Unmanned aerial vehicle inland river mobile platform autonomous landing method based on target motion prediction

Abstract

The invention relates to the technical field of unmanned aerial vehicle autonomous control, and provides an unmanned aerial vehicle inland river mobile platform autonomous landing method based on target motion prediction, and the method comprises the steps: obtaining the real-time motion state information of an unmanned aerial vehicle and an inland river mobile platform through multi-sensor fusion; and a double-layer estimation structure composed of a rough estimation layer and an accurate estimation layer is constructed. On the basis, platform short-time motion trend prediction data is generated according to aperiodic motion characteristics generated by the inland river mobile platform under the action of flow velocity change and backflow disturbance, and dynamic landing trajectory planning and closed-loop control are carried out in combination with kinetic constraints of the unmanned aerial vehicle; safe and stable autonomous landing of the unmanned aerial vehicle under the condition of continuous movement of the inland river mobile platform is realized, so that the stability and success rate of autonomous landing are improved.

Description

Technical Field

[0001] The embodiments of this application relate to the field of unmanned aerial vehicle (UAV) autonomous control technology, and in particular to an autonomous landing method for an unmanned aerial vehicle (UAV) inland waterway mobile platform based on target motion prediction. Background Technology

[0002] With the development of autonomous control and intelligent navigation technologies for unmanned aerial vehicles (UAVs), their use in inland waterway inspection, emergency rescue, material delivery, and hydrological monitoring is becoming increasingly widespread. In these applications, UAVs typically need to autonomously land on mobile vessels navigating inland waterways to achieve mission recovery, energy resupply, or payload exchange. However, compared to stationary land platforms, the motion characteristics of mobile inland waterway platforms are complex and irregular, posing significant challenges to autonomous UAV landing.

[0003] In navigation environments exemplified by inland river basins (such as the Mekong River), ships typically navigate at low speeds. Their motion is influenced by a combination of factors, including changes in river flow velocity, backflow effects, shoreline constraints, and ship maneuvering behavior, exhibiting six-degree-of-freedom motion characteristics dominated by lateral drift, yaw, and non-periodic attitude changes. This type of motion is characterized by small amplitude, low frequency, but long duration and high randomness, making it difficult to accurately describe using traditional steady-state or quasi-static models.

[0004] Existing autonomous landing technologies for unmanned aerial vehicles (UAVs) are mostly based on the assumption of a static or regularly moving platform, typically treating the landing target as a fixed reference point with slow or negligible changes in position and attitude. Under this assumption, UAVs mainly rely on visual sensors, inertial measurement units (IMUs), and satellite positioning systems (GNSS) for relative attitude perception and control. However, in inland waterway environments, the motion of ship platforms exhibits significant non-periodicity and unpredictability, making the above assumptions difficult to hold. This leads to problems such as accumulated alignment errors, control lag, and unstable terminal contact during UAV landing.

[0005] Furthermore, in inland waterway environments commonly present engineering challenges such as GNSS signal obstruction, significant multipath effects, and visual feature degradation. For example, riverside buildings, vegetation, and bridge structures can cause intermittent GNSS signal loss; water surface reflection, fog, and changes in lighting can affect the stable output of visual sensors. These factors make it difficult for a single sensor or simple fusion method to continuously provide highly reliable relative pose information throughout the entire descent process.

[0006] To address the landing challenges of unmanned aerial vehicles (UAVs) on mobile inland waterways, existing research has proposed methods based on visual guidance, multi-sensor fusion, and model predictive control to compensate for the impact of platform motion. However, most current methods focus on estimating the platform's current state and lack effective prediction of its short-term future motion trends, failing to pre-correct the UAV's landing trajectory. In the scenario of low-speed mobile inland waterways, this "passive following" control strategy struggles to cope with persistent lateral drift and yaw disturbances, resulting in a low success rate in capturing the landing window.

[0007] Therefore, there is an urgent need for an autonomous landing technology for UAVs in inland waterway mobile platform environments. This technology should be able to predict the short-term motion trend of the target platform under conditions of unstable multi-sensor information, combined with the platform's motion characteristics, and deeply couple the prediction results with the UAV trajectory planning and control process, thereby achieving safe, stable, and autonomous landing of UAVs on inland waterway mobile platforms. Summary of the Invention

[0008] In view of this, embodiments of this application propose an autonomous landing method for UAVs on inland waterways based on target motion prediction. This method aims to address the characteristics of inland waterway navigation environments, such as low-speed navigation, significant lateral drift, and non-periodic motion of mobile platforms. It integrates multi-source perception, motion prediction, dynamic trajectory planning, and closed-loop control to construct a closed-loop control system for autonomous landing of UAVs on inland waterways, thereby improving the accuracy and stability of autonomous landing in complex inland waterway environments. Specifically, by introducing target motion prediction results, the relative pose estimation, trajectory planning, and control processes are uniformly constrained and guided, enabling the UAV to shift from passively following the mobile platform to prediction-based active compensation landing control.

[0009] To achieve the above objectives, embodiments of this application propose an autonomous landing method for an unmanned aerial vehicle (UAV) mobile platform on an inland waterway based on target motion prediction. The method includes: acquiring real-time motion state information of the UAV and the inland waterway mobile platform through multi-sensor fusion, wherein the multi-sensors include at least a Global Navigation Satellite System (GNSS), an Inertial Measurement Unit (IMU), and a visual sensor; establishing a two-layer estimation structure based on the real-time motion state information, consisting of a coarse estimation layer and a precise estimation layer. The coarse estimation layer uses IMU data and GNSS data to make a preliminary estimate of the relative position, relative attitude, and motion trend between the UAV and the inland waterway mobile platform. The precise estimation layer fuses visual sensor data and IMU data when visual conditions meet preset requirements. The initial estimation results are constrained and corrected to obtain high-precision relative pose estimation results. Based on the estimation results of the two-layer estimation structure, the dynamic characteristics of the inland waterway mobile platform are analyzed to address the non-periodic motion characteristics generated by the platform under the influence of external environmental disturbances. Motion trend prediction data of the inland waterway mobile platform within a preset short time window is generated. Based on the motion trend prediction data and the dynamic constraints of the UAV itself, dynamic landing trajectory planning is performed to generate a landing path that matches the motion trend of the inland waterway mobile platform in real time. Based on the motion trend prediction data and the landing path, the attitude and speed of the UAV are controlled in a closed loop to enable the UAV to complete a safe and stable autonomous landing while the inland waterway mobile platform is in continuous motion.

[0010] To achieve the above objectives, embodiments of this application also propose an autonomous landing system for an unmanned aerial vehicle (UAV) inland waterway mobile platform based on target motion prediction. This system is used to execute the aforementioned autonomous landing method for an UAV inland waterway mobile platform based on target motion prediction. The system includes: a data acquisition module for acquiring real-time motion state information of the UAV and the inland waterway mobile platform based on multi-sensor fusion, wherein the multi-sensors include at least a Global Navigation Satellite System (GNSS), an Inertial Measurement Unit (IMU), and a visual sensor; a dual-layer estimation module for constructing a dual-layer estimation structure consisting of a coarse estimation layer and a precise estimation layer based on real-time motion state information. The coarse estimation layer is used to obtain a preliminary estimate of the relative pose between the UAV and the inland waterway mobile platform based on GNSS data and IMU data when visual information is unavailable or of reduced quality. The precise estimation layer is used to correct the preliminary relative pose estimate based on visual observation information when visual information meets the requirements; and a short-time prediction module for... Based on the estimation results of the dual-layer estimation structure, the motion characteristics of the inland waterway mobile platform under external disturbances such as water flow and wind are analyzed, generating short-term motion trend prediction data for the inland waterway mobile platform. The path planning module is used to dynamically plan the landing trajectory of the UAV based on the short-term motion trend prediction data and the dynamic constraints of the UAV itself, generating a landing trajectory that matches the motion trend of the inland waterway mobile platform. The landing control module is used to perform closed-loop control of the UAV's attitude and speed during the UAV's movement along the landing trajectory, based on the real-time motion status of the inland waterway mobile platform and the aforementioned short-term motion trend prediction data, so as to achieve safe, stable, and autonomous landing of the UAV under the continuous motion conditions of the inland waterway mobile platform.

[0011] Optionally, coarse estimation rapidly estimates the relative pose between the UAV and the inland waterway mobile platform by fusing IMU and GNSS data. This is used to maintain the continuity of positioning results and control input when the visual sensor output is unavailable or the data quality is below a preset threshold. Precise estimation is initiated when the visual sensor output data quality meets preset conditions. Based on the multi-view feature information acquired by the visual sensor, it constrains and corrects the results of the coarse estimation layer to improve the accuracy of the relative pose estimation between the UAV and the inland waterway mobile platform. The dual-layer estimation structure switches between coarse and precise estimation according to the data quality of the visual sensor to improve the reliability of relative pose estimation while ensuring system continuity and safety.

[0012] Optionally, when the data quality output by the visual sensor meets a preset quality threshold, a precise estimation process is initiated. Based on the multi-view feature information acquired by the visual sensor, the relative pose between the UAV and the inland waterway mobile platform is optimized and estimated. The precise estimation corrects the relative pose result obtained by the coarse estimation by constructing visual observation constraints. When the visual observation meets a preset consistency condition, the optimized relative pose estimation result is output. When the data quality output by the visual sensor is lower than the preset quality threshold, the coarse estimation process is maintained, and the relative pose is updated using an approximate solution method based on local visual constraints to ensure the continuity of the relative pose estimation result.

[0013] Optionally, based on the estimation results of the two-layer estimation structure, the historical motion state of the inland waterway mobile platform is modeled to generate motion trend prediction data of the inland waterway mobile platform within a preset time window, taking into account the non-stationary motion characteristics generated by the inland waterway mobile platform under the influence of external environmental disturbances. The motion trend prediction outputs the future pose change trend of the inland waterway mobile platform by performing time series analysis on the historical position and attitude state sequence of the inland waterway mobile platform, which is used to provide a forward-looking motion reference for the autonomous landing process of the UAV. When computing resources are limited or real-time requirements are high, the motion trend prediction adopts an approximate prediction model with low computational complexity to ensure the real-time performance of the prediction process and the stability of the system operation.

[0014] Optionally, based on the motion trend prediction data of the inland waterway mobile platform and the dynamic constraints of the UAV itself, dynamic trajectory planning is performed on the landing process of the UAV to generate a landing trajectory that matches the motion trend of the inland waterway mobile platform in real time. The dynamic trajectory planning comprehensively considers the position, speed, attitude and control input constraints of the UAV during the planning process to ensure that the generated landing trajectory meets the flight safety requirements. During the movement of the UAV along the landing trajectory, the attitude and speed of the UAV are controlled in a closed loop according to the real-time motion status and motion trend prediction data of the inland waterway mobile platform, so that the UAV can complete a safe and stable autonomous landing under the condition of continuous movement of the inland waterway mobile platform.

[0015] Optionally, during the dynamic trajectory planning and closed-loop control process, flight constraints are introduced to ensure the physical feasibility and safety of the landing process. These flight constraints include at least: UAV attitude angle constraints, which limit the roll and pitch angle ranges of the UAV during landing; descent speed constraints, which limit the descent speed of the UAV when approaching the inland waterway mobile platform to avoid deck impact; and relative attitude coordination constraints, which ensure that the predicted position and attitude of the UAV enter the captureable area of ​​the inland waterway mobile platform during the terminal phase of landing. Simultaneously, during the execution of the landing trajectory, a disturbance compensation mechanism is introduced to suppress external disturbances, thereby enhancing the landing robustness of the UAV in complex water surface environments and improving the success rate and safety of autonomous landing.

[0016] Optionally, during dynamic trajectory planning and closed-loop control, a degradation substitution mechanism is set to reduce computational load and ensure continuous system operation. When the motion state of the inland waterway mobile platform changes significantly or the uncertainty of motion trend prediction increases, the prediction time domain of trajectory planning is shortened to improve response speed. When system computing resources are limited or real-time requirements increase, the trajectory planning and control process is switched to a lower computational complexity substitution strategy to maintain basic control over the UAV. When the quality of sensor information deteriorates or some sensors become unavailable, a degradation alignment strategy based on coarse estimation results is enabled to ensure the continuity of relative pose estimation between the UAV and the inland waterway mobile platform. Through the degradation substitution mechanism, the UAV can maintain controlled operation in complex water environments and under multi-source uncertainties, thereby improving the safety and reliability of the autonomous landing process.

[0017] The autonomous landing method for UAVs on inland waterway mobile platforms based on target motion prediction proposed in this application has at least the following advantages compared with traditional UAV autonomous landing technology for ship mobile platforms.

[0018] First, this application addresses the motion characteristics of inland waterway mobile platforms, which are characterized by low-speed navigation and significant lateral drift. Through multi-source sensing fusion and a two-layer estimation structure design, it can maintain the continuity and stability of relative pose estimation between the UAV and the mobile platform even when GNSS signals are unstable or visual conditions are degraded, thereby improving the engineering robustness of the system in inland waterway environments.

[0019] Second, this application predicts the short-term motion trend of the inland waterway mobile platform and uses the prediction results as a constraint for trajectory planning and control, enabling the UAV to change from passively following the platform's motion to actively compensating for the platform's motion, effectively reducing the impact of lateral drift and yaw disturbances on landing accuracy.

[0020] Third, this application adopts a combination of dynamic trajectory planning and closed-loop control to adjust the approach path and attitude control strategy of the UAV in real time during the landing process, so that the UAV can maintain a stable approach channel under the condition of continuous movement of the inland waterway mobile platform, thereby improving the success rate and safety of autonomous landing.

[0021] Fourth, the method of this application is applicable to a variety of inland waterway vessels and similar low-speed mobile platforms, and has good scenario adaptability and engineering application value. Attached Figure Description

[0022] To more clearly illustrate the technical solutions in the embodiments or related technologies of this application, the accompanying drawings used in the description of the embodiments or related technologies of this application will be briefly introduced below. Obviously, the following drawings are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. The drawings described herein are only used to explain this application and are not intended to limit this application.

[0023] Figure 1 This is a flowchart of an autonomous landing method for an unmanned aerial vehicle (UAV) inland waterway mobile platform based on target motion prediction, provided in one embodiment of this application; Figure 2 This is a schematic diagram of the technical route of an autonomous landing method for an unmanned aerial vehicle (UAV) inland waterway mobile platform based on target motion prediction provided in one embodiment of this application; Figure 3 This is a schematic diagram of the technical concept of an autonomous landing method for an unmanned aerial vehicle (UAV) inland waterway mobile platform based on target motion prediction, provided in one embodiment of this application. Figure 4 This is a schematic diagram of a research framework for an autonomous landing method for an unmanned aerial vehicle (UAV) inland waterway mobile platform based on target motion prediction, provided in one embodiment of this application. Figure 5 This is a landing identifier provided in one example of an application for an autonomous landing system for an unmanned aerial vehicle (UAV) inland waterway mobile platform based on target motion prediction. Figure 6 This is a schematic diagram of the structure of an autonomous landing system for an unmanned aerial vehicle (UAV) inland waterway mobile platform based on target motion prediction, provided in another embodiment of this application; Figure 7 This is a schematic diagram of the structure of an electronic device provided in another embodiment of this application. Detailed Implementation

[0024] To make the objectives, technical solutions, and beneficial effects of the embodiments of this application clearer, the specific implementation methods of this application will be described in detail below with reference to the accompanying drawings. Those skilled in the art will understand that equivalent substitutions or modifications made to the following embodiments without departing from the concept of the technical solution of this application should fall within the protection scope of this application.

[0025] This application describes the autonomous landing of unmanned aerial vehicles (UAVs) using a mobile platform in an inland waterway navigation environment. Compared to ocean-going vessels, inland waterway mobile platforms typically exhibit characteristics such as low speed, significant lateral drift, non-periodic motion, and significant susceptibility to river flow velocity and backflow disturbances. Their motion is difficult to describe using simple uniform or periodic models. In such environments, if traditional landing methods that treat the landing platform as a quasi-static target are still employed, it can easily lead to significant lateral deviations and attitude instability during the approach and landing process, thereby affecting landing safety and success rate.

[0026] In this embodiment, when the UAV performs an autonomous landing mission on an inland waterway mobile platform, it first acquires motion state information of itself and the target platform through an onboard multi-sensor system. The multi-sensor system includes at least a Global Navigation Satellite System (GNSS), an Inertial Measurement Unit (IMU), and a visual sensor. The GNSS provides global position information for the UAV and the mobile platform; the IMU acquires the UAV's attitude angular velocity and linear acceleration information; and the visual sensor is an onboard visual sensor mounted on the UAV's fuselage, used for visual perception of the inland waterway mobile platform's deck area and landing markings. The inland waterway mobile platform itself does not have an active visual perception unit.

[0027] Based on the aforementioned multi-source sensing information, this embodiment constructs a two-layer estimation structure to achieve relative pose estimation between the UAV and the inland waterway mobile platform. Specifically, the coarse estimation layer uses GNSS data and IMU data to make a preliminary estimate of the relative position, relative attitude, and motion trend between the UAV and the inland waterway mobile platform, thereby maintaining the continuity of relative state estimation even when GNSS signal quality fluctuates or visual conditions are temporarily unavailable.

[0028] When preset visual conditions are met, the precise estimation layer further incorporates visual sensor data and fuses it with IMU data to constrain and correct the results of the coarse estimation layer, thereby obtaining a high-precision relative pose estimation result. Through the above-mentioned two-layer estimation mechanism, this embodiment can effectively address the perception degradation problem caused by factors such as GNSS instability, illumination changes, and local occlusion in inland waterway environments.

[0029] After obtaining stable and reliable relative pose estimation results, this embodiment further analyzes the short-term motion trend of the inland waterway mobile platform, which is affected by changes in river flow velocity, backflow disturbances, and maneuvering behavior, to predict its motion characteristics. Specifically, based on historical estimated state sequences, a target motion prediction model is constructed to generate motion trend prediction data for the inland waterway mobile platform within a preset time window, which describes the platform's displacement changes and attitude evolution trends in the near future.

[0030] Based on the predicted motion trend data, this embodiment designs a dynamic landing trajectory planning algorithm. The predicted results serve as forward-looking constraints for trajectory planning, combined with the UAV's own dynamic constraints and safety limitations, to generate an optimal landing trajectory in real time that matches the motion trend of the inland waterway mobile platform. By introducing target motion prediction information, the UAV no longer passively follows the platform's movement during landing, but can compensate for the platform's lateral drift and attitude changes in advance. During the landing control phase, this embodiment performs closed-loop control of the UAV's attitude and speed based on the predicted data and planned trajectory. By introducing the prediction results into the control loop, closed-loop coupling between trajectory planning and attitude control is achieved, enabling the UAV to maintain a stable approach to the channel and complete a safe and smooth autonomous landing even under the continuous movement of the inland waterway mobile platform.

[0031] One embodiment of this application proposes an autonomous landing method for an unmanned aerial vehicle (UAV) mobile platform in inland waterways based on target motion prediction. The method flow and implementation details of this embodiment are described below with reference to the accompanying drawings. It should be noted that the following content is only used to explain the technical solution of this embodiment and does not constitute a limitation on the scope of protection of this application. Various equivalent modifications or improvements can be made by those skilled in the art without departing from the technical concept of this application.

[0032] The overall process of the autonomous landing method for UAV inland waterway mobile platforms based on target motion prediction proposed in this embodiment is as follows: Figure 1 As shown. To facilitate understanding of the technical concept and implementation path of this embodiment, before explaining the specific steps, we will first combine... Figure 2 , Figure 3 , Figure 4 and Figure 5 This paper explains the technical approach, technical concept, and overall research framework of this method.

[0033] See Figure 2This embodiment addresses the application scenario of mobile platforms in inland waterway navigation environments. It constructs a motion state perception and prediction system for inland waterway mobile platforms by fusing multi-source sensing information from Global Navigation Satellite System (GNSS), Inertial Measurement Unit (IMU), and visual sensors. Based on the multi-sensor fusion results, a relative motion model between the inland waterway mobile platform and the UAV is established, enabling real-time estimation and short-term prediction of the platform's displacement and attitude changes. Furthermore, through relative pose calculation and dynamic trajectory planning algorithms, the UAV's landing path is updated in real-time, allowing the UAV to actively compensate for relative pose changes caused by low-speed navigation, lateral drift, and non-periodic motion conditions of the inland waterway mobile platform. Finally, closed-loop control enables the UAV to achieve autonomous and stable landing during the continuous movement of the inland waterway mobile platform.

[0034] See Figure 3 The core technical concept of this embodiment lies in improving the accuracy and stability of autonomous landing of UAVs in the complex dynamic environment of inland waterways through a collaborative mechanism of "mobile platform motion prediction and autonomous landing control". First, at the perception layer, real-time motion state information of the UAV and the inland waterway mobile platform is acquired based on multi-sensor fusion, and a two-layer estimation structure is constructed on this basis. The coarse estimation layer, based on GNSS and IMU data, makes a preliminary estimate of the relative position, relative attitude, and motion trend between the UAV and the inland waterway mobile platform to ensure that the system still has continuous and stable state perception capability even under limited visual conditions or strong external interference. The precise estimation layer, when visual conditions meet preset requirements, introduces visual feature constraints and inertial attitude compensation information to further correct the coarse estimation results, thereby achieving high-precision estimation of the six-degree-of-freedom nonlinear relative motion state between the UAV and the inland waterway mobile platform.

[0035] After obtaining highly reliable relative motion state estimation results, this embodiment further constructs a motion prediction and trajectory planning module for the inland waterway mobile platform. By analyzing the dynamic characteristics of the inland waterway mobile platform under the influence of factors such as river flow velocity changes, backflow disturbances, and maneuvering behavior, and combining historical motion state data, a motion trend prediction result for the inland waterway mobile platform within a preset time window is generated. Subsequently, based on the prediction results and the dynamic constraints of the UAV itself, a dynamic trajectory planning algorithm is designed to generate an optimal landing path that matches the motion trend of the inland waterway mobile platform in real time. This path comprehensively considers spatial position deviations, platform attitude changes, and external disturbance factors during the planning process, realizing dynamic correction and adaptive updating of the landing trajectory.

[0036] Finally, the control execution module performs closed-loop control of the UAV's attitude and speed based on the prediction results and the planned path, enabling the UAV to complete a safe and stable autonomous landing while the inland waterway mobile platform continues to move. Through the collaborative design of perception, prediction, planning, and control described above, this embodiment constructs a closed-loop control system for autonomous landing of UAVs on inland waterway mobile platforms, significantly improving the system's real-time performance, robustness, and landing accuracy in complex inland waterway environments.

[0037] See Figure 4 The overall framework of this embodiment adopts a hierarchical structure of "perception-prediction-planning-control," forming a complete autonomous landing technology system for unmanned aerial vehicles (UAVs). Specifically, the perception layer is responsible for multi-source information acquisition and two-layer estimation; the prediction layer is responsible for predicting the movement trend of the inland waterway mobile platform; the planning layer is responsible for dynamic landing path generation; and the control layer is responsible for closed-loop control of attitude and speed. These layers form a closed-loop linkage through information feedback and state updates, thereby enabling the UAV to land autonomously and safely on the inland waterway mobile platform.

[0038] See Figure 5 This illustration shows the nested landing marker structure on the deck area of ​​the inland waterway mobile platform in this embodiment, as well as the visual tracking of the UAV at different distance stages. The outer marker is used for mid-to-long-range target detection, the inner marker is used for precise alignment at close range, and the lateral reference line is used to assist in estimating the heading of the inland waterway mobile platform.

[0039] The following is a detailed description of the specific process of the autonomous landing method for an unmanned aerial vehicle (UAV) inland waterway mobile platform based on target motion prediction proposed in this embodiment, including: Step 11: Acquire real-time motion status information of the UAV and the core mobile platform based on multi-sensor fusion. In this embodiment, the first step in the autonomous landing of the UAV on the inland waterway mobile platform is to acquire real-time motion status information of the UAV and the inland waterway mobile platform based on multi-sensor fusion. The multi-sensor includes at least an airborne Global Navigation Satellite System (GNSS), an Inertial Measurement Unit (IMU), and a visual sensor. The GNSS is used to provide the absolute position information of the UAV and the inland waterway mobile platform in the global coordinate system, the IMU is used to provide inertial measurement information such as angular velocity and linear acceleration of the UAV and the inland waterway mobile platform, and the visual sensor is used to acquire image information of the deck area of ​​the inland waterway mobile platform and extract visual features.

[0040] In this embodiment, the deck area of ​​the inland waterway mobile platform is equipped with visual landing markers for autonomous landing of unmanned aerial vehicles (UAVs). The landing markers adopt a nested structure design, including an outer guidance marker and an inner precision positioning marker. The outer guidance marker is used for initial target detection and coarse alignment of the UAV at medium to long distances, while the inner precision positioning marker is used for precise pose estimation and terminal guidance of the UAV at close ranges.

[0041] The landing markers include a lateral reference line parallel to the heading of the inland waterway mobile platform. This lateral reference line serves as a visual constraint feature in the visual image for estimating the platform's heading, assisting the UAV in visually estimating the platform's motion direction during landing. As the relative distance between the UAV and the inland waterway mobile platform decreases, the visual system automatically switches from the outer guidance marker to the inner precise positioning marker for tracking, thereby improving the accuracy and stability of relative pose estimation during close-range phases.

[0042] It should be noted that in inland waterway navigation environments, factors such as river bends, shoreline obstruction, bridge structures, weather conditions, and water surface reflection mean that visual sensors do not remain consistently available throughout the landing process. Therefore, this embodiment does not rely on a single sensor in its system design. Instead, it acquires the real-time motion status of the UAV and the inland waterway mobile platform through multi-source information fusion, thereby providing a continuous and reliable data foundation for subsequent estimation, prediction, planning, and control.

[0043] Step 12: Establish a two-layer estimation structure consisting of a coarse estimation layer and a precise estimation layer. After acquiring the real-time motion status information of the UAV and the inland waterway mobile platform, this embodiment establishes a two-layer estimation structure consisting of a coarse estimation layer and a precise estimation layer based on the information, which is used to realize continuous estimation and high-precision correction of the relative position, relative attitude and motion trend between the UAV and the inland waterway mobile platform.

[0044] The coarse estimation layer, based on GNSS and IMU data, rapidly estimates the relative motion between the UAV and the inland waterway mobile platform. Let the UAV's position vector in the global coordinate system be... The position vector of the inland waterway mobile platform in the global coordinate system is The rough relative position of the UAV to the inland waterway mobile platform is estimated as follows: In terms of attitude representation, unit quaternions are used to represent the absolute attitudes of the UAV and the inland waterway mobile platform. Let the absolute attitude quaternion of the UAV be... The absolute attitude quaternion of the inland waterway mobile platform is Then the rough relative attitude estimate is: ; in, To represent quaternion multiplication, This represents the inverse of the attitude quaternion of the inland waterway mobile platform. The coarse estimation layer uses either Extended Kalman Filter (EKF) or Unscented Kalman Filter (UKF) to fuse GNSS and IMU data for short-term prediction; its state vector is defined as... The state transition equation is expressed as: ; The observation equation is expressed as: ; in, Indicates IMU measurement input, and These represent process noise and observation noise, respectively. This coarse estimation layer has low computational cost and low latency, and can maintain the continuity and safety margin of relative pose estimation between the UAV and the inland waterway mobile platform when the visual sensor is temporarily unavailable.

[0045] When the data quality output by the vision sensor meets a preset threshold, the precise estimation layer is activated to constrain and correct the coarse estimation results, thereby obtaining a high-precision relative pose estimate. Let the set of three-dimensional feature points on the deck of the inland waterway mobile platform be... The corresponding image plane pixels are The camera projection model is represented as ,in, and Let represent the rotation matrix and translation vector of the UAV relative to the inland waterway's moving plane, respectively. Construct the visual reprojection error cost function: ; in, A robust kernel function is used to suppress the influence of outliers. When the number of matching feature points and the reprojection error meet the conditions, bundle adjustment (BA) is used to solve for the optimal relative transformation; otherwise, a fast approximate solution is performed by combining PnP and RANSAC, thus achieving an engineering trade-off between real-time performance and accuracy.

[0046] Step 13, Short-term motion trend prediction of the inland waterway mobile platform. Based on the two-layer estimation results, this embodiment analyzes the motion characteristics of the inland waterway mobile platform under the influence of river flow velocity changes, backflow disturbances, and maneuvering behavior, generating motion trend prediction data for the inland waterway mobile platform within a preset time window. The historical state sequence of the inland waterway mobile platform is defined as follows: ,in, The future state is predicted using a Long Short-Term Memory (LSTM) network: ; The loss function is defined as: ; When computing power is limited, it can degenerate into a GRU model or a first-order Markov prediction model to ensure that the system has a minimum prediction capability.

[0047] Step 14, Dynamic Landing Trajectory Planning. Based on the prediction results and UAV dynamic constraints, a Model Predictive Control (MPC) problem is constructed. The UAV state vector is defined as... The state transition equation is The cost function is defined as: ; It also introduces attitude, velocity, and safety constraints.

[0048] Step 15: Closed-loop control achieves a safe landing. The control layer adopts a closed-loop control structure combining PID control and adaptive disturbance compensation. The control law is as follows: ; Lyapunov analysis was used to ensure system stability.

[0049] In summary, this embodiment achieves autonomous landing control of an unmanned aerial vehicle (UAV) on an inland waterway mobile platform through a model predictive control (MPC) optimization strategy and an adaptive disturbance compensation mechanism. This enables the UAV to maintain proactive predictive following of the platform's motion state throughout the landing process, effectively overcoming relative attitude misalignment caused by platform motion lag and environmental disturbances. It also significantly reduces the impact of crosswind interference and sudden changes in platform pitch and roll on landing safety. This embodiment addresses the nonlinear motion characteristics of mobile platforms in inland waterway navigation environments caused by factors such as water flow, wind fields, and maneuvering actions. It constructs an autonomous landing technology system integrating multi-source sensing, short-term motion prediction, dynamic trajectory planning, and closed-loop control. Under conditions of strong disturbances and continuous platform motion, it achieves stable, autonomous, and safe end-point guidance and platform landing control. This technical solution effectively solves the key technical challenges that have long existed in the autonomous recovery of UAVs on inland waterway mobile platforms, such as positioning lag, trajectory mismatch, and control instability. It provides a reliable engineering implementation path for autonomous operation, inspection monitoring, and intelligent recovery of UAVs in inland waterways, and also provides an important technical foundation for intelligent unmanned system equipment systems for inland waterway scenarios.

[0050] The autonomous landing method for UAV mobile platforms based on target motion prediction proposed in this embodiment has at least the following advantages compared with traditional UAV autonomous landing technology for ship mobile platforms.

[0051] The proposed autonomous landing method for UAV mobile platforms based on target motion prediction has at least the following advantages compared to existing autonomous landing technologies for mobile platforms (including but not limited to inland waterway vessel platforms, low-speed navigation platforms, etc.).

[0052] First, this embodiment constructs a six-degree-of-freedom motion state estimation and prediction mechanism for mobile platforms by fusing multi-sensory information from Global Navigation Satellite System (GNSS), Inertial Measurement Unit (IMU), and airborne vision sensors. The fusion model utilizes extended Kalman filtering or unscented Kalman filtering to model and constrain nonlinear attitude changes, enabling continuous and stable motion state estimation results even under complex motion conditions such as lateral drift, yaw sway, and slow attitude changes. Compared to technical solutions that rely solely on single GNSS or IMU information, this embodiment significantly improves the robustness and reliability of mobile platform motion state estimation, effectively reducing positioning drift and attitude instability caused by environmental interference or sensor degradation.

[0053] Secondly, this embodiment introduces a two-layer estimation structure consisting of a coarse estimation layer and a precise estimation layer, enabling continuous estimation and high-precision correction of the relative pose between the UAV and the mobile platform. The coarse estimation layer rapidly obtains an initial estimate of the relative position and attitude based on GNSS and IMU data, maintaining control continuity even when the visual sensor is temporarily unavailable or degraded. The precise estimation layer, when visual conditions meet requirements, optimizes the coarse estimation results through visual feature matching, geometric constraints, and inertial attitude compensation, thereby obtaining a more accurate relative pose output. This two-layer estimation structure balances continuity and accuracy requirements in engineering, effectively suppressing the risk of error accumulation and providing stable and reliable pose input for the safe landing of the UAV.

[0054] Third, this embodiment introduces a time-series-based target motion prediction mechanism to proactively estimate the motion trend of the mobile platform within a short time window. By modeling historical motion state sequences and combining them with recurrent neural networks, long short-term memory networks, or simplified dynamic models, the future displacement and attitude changes of the platform are predicted, transforming the UAV's trajectory planning from a traditional passive following method to an active compensation method. This design can adjust the UAV's approach path and descent strategy in advance when there are random disturbances or sudden changes in the platform's motion, significantly improving the fault tolerance and success rate of the autonomous landing process.

[0055] Fourth, this embodiment deeply couples the motion trend prediction results with the dynamic trajectory planning and closed-loop control process. By combining model predictive control with error compensation control, it achieves coordinated regulation of the UAV's attitude, speed, and path. This closed-loop control mechanism can correct control deviations caused by platform motion, wind disturbances, or sensor noise in real time, enabling the UAV to maintain a stable and controllable flight state during approach and contact with the moving platform, thereby effectively improving the safety and stability of autonomous landing.

[0056] It should be noted that the above division of steps and modules is only for clearly describing the technical solution of this embodiment. In specific implementation, the steps can be combined, split, or implemented in an equivalent manner. Any equivalent transformations or non-essential improvements made to the algorithm flow, model structure, or implementation form without changing the core idea and technical effect of this embodiment should fall within the protection scope of this application.

[0057] Another embodiment of this application proposes an autonomous landing system for a mobile UAV platform based on target motion prediction. The system structure and working principle of this embodiment will be described in detail below with reference to the accompanying drawings. It should be noted that the following content is only used to explain the technical solution of this embodiment and does not constitute a limitation on the scope of protection of this application. Figure 6 As shown, the autonomous landing system for a UAV mobile platform based on target motion prediction proposed in this embodiment includes: a data acquisition module 21, a two-layer estimation module 22, a short-term prediction module 23, a path planning module 24, and a landing control module 25. The modules form a closed-loop collaborative working mechanism according to the technical route of "perception-estimation-prediction-planning-control".

[0058] The data acquisition module 21 is used to acquire real-time motion status information of the UAV and mobile platform based on multi-sensor fusion. The multi-sensors include at least an airborne Global Navigation Satellite System (GNSS), an Inertial Measurement Unit (IMU), and a visual sensor. In its implementation, this module is responsible for time synchronization and preliminary preprocessing of raw data from different sensors, and for mapping the UAV's own status information and the mobile platform's observation information to the same reference coordinate system, providing basic data input for subsequent estimation and prediction. It should be noted that the output quality of the visual sensor may fluctuate due to lighting, occlusion, or environmental conditions; the data acquisition module does not require visual information to be always available.

[0059] The dual-layer estimation module 22 is used to construct a dual-layer estimation structure consisting of a coarse estimation layer and a precise estimation layer based on the real-time motion state information output by the data acquisition module 21. The coarse estimation layer primarily uses GNSS and IMU data to quickly and initially estimate the relative position, relative attitude, and motion trend between the UAV and the mobile platform, ensuring continuous positioning and control capabilities even in the event of missing or degraded visual information. The precise estimation layer activates when the visual sensor output meets preset quality conditions. By fusing visual observation data and IMU attitude information, it uses extended Kalman filtering, unscented Kalman filtering, or equivalent nonlinear estimation methods to constrain and correct the coarse estimation results, thereby achieving high-precision estimation of the six-degree-of-freedom nonlinear motion state of the mobile platform. This dual-layer estimation structure achieves a balance between positioning continuity and estimation accuracy in engineering applications, improving the system's robustness in complex environments.

[0060] The short-term prediction module 23 is used to predict the motion trend of the mobile platform within a short time window based on the estimation results output by the dual-layer estimation module 22. In its implementation, this module analyzes the dynamic characteristics of the mobile platform under external disturbances (including but not limited to water flow, wind fields, or operational behavior), and combines time series modeling and motion model fitting methods to generate predicted data on the position and attitude changes of the mobile platform in the short term. This prediction result is input as forward-looking information into the path planning module, enabling the UAV to perceive the motion trend of the target platform in advance, thereby avoiding passive following based solely on the current state.

[0061] The path planning module 24 is used to design a dynamic landing trajectory planning algorithm based on the motion trend prediction data output by the short-term prediction module 23 and the dynamic constraints of the UAV itself, and to generate the optimal landing path that matches the motion trend of the mobile platform in real time. During the planning process, the module comprehensively considers the speed limit, attitude constraint, descent angle constraint of the UAV and the reachability area constraint of the mobile platform, and dynamically optimizes the approach path of the UAV so that the generated landing trajectory satisfies both physical feasibility and has a good safety margin.

[0062] The landing control module 25 is used to perform closed-loop control of the UAV's attitude and speed based on the prediction results of the short-term prediction module 23 and the optimal landing path generated by the path planning module 24. In specific implementations, this module can employ PID control, model predictive control, or a combination of both control strategies to convert the planned trajectory into actual control commands for the UAV, and continuously correct control errors based on real-time feedback, thereby achieving a safe and stable landing of the UAV under continuous motion conditions on the mobile platform.

[0063] It is worth mentioning that each module in this embodiment is a functional logic module. In actual engineering implementation, it can be implemented by independent hardware units, different software modules on the same processor, or multiple physical modules working together. This system embodiment corresponds to the aforementioned method embodiment. Each module forms a complete closed-loop autonomous landing system through hierarchical collaboration, enabling the UAV to actively compensate and dynamically adjust based on motion prediction results when the motion state of the mobile platform is constantly changing. This significantly improves the success rate, safety, and engineering adaptability of autonomous landing. Furthermore, to highlight the innovative aspects of this application, this embodiment does not introduce modules that are not closely related to solving the technical problems proposed in this application. However, this does not mean that other modules do not exist in this embodiment.

[0064] It is not difficult to see that this embodiment is a system embodiment corresponding to the above method embodiments, and this embodiment can be implemented in conjunction with the above method embodiments. The relevant technical details and technical effects mentioned in the above method embodiments are still valid in this embodiment, and will not be repeated here to reduce repetition. Accordingly, the relevant technical details mentioned in this embodiment can also be applied to the above method embodiments.

[0065] This application proposes an autonomous landing method for unmanned aerial vehicle (UAV) mobile platforms on inland waterways based on target motion prediction. This method integrates motion prediction, navigation fusion, intelligent perception, and autonomous control of inland waterway mobile platforms. With the continuous development of intelligent and autonomous UAV operation scenarios on inland waterways, this technology has become an important development direction for inland waterway inspection, collaborative operations of inland waterway vessels, and the application of unmanned systems in waterways. Its core idea lies in real-time modeling and short-term prediction of the dynamic motion state (including six degrees of freedom motion such as sway, roll, pitch, and yaw) of the inland waterway mobile platform under complex water flow, wind field, and manipulation disturbance conditions. Combined with multi-source information fusion from visual sensing, inertial navigation systems (INS), and global navigation satellite systems (GNSS), this method achieves high-precision, autonomous landing and relative navigation control of the UAV on a continuously moving platform. Through predictive compensation for the platform's dynamic trends, this technology can effectively reduce the impact of water flow disturbances and platform attitude changes on the landing process, thereby significantly improving the safety and mission reliability of autonomous landing of UAVs in complex inland waterway environments.

[0066] In inland waterway unmanned aerial vehicle (UAV) operation systems, this method can be applied to the autonomous recovery and landing control of UAVs on inland waterway vessels or other mobile platforms. By estimating and predicting the trajectory and attitude of the mobile platform in real time, the UAV can perform trajectory compensation and attitude adjustment before the platform's motion reaches its extreme value, achieving flexible and autonomous landing on the mobile platform deck. This reduces reliance on manual guidance, fixed visual markers, or manual operation, significantly improving the automation level and success rate of UAV recovery in inland waterway operation environments.

[0067] In multi-platform collaborative operation scenarios, this method can be applied to UAV dynamic transfer, material delivery, and information exchange tasks under multi-vessel collaborative conditions in inland waterways. UAVs can rely on inland waterway mobile platform motion prediction models to perceive and predict the relative motion relationship between the take-off and landing platforms and the receiving platform in short time, thereby completing autonomous take-off, landing, and docking operations under multi-platform conditions. This ensures the continuity of collaborative operations in inland waterways and the stability of task coordination, and is suitable for applications such as inland waterway inspection, emergency response, and joint operations.

[0068] In inland waterway emergency rescue and logistical support missions, this method can be used for the autonomous landing and material delivery of mobile supply vessels, rescue vessels, or emergency operation platforms by UAVs. By making short-term predictions of the motion status of mobile platforms in inland waterways, UAVs can dynamically plan safe landing windows and approach trajectories, achieving precise alignment and safe landing of continuously moving platforms, significantly improving the timeliness and safety of emergency response and material delivery missions in inland waterways.

[0069] In inland waterway unmanned system mission scenarios, this method can effectively support the autonomous recovery and repetitive operation of various types of inland waterway operation drones (such as inspection drones, transport drones, and monitoring drones), significantly reducing human intervention and environmental dependence, and improving the mission continuity and response efficiency of drone systems in complex inland waterways. Simultaneously, under multi-platform collaborative operation conditions, this technology can support the dynamic transfer and energy replenishment of drones between different inland waterway mobile platforms, enhancing the overall intelligence level of inland waterway unmanned systems.

[0070] In inland waterway unmanned system mission scenarios, this method can effectively support the autonomous recovery and repetitive operation of various types of inland waterway operation drones (such as inspection drones, transport drones, and monitoring drones), significantly reducing human intervention and environmental dependence, and improving the mission continuity and response efficiency of drone systems in complex inland waterways. Simultaneously, under multi-platform collaborative operation conditions, this technology can support the dynamic transfer and energy replenishment of drones between different inland waterway mobile platforms, enhancing the overall intelligence level of inland waterway unmanned systems.

[0071] In inland waterway emergency response and waterway operation support scenarios, this method enables UAVs to autonomously land on mobile vessels, supply platforms, or temporary floating platforms and deliver supplies, providing efficient and safe technical support for inland waterway emergency response, patrol and support, and disaster response. Simultaneously, in the field of inland waterway monitoring and waterway information collection, UAVs can achieve automatic take-off and landing, data transmission, and collaborative observation based on inland waterway mobile platform motion prediction and relative navigation algorithms, significantly improving the information acquisition capabilities and situational awareness level of inland waterways.

[0072] Furthermore, this method can be extended to inland waterway UAV docking, energy replenishment, and collaborative operation systems. By predicting the relative motion state between the inland waterway mobile platform and the UAV, the system can achieve precise spatial docking and attitude synchronization, providing reliable technical support for long-endurance and continuous operation of inland waterway UAVs, and is suitable for inland waterway inspection, logistics transportation, and long-term monitoring tasks.

[0073] For future smart inland waterways, intelligent shipping, and unmanned inland waterway operation systems, this method can be further combined with intelligent algorithms such as deep learning, reinforcement learning, and edge computing to achieve higher-precision prediction and adaptive control of the motion status of mobile platforms on inland waterways by unmanned aerial vehicles (UAVs). By constructing an integrated system architecture that coordinates air, ship, and platform, this technology will become an important foundational technology supporting future unmanned inland waterway systems, intelligent waterway management, and intelligent waterway construction, possessing significant engineering application value and promising prospects for industrial promotion.

[0074] Another embodiment of this application provides an electronic device, such as Figure 7 As shown, it includes a processor 31 and a memory 32. The memory 32 stores instructions that the processor 31 can execute. When the processor 31 is configured to execute the instructions, the electronic device can realize an autonomous landing method for a UAV mobile platform based on target motion prediction as described in the above method embodiment.

[0075] The memory and processor are connected via a bus, which includes any number of interconnecting buses and bridges, connecting various circuits of one or more processors and the memory. The bus can also connect various other circuits such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art and will not be described further herein. The bus interface provides an interface between the bus and the transceiver. The transceiver can be a single component or multiple components, such as multiple receivers and transmitters, providing a unit for communicating with various other devices over a transmission medium. Data processed by the processor is transmitted over the wireless medium via an antenna, which further receives data and transmits it to the processor.

[0076] The processor manages the bus and general processing, and also provides various functions, including timing, peripheral interfaces, voltage regulation, power management, and other control functions. Memory is used to store data used by the processor during operation.

[0077] Another embodiment of this application proposes a computer-readable storage medium storing a computer program that, when executed by a processor, enables an autonomous landing method for an unmanned aerial vehicle (UAV) inland waterway mobile platform based on target motion prediction as described in the above method embodiments.

[0078] That is, those skilled in the art will understand that all or part of the steps in the above method embodiments can be implemented by a program instructing related hardware. The program is stored in a storage medium and includes several instructions to cause a device (such as a microcontroller, chip, etc.) or processor to execute all or part of the steps of the method described in the method embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory, random access memory, magnetic disks, or optical disks.

[0079] It will be understood by those skilled in the art that the above embodiments are specific implementations of this application, and various changes in form and detail can be made in practical applications without departing from the spirit and scope of this application. For those skilled in the art, several improvements and modifications can be made without departing from the principles of this application, and these improvements and modifications are also considered to be within the scope of protection of this application.

Claims

1. A method for autonomous landing of an unmanned aerial vehicle (UAV) inland waterway mobile platform based on target motion prediction, characterized in that, The method includes: The real-time motion status information of UAVs and inland waterway mobile platforms is obtained based on multi-sensor fusion, wherein the multi-sensor includes at least a Global Navigation Satellite System (GNSS), an Inertial Measurement Unit (IMU), and a visual sensor. Based on real-time motion state information, a two-layer estimation structure consisting of a coarse estimation layer and a precise estimation layer is established. The coarse estimation layer uses IMU data and GNSS data to make a preliminary estimate of the relative position, relative attitude, and motion trend between the UAV and the inland waterway mobile platform. When the visual conditions meet the preset requirements, the precise estimation layer fuses visual sensor data and IMU data to constrain and correct the preliminary estimation results in order to obtain a high-precision relative pose estimation result. Based on the estimation results of the two-layer estimation structure, the dynamic characteristics of the inland waterway mobile platform are analyzed in response to the non-periodic motion characteristics generated by the inland waterway mobile platform under the influence of external environmental disturbances, and motion trend prediction data of the inland waterway mobile platform within a preset short time window are generated. Based on motion trend prediction data and the dynamic constraints of the UAV itself, dynamic landing trajectory planning is carried out to generate a landing path that matches the motion trend of the inland waterway mobile platform in real time. Based on motion trend prediction data and landing path, the attitude and speed of the UAV are controlled in a closed loop, enabling the UAV to complete a safe and stable autonomous landing while the inland waterway mobile platform is in continuous motion.

2. The autonomous landing method for an unmanned aerial vehicle (UAV) inland waterway mobile platform based on target motion prediction according to claim 1, characterized in that, A rough estimation is performed by fusing IMU data and GNSS data to quickly estimate the relative pose between the UAV and the inland waterway mobile platform. This is used to maintain the continuity of positioning results and control input when the visual sensor output is unavailable or the data quality is below a preset threshold. The precise estimation is initiated when the output data quality of the visual sensor meets the preset conditions. Based on the multi-view feature information obtained by the visual sensor, the results of the coarse estimation layer are constrained and corrected to improve the accuracy of the relative pose estimation between the UAV and the inland waterway mobile platform. The dual-layer estimation structure switches between coarse and precise estimation based on the data quality of the visual sensor, thereby improving the reliability of relative pose estimation while ensuring system continuity and security.

3. The autonomous landing method for an unmanned aerial vehicle (UAV) mobile platform in inland waterways based on target motion prediction according to claim 2, characterized in that, When the data quality output by the visual sensor meets the preset quality threshold, the accurate estimation process is started. Based on the multi-view feature information obtained by the visual sensor, the relative pose between the UAV and the inland waterway mobile platform is optimized and estimated. Accurate estimation corrects the relative pose result obtained by coarse estimation by constructing visual observation constraints. When the visual observation meets the preset consistency condition, the optimized relative pose estimation result is output. When the data quality output by the visual sensor is lower than the preset quality threshold, the coarse estimation process is maintained, and the relative pose is updated by an approximate solution method based on local visual constraints to ensure the continuity of the relative pose estimation result.

4. The autonomous landing method for an unmanned aerial vehicle (UAV) mobile platform in inland waterways based on target motion prediction according to claim 3, characterized in that, Based on the estimation results of the two-layer estimation structure, the historical motion state of the inland waterway mobile platform is modeled to generate motion trend prediction data of the inland waterway mobile platform within a preset time window, taking into account the non-stationary motion characteristics of the inland waterway mobile platform under the influence of external environmental disturbances. Motion trend prediction involves performing time series analysis on the historical position and attitude state sequences of the inland waterway mobile platform to output the future pose change trend of the inland waterway mobile platform, which can be used to provide a forward-looking motion reference for the autonomous landing process of the UAV. When computational resources are limited or real-time requirements are high, motion trend prediction adopts an approximate prediction model with low computational complexity to ensure the real-time nature of the prediction process and the stability of system operation.

5. The autonomous landing method for an unmanned aerial vehicle (UAV) mobile platform in inland waterways based on target motion prediction according to claim 4, characterized in that, Based on the motion trend prediction data of the inland waterway mobile platform and the dynamic constraints of the UAV itself, dynamic trajectory planning is performed on the UAV's landing process to generate a landing trajectory that matches the motion trend of the inland waterway mobile platform in real time. The dynamic trajectory planning comprehensively considers the UAV's position, speed, attitude, and control input constraints during the planning process to ensure that the generated landing trajectory meets flight safety requirements. During the UAV's movement along the landing trajectory, the attitude and speed of the UAV are controlled in a closed loop according to the real-time motion status and motion trend prediction data of the inland waterway mobile platform, enabling the UAV to complete a safe and stable autonomous landing under the condition of continuous movement of the inland waterway mobile platform.

6. The autonomous landing method for an unmanned aerial vehicle (UAV) mobile platform in inland waterways based on target motion prediction according to claim 5, characterized in that, In the dynamic trajectory planning and closed-loop control process, flight constraints are introduced to ensure the physical feasibility and safety of the landing process. These constraints include at least: UAV attitude angle constraints, which limit the roll and pitch angle ranges of the UAV during landing; descent speed constraints, which limit the descent speed of the UAV when approaching the inland waterway mobile platform to avoid deck impact; and relative attitude coordination constraints, which ensure that the predicted position and attitude of the UAV enter the captureable area of ​​the inland waterway mobile platform during the terminal phase of landing. At the same time, during the execution of the landing trajectory, a disturbance compensation mechanism is introduced to suppress external disturbances, thereby enhancing the landing robustness of the UAV in complex water surface environments and improving the success rate and safety of autonomous landing.

7. The autonomous landing method for an unmanned aerial vehicle (UAV) inland waterway mobile platform based on target motion prediction according to claim 6, characterized in that, In the process of dynamic trajectory planning and closed-loop control, a degradation substitution mechanism is set up to reduce the computational load and ensure the continuous operation of the system; when the amplitude of the motion state change of the inland waterway mobile platform increases or the uncertainty of motion trend prediction increases, the prediction time domain of trajectory planning is shortened to improve the response speed. When system computing resources are limited or real-time requirements increase, the trajectory planning and control process is switched to an alternative strategy with lower computational complexity in order to maintain basic control over the UAV. When the quality of sensor information deteriorates or some sensors become unavailable, a degraded alignment strategy based on coarse estimation results is activated to ensure the continuity of relative pose estimation between the UAV and the inland waterway mobile platform. By employing a degradation and replacement mechanism, drones can maintain controlled operation even in complex water environments and under multi-source uncertainties, thereby improving the safety and reliability of the autonomous landing process.

8. An autonomous landing system for an unmanned aerial vehicle (UAV) inland waterway mobile platform based on target motion prediction, characterized in that, The system includes: The data acquisition module is used to acquire real-time motion status information of UAVs and inland waterway mobile platforms based on multi-sensor fusion. The multi-sensors include at least Global Navigation Satellite System (GNSS), Inertial Measurement Unit (IMU), and visual sensors. The dual-layer estimation module is used to construct a dual-layer estimation structure consisting of a coarse estimation layer and a precise estimation layer based on real-time motion state information. The coarse estimation layer is used to obtain a preliminary estimate of the relative pose between the UAV and the inland waterway mobile platform based on GNSS data and IMU data when visual information is unavailable or of reduced quality. The precise estimation layer is used to correct the preliminary estimate of the relative pose based on visual observation information when visual information meets the requirements. The short-term prediction module is used to analyze the motion characteristics of the inland waterway mobile platform under the influence of external disturbances such as water flow and wind field, based on the estimation results of the two-layer estimation structure, and generate short-term motion trend prediction data of the inland waterway mobile platform. The path planning module is used to perform dynamic trajectory planning for the landing process of the UAV based on short-term motion trend prediction data and the dynamic constraints of the UAV itself, and generate a landing trajectory that matches the motion trend of the inland waterway mobile platform. The landing control module is used to perform closed-loop control of the UAV's attitude and speed during its movement along the landing trajectory, based on the real-time motion status of the inland waterway mobile platform and the short-term motion trend prediction data, so as to achieve safe, stable and autonomous landing of the UAV under the continuous motion conditions of the inland waterway mobile platform.

9. The autonomous landing system for an unmanned aerial vehicle (UAV) inland waterway mobile platform based on target motion prediction according to claim 1, characterized in that, The landing area of ​​the inland waterway mobile platform is equipped with visual landing markers. The visual landing markers have a nested structure, including an outer guidance marker and an inner precision positioning marker. The outer guidance marker is used for initial target detection and coarse alignment of the UAV at medium and long distances, while the inner precision positioning marker is used for precise relative pose estimation and terminal guidance of the UAV at close distances.

10. The autonomous landing system for an unmanned aerial vehicle (UAV) inland waterway mobile platform based on target motion prediction according to claim 9, characterized in that, The visual landing marker includes a lateral reference line that aligns with the heading of the inland waterway mobile platform. This lateral reference line provides visual constraints on the direction of motion of the inland waterway mobile platform during the UAV's visual perception process, assisting the UAV in estimating the heading of the inland waterway mobile platform and using it for attitude alignment during landing.

Images