft and the ship landing point inferred by a deep learning model. The status output module is used to output an aircraft-ship relative status estimate based on the first relative status estimate and the second relative status estimate; The takeoff and landing control module is used to generate control commands to drive the aircraft based on the aircraft-ship relative state estimation, and drive the aircraft according to the control commands.

[0007] Thirdly, the present invention provides an autonomous take-off and landing control system for shipborne aircraft, including a shipborne subsystem and an airborne subsystem; The shipborne subsystem is used to acquire the ship's navigation and positioning data and transmit the ship's navigation and positioning data to the aircraft; the ship's navigation and positioning data includes the ship's raw GNSS / INS observation data and RTCM differential corrections, as well as the ship's navigation information calculated based on the ship's raw GNSS / INS observation data; The airborne subsystem includes: Airborne data link communication module, used to receive data from shipborne subsystems; Airborne flight control computer, including: The relative state acquisition module is used to acquire a first relative state estimate and a second relative state estimate of the aircraft relative to the ship landing point. The first relative state estimate is the relative position and velocity estimate of the aircraft and the ship landing point calculated by kinematics and coordinate transformation. The second relative state estimate is the relative position and velocity estimate of the aircraft and the ship landing point inferred by a deep learning model. The status output module is used to output an aircraft-ship relative status estimate based on the first relative status estimate and the second relative status estimate; The takeoff and landing control module is used to generate control commands to drive the aircraft based on the aircraft-ship relative state estimation, and drive the aircraft according to the control commands.

[0008] Fourthly, the present invention provides an electronic device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the above-described autonomous take-off and landing control method for shipborne aircraft.

[0009] Fifthly, the present invention provides a storage medium, which is a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the steps of the above-described autonomous take-off and landing control method for shipborne aircraft.

[0010] Sixthly, the present invention provides a computer program product comprising a computer program that, when executed by a processor, implements the steps of the above-described autonomous take-off and landing control method for shipborne aircraft.

[0011] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention employs a kinematics-and-coordinate transformation-based method and a deep learning model for parallel computation, creating dual redundancy at the perception layer. This effectively mitigates the systemic failure risks caused by model mismatch, signal interference, or multipath effects inherent in single-method approaches. Furthermore, by selecting the optimal aircraft state estimate relative to the ship's landing point based on the reliability indices of both the kinematics-and-coordinate transformation-based and deep learning model-based computations, the invention significantly enhances the system's fault tolerance and reliability, ensuring safety. Moreover, higher accuracy can be achieved by leveraging the stability of the kinematics-and-coordinate transformation-based computation and the optimization capabilities of the deep learning model. Additionally, by integrating relative pose observations and ranging values ​​from sensors, the system can maintain stable and smooth relative state estimations even when GNSS signals are obstructed by the ship's hull or severely degraded by multipath propagation on the sea surface, ensuring navigation continuity throughout the approach, hovering, and touchdown maneuvers. Attached Figure Description

[0012] Figure 1 This is a flowchart of the autonomous take-off and landing control method for shipborne aircraft according to Embodiment 1 of the present invention.

[0013] Figure 2 This is a flowchart of the autonomous take-off and landing control method for shipborne aircraft according to Embodiment 2 of the present invention.

[0014] Figure 3 This is a schematic diagram of the autonomous take-off and landing control system for shipborne aircraft according to Embodiment 4 of the present invention.

[0015] Figure 4 This is a schematic diagram of the equipment structure of the hardware operating environment involved in the autonomous take-off and landing control method for shipborne aircraft in this embodiment of the invention. Detailed Implementation

[0016] The specific embodiments of the present invention will be described in further detail below with reference to the accompanying drawings and examples. The following examples are for illustrative purposes only and are not intended to limit the scope of the invention.

[0017] Example 1 like Figure 1 As shown, a preferred embodiment of the present invention provides a method for autonomous takeoff and landing control of a shipborne aircraft, comprising: S101. Obtain a first relative state estimate and a second relative state estimate of the aircraft relative to the ship landing point. The first relative state estimate is the relative position and velocity estimate of the aircraft and the ship landing point calculated by kinematics and coordinate transformation. The second relative state estimate is the relative position and velocity estimate of the aircraft and the ship landing point inferred by a deep learning model. S102. Based on the first relative state estimate and the second relative state estimate, output the aircraft-ship relative state estimate; S103. Generate control commands to drive the aircraft based on the aircraft-ship relative state estimation, and drive the aircraft according to the control commands.

[0018] It should be noted that the aircraft described in this invention may be a drone, a manned aircraft, etc.

[0019] Specifically, before obtaining the first relative state estimate and the second relative state estimate of the aircraft relative to the ship's landing point, the process includes: Acquire raw GNSS / INS observation data and RTCM differential corrections for the ship; The navigation information of the ship is calculated based on the raw GNSS / INS observation data of the ship. The navigation information of the ship includes the ship's absolute position, speed, attitude, linear acceleration and angular velocity information. Acquire raw GNSS / INS observation data of the aircraft; The aircraft's absolute navigation information, including its absolute position and absolute velocity, is obtained by calculating the RTCM differential corrections and the aircraft's raw GNSS / INS observation data. The RTCM differential corrections received from the ship are used to perform real-time differential corrections on the aircraft's GNSS observations, thereby outputting the aircraft's high-precision absolute position, velocity, and attitude information, providing an accurate absolute positioning reference for subsequent relative state calculations. Based on the navigation information of the ship and the absolute navigation information of the aircraft itself, the first relative state estimate is calculated through kinematics and coordinate transformation; Based on the aircraft's raw GNSS / INS observation data, the aircraft's own absolute navigation information, the ship's raw GNSS / INS observation data, and the ship's navigation information, a second relative state estimate is output through deep learning model inference.

[0020] In some embodiments, the calculation of the first relative state estimate based on the ship's navigation information and the aircraft's own absolute navigation information through kinematics and coordinate transformation includes: Obtain the pre-calibrated pole offset at the ship's landing point; Calculate the attitude rotation matrix from the ship coordinate system to the navigation coordinate system based on the ship's attitude in the ship's navigation information; The absolute position of the ship's landing point in the navigation coordinate system is calculated based on the offset of the ship's landing point arm, the attitude rotation matrix, and the ship's absolute position in the ship's navigation information. The relative position of the aircraft to the ship's landing point is calculated based on the aircraft's absolute position in its own absolute navigation information and the absolute position of the ship's landing point in the navigation coordinate system. The relative velocity of the aircraft to the landing point of the ship is calculated based on the absolute velocity of the aircraft, the velocity of the ship in the navigation information of the ship, the attitude rotation matrix, the angular velocity information in the navigation information of the ship, and the offset of the control arm at the landing point of the ship. The first relative state estimate is obtained based on the relative position of the aircraft with respect to the ship's landing point and the relative speed of the aircraft with respect to the ship's landing point.

[0021] Specifically, the calculation of the first relative state estimate based on the ship's navigation information and the aircraft's own absolute navigation information, through kinematics and coordinate transformation, includes: Obtain the pre-calibrated pole arm offset at the ship landing point ; Calculate the attitude rotation matrix from the ship coordinate system to the navigation coordinate system based on the ship's attitude from the navigation information. ; Calculate the absolute position of the ship's landing point in the navigation coordinate system. : , in, The absolute position of the vessel in the vessel's navigation information; Calculate the relative position of the aircraft with respect to the ship's landing point : , in, This refers to the absolute position of the aircraft; Calculate the linear velocity compensation caused by the ship's rotational motion at the landing point, and further calculate the relative velocity of the aircraft with respect to the landing point. : , in, For the absolute speed of the aircraft, The speed of the ship in the navigation information of the ship. The angular velocity information in the ship's navigation information; based on the pre-calibrated offset of the ship's landing point boom. Angular velocity information in the ship's navigation information Calculate the linear velocity compensation caused by the ship's rotational motion at the landing point; then, based on the linear velocity compensation caused by the ship's rotational motion at the landing point and the attitude rotation matrix... The ship's speed in the navigation information is used to calculate the linear velocity of the landing point in the navigation coordinate system; based on the linear velocity of the landing point in the navigation coordinate system and the aircraft's absolute speed, the relative velocity of the aircraft with respect to the landing point is calculated. A navigation coordinate system is the coordinate system that an aircraft refers to and uses when calculating its own navigation information (position, velocity, attitude).

[0022] The first relative state estimate is obtained. : .

[0023] In other embodiments, based on the aircraft's raw GNSS / INS observation data, the aircraft's own absolute navigation information, the ship's raw GNSS / INS observation data, and the ship's navigation information, a second relative state estimate is output through deep learning model inference, including: Obtain the pre-calibrated ship landing point arm offset; the ship landing point arm offset refers to the fixed spatial vector from the phase center of the shipborne GNSS antenna to the actual landing point of the aircraft in the ship's body coordinate system. It is fixed information that is transmitted from the ship to the aircraft. The GNSS / INS raw observation data of the aircraft, the absolute navigation information of the aircraft itself, the GNSS / INS raw observation data of the ship, and the navigation information of the ship are synchronized in time to obtain a synchronized time-series observation sequence. The ship landing point arm offset and the synchronized time-series observation sequence are input into a trained deep learning model, and the trained deep learning model outputs a second relative state estimate.

[0024] In addition, this embodiment uses aircraft-ship relative state estimation as a supervision label to construct a replay sample set; Based on the replay sample set, the parameters of the deep learning model are updated using incremental learning or online fine-tuning strategies; wherein, the covariance of the aircraft-ship relative state estimate is only used for model training when it is lower than a preset threshold.

[0025] The deep learning model in this embodiment can perform incremental learning, enabling continuous adaptive optimization and improving long-term accuracy. Furthermore, during training, it uses aircraft-ship relative state estimates below a preset threshold to exclude low-quality data from moments of GNSS lock-up, visual blur, and UWB multipath propagation, preventing the model from learning incorrect "pseudo-labels." The aircraft-ship relative state estimate in this embodiment is output through an extended Kalman filter (EKF), which also outputs the corresponding covariance. The preset threshold is set based on one or more components of the covariance, such as a position accuracy threshold or a velocity accuracy threshold. Optionally, the deep learning model in this embodiment employs neural networks, such as MLP (Multilayer Perceptron), LSTM (Long Short-Term Memory), or RNN (Recurrent Neural Network).

[0026] The step of outputting an aircraft-ship relative state estimate based on the first relative state estimate and the second relative state estimate includes: The first relative state estimate and the second relative state estimate are fused to obtain the main relative state estimate; The relative pose of the aircraft to the ship's landing point is acquired through sensors, and / or the distance measurement of the aircraft relative to the ship's landing point. Based on the main relative state estimate, the relative pose observation and / or the ranging value, the aircraft-ship relative state estimate is output.

[0027] In this embodiment, in obtaining the first relative state estimate and the second relative state estimate of the aircraft relative to the ship landing point, the confidence level of the first relative state estimate is obtained while calculating the second relative state estimate, and the uncertainty level of the second relative state estimate is obtained while calculating the first relative state estimate. In the process of fusing the first relative state estimate and the second relative state estimate to obtain the main relative state estimate, the main relative state estimate is output based on the first relative state estimate, the confidence level, the second relative state estimate, and the uncertainty, according to a preset threshold.

[0028] Confidence score is an indicator reflecting the reliability of the first relative state estimate. It is a comprehensive confidence score calculated based on real-time GNSS / INS solution quality indicators from both the aircraft and ship, such as satellite count, signal-to-noise ratio, differential state, and inertial navigation error. Uncertainty is an indicator reflecting the reliability of the second relative state estimate inferred by the deep learning model. Its calculation is achieved through estimation methods built into the model, such as using a Bayesian neural network to output the prediction variance, or adding an uncertainty estimation branch to the model output layer.

[0029] In some embodiments, the step of outputting the primary relative state estimate based on a preset threshold includes: The preset thresholds include a confidence threshold, an uncertainty threshold, and a consistency threshold; wherein... This is the estimate of the first relative state. For the second relative state estimation, For the estimation of the main relative state, The confidence threshold is... The uncertainty threshold is... The consistency threshold; If the confidence level Greater than the confidence threshold And the uncertainty Less than the uncertainty threshold And the first relative state estimation Compared with the second relative state estimate The difference is less than the consistency threshold Then calculate the residual. Output the main relative state estimate ,in Configurable compensation gain; If the confidence level Greater than the confidence threshold However, the aforementioned uncertainty Greater than or equal to the uncertainty threshold or the first relative state estimate Compared with the second relative state estimate The difference exceeds the consistency threshold Then the first relative state estimate is output as the main relative state estimate. ; If the confidence level Less than or equal to the confidence threshold However, the aforementioned uncertainty Less than the uncertainty threshold The main relative state estimation Using the second relative state estimation Or use the first relative state estimation Compared with the second relative state estimate The weighted average; If the confidence level Less than or equal to the confidence threshold And the uncertainty Greater than or equal to the uncertainty threshold Then the main relative state estimation Invalid; a sensor-based navigation scheme is adopted. The sensor-based navigation scheme described in this embodiment refers to planning the aircraft's motion solely based on relative pose observations measured by the aircraft's visual sensors and / or distance measurements measured by its UWB sensors.

[0030] Right now: when > ,and < ,and and The difference is less than Then calculate the residual. Output the main relative state estimate ,in Configurable compensation gain; when > ,but ≥ or and The difference exceeds Then output the main relative state estimate. ; when ≤ ,but < Main relative state estimation = Or, the first relative state estimation can be used. Compared with the second relative state estimate The weighted average; when ≤ ,and ≥ Then the main relative state estimation Ineffective; a sensor-based navigation scheme based on the aircraft's sensors is adopted.

[0031] In the process of outputting the aircraft-ship relative state estimate based on the master relative state estimate, the relative pose observations, and / or the ranging values, Using the primary relative state estimate as the primary observation, and fusing the relative pose observation and / or the ranging value, an aircraft-ship relative state estimate is output through a filtering algorithm. This embodiment employs an extended Kalman filter as the filtering algorithm.

[0032] Specifically, in outputting the aircraft-ship relative state estimate by using the main relative state estimate as the main observation value and fusing the relative pose observation value and / or the ranging value, the aircraft-ship relative state estimate is: ;include: Construct an extended Kalman filter with relative position and relative velocity as state variables; The extended Kalman filter performs a prediction step based on the previous state. and error covariance Predict the state at the current moment. and error covariance Among them, the state of the previous moment and error covariance , which represents the final state value and error covariance obtained from the previous filtering cycle; The extended Kalman filter performs an update step to estimate the master relative state. As the primary observation, the state at the current moment is used as the primary observation. and error covariance The first update is performed based on the prior, resulting in the first state. , Among them, the observation noise covariance Based on the confidence level The aforementioned uncertainty To dynamically set; if the relative pose observation value is valid, in this embodiment, the relative pose observation value is measured by the visual sensor on the aircraft, that is, when the visual sensor is valid, the relative pose observation value is introduced as the observation value, in the first state. A second update is performed based on the prior, resulting in the second state. , Among them, the observation noise covariance The relative pose observation value is obtained during generation; if the ranging value is valid, in this embodiment, the relative pose observation value is measured by a UWB sensor on the aircraft, that is, when the UWB sensor is valid, the ranging value is introduced as the observation value, in the second state. The third update is performed based on the prior, resulting in the third state. , Among them, the observation noise covariance Obtained during the generation of the distance measurement value; If both the relative pose observation and the ranging value are invalid, then the aircraft-ship relative state estimation... The first state If the relative pose observation is valid and the ranging value is invalid, then the aircraft-ship relative state estimation... The second state If both the relative pose observation and the ranging value are valid, then the aircraft-ship relative state estimation... The third state .

[0033] Further, the step of generating control commands to drive the aircraft based on the aircraft-ship relative state estimation, and driving the aircraft according to the control commands, includes: The landing point of the ship is predicted based on the ship's navigation and positioning data; Based on the motion prediction and the aircraft-ship relative state estimation, control commands to drive the aircraft are generated based on feedforward-feedback control. Drive the aircraft according to the control commands.

[0034] Specifically, the step of predicting the landing point of the ship based on real-time ship motion information, and then basing the motion prediction on the aircraft-ship relative state estimation... Based on feedforward-feedback control, control commands are generated to drive the aircraft, and the aircraft is driven to complete takeoff, approach, hovering tracking, and precise landing according to the control commands, including: The real-time motion information of the ship is obtained by using the ship's navigation information and predicting the ship's landing point based on the navigation information. The predicted trajectory of the ship's landing point within a given time period; Based on the aircraft's flight mission and the estimated relative state between the aircraft and the ship, the desired relative motion trajectory is generated; The predicted trajectory is used to perform feedforward compensation on the desired relative motion trajectory to obtain the motion-compensated desired trajectory. Calculate the error between the current estimated relative state of the aircraft and the expected trajectory after motion compensation, the error including position error and velocity error; The feedback control quantity is calculated based on the position error and the velocity error; The absolute acceleration of the ship's landing point is calculated based on rigid body kinematics using the pre-calibrated arm offset of the ship's landing point and the ship's navigation information. The feedforward control quantity used to counteract the ship's motion acceleration is calculated based on the absolute acceleration at the ship's landing point. The control command is synthesized based on the feedback control quantity and the feedforward control quantity.

[0035] Specifically: the motion prediction of the ship's landing point based on the ship's real-time motion information, and the estimation of the relative state between the aircraft and the ship... Based on feedforward-feedback control, control commands are generated to drive the aircraft, and the aircraft is driven to complete takeoff, approach, hovering tracking, and precise landing according to the control commands, including: The real-time motion information of the ship is obtained by using the ship's navigation information and predicting the ship's landing point based on the navigation information. The predicted trajectory of the ship's landing point within a given time period; Based on the aircraft's flight mission and the aircraft-ship relative state estimation Generate the desired relative motion trajectory; The predicted trajectory is used to perform feedforward compensation on the desired relative motion trajectory to obtain the motion-compensated desired trajectory. Calculate the current aircraft-ship relative state estimate The error between the motion-compensated desired trajectory and the error, the error including position error. Speed ​​error ; Calculate feedback control quantity : ,in, It is a proportional gain matrix. It is the differential gain matrix; Using the pre-calibrated ship landing point pole offset The navigation information of the ship is used to calculate the absolute acceleration of the ship's landing point based on rigid body kinematics. ; Calculate the feedforward control quantity used to counteract the ship's acceleration. : ; Synthesize the control commands : .

[0036] This embodiment uses feedback control. To ensure system stability, eliminate steady-state errors, and exhibit robustness to unmodeled disturbances, feedforward control is employed. This approach can proactively counteract the effects of ship swaying, allowing the aircraft to "follow the deck" rather than passively chase it. Since feedback control only acts when errors occur, relying solely on it results in a slow response, potentially lagging under severe ship motion. Conversely, using only feedforward control depends on model accuracy and cannot handle unpredictable disturbances. This embodiment combines both approaches, actively compensating for known disturbances while correcting residual errors through feedback, achieving high-precision and robust tracking control.

[0037] Example 2 like Figure 2 As shown, this embodiment of the invention provides a method for autonomous takeoff and landing control of shipborne aircraft, including: S201. Obtain the raw GNSS / INS observation data and RTCM differential corrections of the vessel; S202. The navigation information of the ship is calculated based on the raw GNSS / INS observation data of the ship. The navigation information of the ship includes the ship's absolute position, speed, attitude, linear acceleration and angular velocity information. S203. Acquire raw GNSS / INS observation data of the aircraft; S204. Based on the RTCM differential correction and the aircraft's raw GNSS / INS observation data, the aircraft's absolute navigation information is obtained, which includes the aircraft's absolute position and absolute speed. S205. Acquire the relative pose observation value of the visual sensor and the ranging value of UWB. The relative pose observation value of the visual sensor is the relative pose observation value of the aircraft relative to the visual guidance mark on the ship obtained by the visual sensor on the aircraft. The ranging value of UWB is the ranging value of the aircraft relative to the UWB positioning base station on the ship obtained by the UWB sensor on the aircraft. S206. Based on the navigation information of the ship and the absolute navigation information of the aircraft itself, calculate the first relative state estimate and its confidence level through kinematics and coordinate transformation; S207. Based on the aircraft's raw GNSS / INS observation data, the aircraft's own absolute navigation information, the ship's raw GNSS / INS observation data, and the ship's navigation information, a second relative state estimate and its uncertainty are output through deep learning model reasoning. S208. Based on the first relative state estimate, the confidence level, the second relative state estimate, and the uncertainty, output the main relative state estimate according to a preset threshold. S209. Using the main relative state estimate as the main observation value, and fusing the relative pose observation value of the visual sensor and / or the ranging value of the UWB, output the aircraft-ship relative state estimate through a filtering algorithm. S210. Based on the aircraft-ship relative state estimation and the real-time navigation information of the ship, predict the motion of the aircraft's landing point on the ship to obtain the landing point motion prediction result. Based on the landing point motion prediction result and the aircraft's takeoff and landing mission, generate the expected motion trajectory after motion compensation. Based on the expected motion trajectory after compensation and feedforward-feedback control, generate control commands. Drive the aircraft to complete takeoff, approach, hovering tracking and precise landing according to the control commands.

[0038] Example 3 This invention provides an aircraft, comprising: Airborne data link communication module, used to receive data from the ship; Airborne flight control computer, the airborne flight control computer comprising: The relative state acquisition module is used to acquire a first relative state estimate and a second relative state estimate of the aircraft relative to the ship landing point. The first relative state estimate is the relative position and velocity estimate of the aircraft and the ship landing point calculated by kinematics and coordinate transformation. The second relative state estimate is the relative position and velocity estimate of the aircraft and the ship landing point inferred by a deep learning model. The status output module is used to output an aircraft-ship relative status estimate based on the first relative status estimate and the second relative status estimate; The takeoff and landing control module is used to generate control commands to drive the aircraft based on the aircraft-ship relative state estimation, and drive the aircraft according to the control commands.

[0039] Specifically, the aircraft in this embodiment includes: An airborne data link communication module is used to receive data from the ship; the ship's data includes the ship's raw GNSS / INS observation data, the RTCM differential corrections, and the ship's navigation information; The airborne auxiliary guidance module is used to acquire relative pose observations and UWB ranging values ​​through sensors; Airborne flight control computer, the airborne flight control computer comprising: The relative state acquisition module includes a physics calculation unit and a deep learning calculation unit. The physics calculation unit is used to acquire a first relative state estimate of the aircraft relative to the ship landing point. The first relative state estimate is an estimate of the relative position and velocity of the aircraft and the ship landing point calculated through kinematics and coordinate transformation. The deep learning calculation unit is used to acquire a second relative state estimate of the aircraft relative to the ship landing point. The second relative state estimate is an estimate of the relative position and velocity of the aircraft and the ship landing point inferred through a deep learning model. The judgment module is used to fuse the first relative state estimate and the second relative state estimate to obtain the main relative state estimate; The state output module is used to output the aircraft-ship relative state estimate by taking the main relative state estimate as the main observation value and fusing the relative pose observation value and / or the ranging value. The takeoff and landing control module is used to predict the landing point of the ship based on the real-time motion information of the ship, and generate control commands to drive the aircraft based on the motion prediction and the relative state estimation of the aircraft and the ship, and drive the aircraft to complete takeoff, approach, hovering tracking and precise landing according to the control commands.

[0040] Example 4 like Figure 3 As shown, this embodiment of the invention provides an autonomous take-off and landing control system for shipborne aircraft, including a shipborne subsystem and an airborne subsystem; The shipborne subsystem is used to acquire the ship's navigation and positioning data and transmit the ship's navigation and positioning data to the aircraft; the ship's navigation and positioning data includes the ship's raw GNSS / INS observation data and RTCM differential corrections, as well as the ship's navigation information calculated based on the ship's raw GNSS / INS observation data; The airborne subsystem includes: Airborne data link communication module, used to receive data from shipborne subsystems; Airborne flight control computer, including: The relative state acquisition module is used to acquire a first relative state estimate and a second relative state estimate of the aircraft relative to the ship landing point. The first relative state estimate is the relative position and velocity estimate of the aircraft and the ship landing point calculated by kinematics and coordinate transformation. The second relative state estimate is the relative position and velocity estimate of the aircraft and the ship landing point inferred by a deep learning model. The status output module is used to output an aircraft-ship relative status estimate based on the first relative status estimate and the second relative status estimate; The takeoff and landing control module is used to generate control commands to drive the aircraft based on the aircraft-ship relative state estimation, and drive the aircraft according to the control commands.

[0041] Specifically, the shipborne subsystem includes: The shipborne data acquisition module is used to acquire the ship's raw GNSS / INS observation data and RTCM differential corrections; The shipborne integrated navigation module is used to calculate the ship's navigation information based on the ship's raw GNSS / INS observation data. The ship's navigation information includes the ship's absolute position, speed, attitude, and angular velocity information. The shipborne data link communication module is used to transmit the ship's raw GNSS / INS observation data, the RTCM differential corrections, and the ship's navigation information to the aircraft. Shipborne auxiliary guidance module, located in the aircraft landing area on the ship, including visual guidance markers and / or UWB positioning base stations; The airborne subsystem includes: Airborne data acquisition module, used to acquire raw GNSS / INS observation data of the aircraft; The airborne integrated navigation module is used to calculate and obtain the aircraft's own absolute navigation information based on the RTCM differential correction and the aircraft's raw GNSS / INS observation data. The aircraft's own absolute navigation information includes the aircraft's absolute position and the aircraft's absolute speed. Airborne data link communication module, used to receive data from shipborne subsystems; An airborne auxiliary guidance module is used to acquire relative pose observation values ​​from a visual sensor and ranging values ​​from UWB. The relative pose observation values ​​from the visual sensor are obtained by the visual sensor on the aircraft relative to the visual guidance mark on the ship. The ranging values ​​from the UWB are obtained by the UWB sensor on the aircraft relative to the UWB positioning base station on the ship. Airborne flight control computer, including: The relative state acquisition module includes a physics calculation unit and a deep learning calculation unit. The physics calculation unit calculates a first relative state estimate and its confidence level based on the navigation information of the ship and the absolute navigation information of the aircraft through kinematics and coordinate transformation. The deep learning calculation unit outputs a second relative state estimate and its uncertainty based on the raw GNSS / INS observation data of the aircraft, the absolute navigation information of the aircraft, the raw GNSS / INS observation data of the ship, and the navigation information of the ship through deep learning model inference. The judgment module is used to output a main relative state estimate based on the first relative state estimate, the confidence level, the second relative state estimate, and the uncertainty, according to a preset threshold. The state output module is used to take the main relative state estimate as the main observation value, and fuse the relative pose observation value of the visual sensor and / or the ranging value of the UWB, and output the aircraft-ship relative state estimate through a filtering algorithm. The takeoff and landing control module is used to predict the motion of the aircraft's landing point on the ship based on the relative state estimation of the aircraft and the real-time navigation information of the ship, to obtain the landing point motion prediction result, to generate the expected motion trajectory after motion compensation based on the landing point motion prediction result and the aircraft's takeoff and landing mission, to generate control commands based on the expected motion trajectory after compensation and feedforward-feedback control, and to drive the aircraft to complete takeoff, approach, hovering tracking and precise landing according to the control commands.

[0042] Example 5 like Figure 4 As shown, this embodiment of the invention also provides an electronic device, which includes: a memory 406, a processor 405, and a computer program stored in the memory 406 and executable on the processor 405. The computer program is configured to implement the steps of the above-described autonomous take-off and landing control method for shipborne aircraft and can achieve the same technical effect. To avoid repetition, it will not be described again here.

[0043] For details, see Figure 4 The present invention also provides an electronic device, including a bus 401, a transceiver 402, an antenna 403, a bus interface 404, a processor 405, and a memory 406.

[0044] The transceiver 402 is used to acquire unstructured data, which includes at least one of data obtained based on user input information and data obtained based on configuration file scanning. The processor 405 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 405 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, digital signal processors (DSPs), and any suitable processor, controller, microcontroller, etc. Processor 405 performs the various methods and processes described above, such as autonomous take-off and landing control methods for shipborne aircraft.

[0045] exist Figure 4In this context, a bus architecture (represented by bus 401) is used. Bus 401 can include any number of interconnected buses and bridges, linking various circuits including one or more processors represented by processor 405 and memory represented by memory 406. Bus 401 can also link various other circuits such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art and therefore will not be described further herein. Bus interface 404 provides an interface between bus 401 and transceiver 402. Transceiver 402 can be a single element or multiple elements, such as multiple receivers and transmitters, providing a unit for communicating with various other devices over a transmission medium. Data processed by processor 405 is transmitted over a wireless medium via antenna 403, which further receives data and transmits data to processor 405.

[0046] Processor 405 is responsible for managing bus 401 and general processing, and can also provide various functions, including timing, peripheral interface, voltage regulation, power management, and other control functions. Memory 406 can be used to store data used by processor 405 during operation.

[0047] Optionally, the processor 405 may be a central processing unit (CPU), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or a complex programmable logic device (CPLD).

[0048] Example 6 This invention also provides a storage medium, which is a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it implements the steps of the above-described autonomous takeoff and landing control method for shipborne aircraft and achieves the same technical effect. To avoid repetition, it will not be described again here. The computer-readable storage medium may be a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk, etc.

[0049] Example 7 This invention also provides a computer program product, which includes a computer program. When the computer program is executed by a processor, it implements the steps of the above-described autonomous take-off and landing control method for shipborne aircraft, achieving the same technical effect. To avoid repetition, it will not be described again here.

[0050] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element. Furthermore, it should be noted that the scope of the methods and apparatuses in the embodiments of the present invention is not limited to performing functions in the order discussed, but may also include performing functions substantially simultaneously or in the reverse order, depending on the functions involved. For example, the described methods may be performed in a different order than described, and various steps may be added, omitted, or combined. Additionally, features described with reference to certain examples may be combined in other examples.

[0051] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal (which may be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0052] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of the present invention without departing from the spirit and scope of the claims, and all of these forms are within the protection scope of the present invention.

Claims

1. A method for autonomous takeoff and landing control of shipborne aircraft, characterized in that, include: The first relative state estimate and the second relative state estimate of the aircraft relative to the ship landing point are obtained. The first relative state estimate is the relative position and velocity estimate of the aircraft and the ship landing point calculated by kinematics and coordinate transformation. The second relative state estimate is the relative position and velocity estimate of the aircraft and the ship landing point inferred by a deep learning model. Based on the first relative state estimate and the second relative state estimate, the aircraft-ship relative state estimate is output. Control commands for driving the aircraft are generated based on the aircraft-ship relative state estimation, and the aircraft is driven according to the control commands.

2. The autonomous takeoff and landing control method for shipborne aircraft according to claim 1, characterized in that, Before obtaining the first and second relative state estimates of the aircraft relative to the ship's landing point, the process includes: Acquire raw GNSS / INS observation data and RTCM differential corrections for the ship; The navigation information of the vessel is calculated based on the raw GNSS / INS observation data of the vessel. Acquire raw GNSS / INS observation data of the aircraft; The absolute navigation information of the aircraft itself is obtained by calculating the RTCM differential correction and the raw GNSS / INS observation data of the aircraft. Based on the navigation information of the ship and the absolute navigation information of the aircraft itself, the first relative state estimate is calculated through kinematics and coordinate transformation; Based on the aircraft's raw GNSS / INS observation data, the aircraft's own absolute navigation information, the ship's raw GNSS / INS observation data, and the ship's navigation information, a second relative state estimate is output through deep learning model inference.

3. The method according to claim 2, characterized in that, The calculation of the first relative state estimate based on the ship's navigation information and the aircraft's own absolute navigation information, through kinematics and coordinate transformation, includes: Obtain the pre-calibrated pole offset at the ship's landing point; Calculate the attitude rotation matrix from the ship coordinate system to the navigation coordinate system based on the ship's navigation information; Based on the offset of the landing point arm, the attitude rotation matrix, and the navigation information of the ship, calculate the absolute position of the ship's landing point in the navigation coordinate system; The relative position of the aircraft to the ship's landing point is calculated based on the aircraft's own absolute navigation information and the absolute position of the ship's landing point in the navigation coordinate system. The relative velocity of the aircraft to the landing point of the ship is calculated based on the aircraft's own absolute navigation information, the ship's navigation information, the attitude rotation matrix, and the ship's landing point arm offset. The first relative state estimate is obtained based on the relative position of the aircraft with respect to the ship's landing point and the relative speed of the aircraft with respect to the ship's landing point.

4. The method according to claim 2, characterized in that, Based on the aircraft's raw GNSS / INS observation data, the aircraft's own absolute navigation information, the ship's raw GNSS / INS observation data, and the ship's navigation information, a second relative state estimate is output through deep learning model inference, including: Obtain the pre-calibrated pole offset at the ship's landing point; The GNSS / INS raw observation data of the aircraft, the absolute navigation information of the aircraft itself, the GNSS / INS raw observation data of the ship, and the navigation information of the ship are synchronized in time to obtain a synchronized time-series observation sequence. The ship landing point arm offset and the synchronized time-series observation sequence are input into a trained deep learning model, and the trained deep learning model outputs a second relative state estimate.

5. The method according to claim 1, characterized in that, The step of outputting an aircraft-ship relative state estimate based on the first relative state estimate and the second relative state estimate includes: The first relative state estimate and the second relative state estimate are fused to obtain the main relative state estimate; The relative pose of the aircraft to the ship's landing point is acquired through sensors, and / or the distance measurement of the aircraft relative to the ship's landing point. Based on the main relative state estimate, the relative pose observation and / or the ranging value, the aircraft-ship relative state estimate is output.

6. The method according to claim 5, characterized in that, The process of fusing the first relative state estimate and the second relative state estimate to obtain the main relative state estimate includes: In obtaining the first relative state estimate and the second relative state estimate of the aircraft relative to the ship landing point, the confidence level of the first relative state estimate is obtained while calculating the uncertainty of the second relative state estimate. In the process of fusing the first relative state estimate and the second relative state estimate to obtain the main relative state estimate, the main relative state estimate is output based on the first relative state estimate, the confidence level, the second relative state estimate, and the uncertainty, according to a preset threshold.

7. The method according to claim 5, characterized in that, In the process of outputting the aircraft-ship relative state estimate based on the master relative state estimate, the relative pose observations, and / or the ranging values, The aircraft-ship relative state estimate is output through a filtering algorithm, using the main relative state estimate as the main observation value and fusing the relative pose observation value and / or the ranging value.

8. The method according to claim 2, characterized in that, The step of generating control commands to drive the aircraft based on the aircraft-ship relative state estimation, and driving the aircraft according to the control commands, includes: The landing point of the ship is predicted based on the ship's navigation and positioning data; Based on the motion prediction and the aircraft-ship relative state estimation, control commands to drive the aircraft are generated based on feedforward-feedback control. Drive the aircraft according to the control commands.

9. An aircraft, characterized in that, include: Airborne data link communication module, used to receive data from the ship; Airborne flight control computer, the airborne flight control computer comprising: The relative state acquisition module is used to acquire a first relative state estimate and a second relative state estimate of the aircraft relative to the ship landing point. The first relative state estimate is the relative position and velocity estimate of the aircraft and the ship landing point calculated by kinematics and coordinate transformation. The second relative state estimate is the relative position and velocity estimate of the aircraft and the ship landing point inferred by a deep learning model. The status output module is used to output an aircraft-ship relative status estimate based on the first relative status estimate and the second relative status estimate; The takeoff and landing control module is used to generate control commands to drive the aircraft based on the aircraft-ship relative state estimation, and drive the aircraft according to the control commands.

10. An autonomous take-off and landing control system for shipborne aircraft, characterized in that, Including shipborne subsystems and airborne subsystems; The shipborne subsystem is used to acquire the ship's navigation and positioning data and transmit the ship's navigation and positioning data to the aircraft; the ship's navigation and positioning data includes the ship's raw GNSS / INS observation data and RTCM differential corrections, as well as the ship's navigation information calculated based on the ship's raw GNSS / INS observation data; The airborne subsystem includes: Airborne data link communication module, used to receive data from shipborne subsystems; Airborne flight control computer, including: The relative state acquisition module is used to acquire a first relative state estimate and a second relative state estimate of the aircraft relative to the ship landing point. The first relative state estimate is the relative position and velocity estimate of the aircraft and the ship landing point calculated by kinematics and coordinate transformation. The second relative state estimate is the relative position and velocity estimate of the aircraft and the ship landing point inferred by a deep learning model. The status output module is used to output an aircraft-ship relative status estimate based on the first relative status estimate and the second relative status estimate; The takeoff and landing control module is used to generate control commands to drive the aircraft based on the aircraft-ship relative state estimation, and drive the aircraft according to the control commands.

11. An electronic device, characterized in that, The device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the autonomous take-off and landing control method for a shipborne aircraft as described in any one of claims 1 to 8.

12. A storage medium, characterized in that, The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, it implements the steps of the autonomous take-off and landing control method for shipborne aircraft as described in any one of claims 1 to 8.

13. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, implements the steps of the autonomous take-off and landing control method for shipborne aircraft as described in any one of claims 1 to 8.

Images