Quadrotor unmanned aerial vehicle attitude estimation method and device based on improved UKF, and medium

Abstract

The invention discloses a quadrotor unmanned aerial vehicle attitude estimation method and device based on an improved UKF and a medium, and the method comprises the steps: constructing a kinetic model of a quadrotor unmanned aerial vehicle, deducing an attitude angle kinetic equation through a Newton-Euler method, and superposing a wind shear model and a turbulent wind field model; the method comprises the following steps: collecting multi-source sensor data, and performing hierarchical filtering processing on the multi-source sensor data through a multi-source data fusion and dynamic noise suppression technology; and performing attitude estimation on the quad-rotor unmanned aerial vehicle by adopting a UKF (Unscented Kalman Filter) algorithm of an improved adaptive factor based on the deduced and superposed kinetic model and the processed multi-source sensor data. Through the above method, the method is closer to a real flight environment, the influence of environment sudden change on data quality is inhibited, and high-precision and high-stability attitude estimation is realized.

Description

Technical Field

[0001] The present application relates to the technical field of attitude estimation, and in particular to a method, device and medium for estimating the attitude of a quadrotor drone based on an improved UKF. Background Art

[0002] Quadrotors play an important role in emergency rescue, agricultural monitoring, industrial inspection and other fields due to their flexibility and wide applicability, and accurate attitude estimation is the core foundation for their stable flight and mission execution. At present, the attitude estimation method based on the Unscented Kalman Filter (UKF) has become the mainstream solution due to its nonlinear processing capabilities, but its adaptability in complex dynamic environments is still significantly insufficient. The traditional UKF algorithm uses a preset fixed noise covariance matrix, which is difficult to cope with sudden sensor interference or model errors during flight. This type of static parameter setting causes the algorithm's estimation accuracy to drop sharply when the environment changes suddenly, and it needs to rely on manual experience to repeatedly adjust the parameters. It lacks real-time calibration capabilities under load changes or external disturbances, which greatly limits the reliability and operational efficiency of the drone.

[0003] Furthermore, existing technologies have significant deficiencies in multi-source sensor data fusion and noise suppression. Although data from sensors such as magnetometers, IMUs, and GPS is widely collected, traditional methods inadequately compensate for drift in low-frequency signals, have limited effectiveness in suppressing high-frequency vibration noise, and lack a dynamic weight allocation mechanism. For example, when magnetometer data is distorted by strong magnetic field interference, the system cannot quickly reduce its weight, resulting in accumulated deviations in the fused data. Furthermore, sensor noise assessment is often based on fixed thresholds or offline calibration, which cannot adapt to environmental changes in real time, resulting in insufficient robustness of the attitude estimator in complex scenarios.

[0004] Furthermore, existing UAV dynamics models often ignore the effects of real-world wind field interference or employ only a simplified single wind field model, resulting in significant deviations from the actual flight environment. In complex airflows characterized by superimposed wind shear and turbulence, traditional models struggle to accurately describe the dynamic behavior of UAVs, further amplifying attitude estimation errors. Existing improvements often focus on local optimization at the algorithmic level, failing to systematically enhance estimation accuracy and anti-interference capabilities from a model-data-algorithm synergy perspective. Summary of the Invention

[0005] The embodiments of the present application provide a quadrotor drone attitude estimation method, device and medium based on an improved UKF to solve the above technical problems.

[0006] On the one hand, the embodiment of the present application provides a quadrotor drone attitude estimation method based on an improved UKF, comprising:

[0007] Construct a dynamic model of a quadrotor drone, derive the attitude angle dynamic equations using the Newton-Euler method, and superimpose the wind shear model and turbulent wind field model;

[0008] Collecting multi-source sensor data and performing hierarchical filtering on the multi-source sensor data through multi-source data fusion and dynamic noise suppression technology;

[0009] Based on the derived superimposed dynamic model and the processed multi-source sensor data, the attitude of the quadrotor UAV is estimated by adopting the UKF algorithm with improved adaptive factor.

[0010] In one implementation of the present application, hierarchical filtering is performed on the multi-source sensor data using multi-source data fusion and dynamic noise suppression technology, specifically including:

[0011] Determining low-frequency signals and high-frequency signals in the multi-source sensor data;

[0012] Based on the geomagnetic model, drift compensation is performed on the low-frequency signal, and sliding average filtering is used to suppress vibration noise in the high-frequency signal;

[0013] The noise level of each sensor is dynamically evaluated to assign a weight coefficient to each sensor in real time.

[0014] In one implementation of the present application, the noise level of each sensor is dynamically evaluated to assign a weight coefficient to each sensor in real time, specifically including:

[0015] Set an independent sliding time window for each sensor and calculate the variance and residual of the data within the sliding time window;

[0016] Based on the variance and the residual, the instantaneous noise intensity of the quadrotor drone is evaluated, and the sensor weight is adjusted in inverse proportion to the noise intensity.

[0017] In one implementation of the present application, based on the geomagnetic model, drift compensation is performed on the low-frequency signal, and sliding average filtering is used to suppress vibration noise in the high-frequency signal, specifically including:

[0018] In view of the magnetometer's electromagnetic interference characteristics, a magnetometer calibration equation is constructed based on the geomagnetic field model, and the least squares method is used to correct the magnetometer's yaw angle data in real time to achieve drift compensation.

[0019] For the gyroscope and accelerometer, the original data is divided into windows according to the time series, and the data mean within the window is calculated. The current sampling value is replaced by the data mean to suppress the vibration noise in the high-frequency signal.

[0020] In one implementation of the present application, based on the derived superimposed dynamic model and the processed multi-source sensor data, and using the UKF algorithm with an improved adaptive factor, the attitude estimation of the quadrotor drone is performed, specifically including:

[0021] According to the trace difference between the predicted covariance matrix and the updated covariance matrix, the process noise covariance matrix is dynamically adjusted;

[0022] Dynamically adjusting the observation noise covariance matrix according to a ratio between a trace of the innovation covariance matrix and a trace of the observation noise covariance matrix;

[0023] The weight ratio of historical data and current data is controlled by the smoothing factor.

[0024] In one implementation of the present application, the process noise covariance matrix is dynamically adjusted according to the trace difference between the predicted covariance matrix and the updated covariance matrix, specifically including:

[0025] Calculating the trace difference between the predicted covariance matrix of the current sampling step and the updated covariance matrix of the previous sampling step, and comparing the trace difference with a preset difference threshold;

[0026] When the trace difference exceeds the preset difference threshold, scaling the process noise covariance matrix according to the ratio, and limiting the minimum scaling ratio of the process noise covariance matrix by a max function;

[0027] A smoothing factor is added to blend the original process noise covariance matrix and the scaled process noise covariance matrix proportionally.

[0028] In one implementation of the present application, the observation noise covariance matrix is dynamically adjusted according to the ratio between the trace of the innovation covariance matrix and the trace of the observation noise covariance matrix, specifically including:

[0029] defining an innovation vector, and calculating a ratio between a trace of an innovation covariance matrix corresponding to the innovation vector and a trace of an observation noise covariance matrix, and comparing the ratio with a preset ratio threshold;

[0030] If the ratio exceeds the threshold ratio threshold, the observation noise covariance matrix is scaled proportionally, and the minimum scaling ratio of the observation noise covariance matrix is limited by a max function;

[0031] A smoothing factor is added to fuse the original over-observation noise covariance matrix and the scaled observation noise covariance matrix proportionally.

[0032] In one implementation of the present application, the attitude angle dynamics equation is derived by the Newton-Euler method, and the wind shear model and the turbulent wind field model are superimposed, specifically including:

[0033] According to the Newton-Euler method, the attitude angle dynamic equation of the quadrotor drone is expressed as follows:

[0034]

[0035] Wherein, U1 represents the input value of the roll angle channel, U2 represents the input value of the pitch angle channel, U3 represents the input value of the yaw angle channel, b represents the anti-torque coefficient of the propeller motor, and Ω1 to Ω4 represent the rotation speeds of the four rotors;

[0036]

[0037] Among them, l represents the farthest distance between the arm and the center of mass of the quadrotor drone, I r The moment of inertia of the rotor rotating around the axis, k i (i=1,2,3) represent the air resistance coefficient of the rotation around the corresponding coordinate axis in the carrier coordinate system, Represents the attitude angle component of the quadrotor drone in the inertial coordinate system, V=[p,q,r] T Represents the angular velocity component of the quadrotor drone in the carrier coordinate system, P = [x, y, z] T Represents the position component of the quadrotor UAV in the carrier coordinate system, I x , I y , I z represent the moment of inertia about the x, y, and z axes respectively.

[0038] On the other hand, an embodiment of the present application further provides a quadrotor drone attitude estimation device based on an improved UKF, the device comprising:

[0039] at least one processor;

[0040] and, a memory communicatively coupled to the at least one processor;

[0041] The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the above-mentioned quadrotor drone attitude estimation method based on the improved UKF.

[0042] On the other hand, an embodiment of the present application further provides a non-volatile computer storage medium storing computer executable instructions, which, when executed, implements the above-mentioned quadrotor drone attitude estimation method based on the improved UKF.

[0043] The embodiments of the present application provide a quadrotor drone attitude estimation method, device, and medium based on an improved UKF, which have at least the following beneficial effects:

[0044] By superimposing wind shear and turbulent wind field models in the dynamic model, the UAV modeling is made closer to the real flight environment, effectively reducing the model mismatch error, providing a more accurate dynamic basis for subsequent attitude estimation, especially in complex airflow scenarios such as strong winds and turbulence, significantly improving the reliability of state prediction; through differentiated processing of multi-source sensor data, combined with real-time evaluation of noise levels and dynamic adjustment of weights, it can adaptively suppress the impact of environmental mutations on data quality, ensuring that the fused observations have both high precision and strong anti-interference capabilities; based on the trace difference of the predicted covariance and the updated covariance, and the trace ratio of the new information covariance to the observation noise, the Q / R matrix is corrected in real time, and the weights of historical and current data are balanced in combination with the smoothing factor, so that the algorithm can quickly respond to sensor noise mutations or model errors, avoid divergence of estimation results, reduce dependence on manual parameter adjustment, and achieve high-precision and high-stability attitude estimation. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0046] Figure 1 A flow chart of a quadrotor drone attitude estimation method based on an improved UKF provided in an embodiment of the present application;

[0047] Figure 2 Schematic diagram of the internal structure of the quadrotor drone attitude estimation device based on the improved UKF provided in an embodiment of the present application. DETAILED DESCRIPTION

[0048] To make the purpose, technical solutions, and advantages of this application more clear, the technical solutions of this application will be clearly and completely described below in conjunction with the specific embodiments of this application and the corresponding drawings. Obviously, the embodiments described are only part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0049] The following describes in detail the technical solutions provided by various embodiments of the present application in conjunction with the accompanying drawings.

[0050] Figure 1 A flowchart of a quadrotor drone attitude estimation method based on an improved UKF is provided in an embodiment of the present application.

[0051] The analysis method involved in the embodiments of the present application can be implemented by a terminal device or a server, and the present application does not impose any special restrictions on this. For ease of understanding and description, the following embodiments are described in detail using a server as an example.

[0052] It should be noted that the server can be a single device or a system composed of multiple devices, that is, a distributed server, and this application does not make any specific restrictions on this.

[0053] like Figure 1 As shown, the quadrotor drone attitude estimation method based on the improved UKF provided in the embodiment of the present application includes:

[0054] Step 101: Construct a dynamic model of the quadrotor drone, derive the attitude angle dynamic equation using the Newton-Euler method, and superimpose a wind shear model and a turbulent wind field model.

[0055] In this embodiment, the construction of the dynamic model is the basis of attitude estimation. It should be noted that the attitude angle dynamic equation of the quadcopter is derived based on the Newton-Euler method, which establishes the roll angle by analyzing the force and torque balance relationship between the rigid body in the inertial system and the carrier coordinate system. Dynamic equations for pitch angle θ and yaw angle ψ. Specifically, the dynamic equation for roll angle is determined by the balance between the torque generated by the rotor lift difference and the air resistance and inertial force. The equations for pitch angle and yaw angle correspond to the torque effects on the front and rear rotors and the left and right rotors, respectively. For example, the input value of the roll angle channel is proportional to the square difference in rotor speed, while the air resistance coefficient and the UAV structural parameters jointly affect the calculation of angular acceleration, such as the length of the aircraft arm and the moment of inertia.

[0056] It is understandable that traditional models often ignore wind field interference or only use a single simplified model, resulting in the accumulation of prediction errors in complex airflow environments. To this end, this step further superimposes the wind shear model and the turbulent wind field model to improve the model accuracy. Specifically, the wind shear model uses a logarithmic model to describe the gradient change of wind speed with altitude during low-altitude flight, which is suitable for wind field simulation when the UAV is operating near the ground. The turbulent wind field model generates random disturbance components based on the spectrum model to simulate the random fluctuation characteristics of the instantaneous wind speed. It should be noted that the forces of the two wind field models are integrated into the dynamic equation after vector superposition. For example, when the UAV encounters lateral wind shear, the model will dynamically correct the predicted value of the roll angle according to the current altitude and wind speed gradient, while the turbulent component generates random disturbance torque through spectrum analysis to enhance the model's adaptability to instantaneous airflow mutations.

[0057] In one embodiment, the attitude angle dynamics equation of the quadrotor drone is derived according to the Newton-Euler method and is expressed as follows:

[0058]

[0059] Wherein, U1 represents the input value of the roll angle channel, U2 represents the input value of the pitch angle channel, U3 represents the input value of the yaw angle channel, b represents the anti-torque coefficient of the propeller motor, and Ω1 to Ω4 represent the rotation speeds of the four rotors;

[0060]

[0061] Among them, l represents the farthest distance between the arm and the center of mass of the quadrotor drone, I r The moment of inertia of the rotor rotating around the axis, k i (i=1,2,3) represent the air resistance coefficient of the rotation around the corresponding coordinate axis in the carrier coordinate system, Represents the attitude angle component of the quadrotor drone in the inertial coordinate system, V=[p,q,r] T Represents the angular velocity component of the quadrotor drone in the carrier coordinate system, P = [x, y, z] T Represents the position component of the quadrotor UAV in the carrier coordinate system, I x , I y , I z represent the moment of inertia about the x, y, and z axes respectively.

[0062] In one embodiment, for real-world interference, primarily wind field effects, two wind fields, wind shear and turbulent wind fields, are considered for a quadcopter environment. The wind shear model uses the logarithmic model proposed by Prandtl, applicable to an altitude range of 30-100m, and the turbulent wind field model uses the Dryden model. The wind field turbulence model, a superposition of the two wind fields, is incorporated into the quadcopter's dynamic model, allowing the drone model to take into account the influence of wind field factors, resulting in a more accurate model and serving as the drone model for attitude estimation.

[0063] Step 102: Collect multi-source sensor data and perform hierarchical filtering on the multi-source sensor data through multi-source data fusion and dynamic noise suppression technology.

[0064] In this embodiment, the fusion and preprocessing of multi-source sensor data are key to ensuring the reliability of observations. It should be noted that the sensor data includes the output signals of the magnetometer, inertial measurement unit (IMU), visual odometry, barometer, and GPS module, and these data have significant differences in frequency, noise characteristics, and susceptibility to interference. Specifically, in order to solve the problems of insufficient low-frequency drift compensation and limited high-frequency noise suppression in traditional methods, the system first performs differential processing on the data through a hierarchical filtering strategy.

[0065] For low-frequency signals, such as the yaw angle data output by a magnetometer, for example, the magnetometer will slowly drift when it is affected by an uneven geomagnetic field or interference from metal objects. To this end, a calibration equation is constructed based on the geomagnetic field model, and the least squares method is used to fit the residual between the magnetometer measurement value and the theoretical geomagnetic vector in real time to dynamically correct the yaw angle deviation. For example, when a quadrotor drone flies close to a high-voltage transmission line, the system automatically detects abnormal offsets in the magnetometer data and triggers calibration to avoid accumulation of yaw angle errors.

[0066] For high-frequency signals, such as the gyroscope and accelerometer data from the IMU, it is understandable that motor vibration will introduce high-frequency noise. Using sliding average filtering technology, the original data sequence is divided into time windows, and the mean of the data within the window is calculated to replace the current sample value. For example, when the drone accelerates or decelerates suddenly, the high-frequency noise of the accelerometer is effectively suppressed through sliding window mean filtering, ensuring the smoothness of the angular velocity and linear acceleration data. Through statistical variance analysis, a data temporary storage mechanism is triggered for abnormal conditions such as sudden changes in ambient air pressure, loss of GPS signals, and power supply noise interference, preventing dirty data from entering the fusion process.

[0067] Furthermore, dynamic noise suppression technology evaluates the sensor noise level in real time and assigns weight coefficients. It should be noted that the system sets an independent sliding time window for each sensor, calculates the variance and residual of the data in the window, and evaluates the instantaneous noise intensity. The higher the noise intensity, the lower the weight coefficient of the sensor. For example, when the magnetometer is close to a strong magnetic field source such as a building or large metal equipment, the noise level of the magnetometer increases significantly, and the weight coefficient of the magnetometer is quickly reduced. When the GPS signal jumps due to occlusion by a building, the system identifies the sudden increase in noise through variance calculation, automatically reduces the fusion weight of the GPS data, and at the same time increases the weight ratio of the IMU and visual odometry to ensure the continuity of attitude estimation.

[0068] Step 103: Based on the derived superimposed dynamic model and the processed multi-source sensor data, and using the UKF algorithm with an improved adaptive factor, the attitude of the quadrotor UAV is estimated.

[0069] In this embodiment, the UKF algorithm with an improved adaptive factor is the core of attitude estimation. It should be noted that the traditional UKF algorithm uses a fixed noise covariance matrix (Q / R), making it difficult to adapt to sudden changes in model errors and sensor noise in dynamic environments. Therefore, this solution improves the algorithm's adaptability by dynamically adjusting the Q / R matrix and introducing a smoothing factor.

[0070] Specifically, the adjustment of the process noise covariance matrix (Q) is based on the trace difference between the predicted covariance matrix and the updated covariance matrix. It can be understood that the trace difference reflects the degree of accumulation of the model prediction error. If the difference exceeds the preset threshold, it indicates that the process noise is underestimated, and the process noise covariance Q matrix needs to be scaled down, and the minimum scaling ratio is limited by the max function to avoid over-compression. For example, when the UAV suddenly encounters strong turbulence, the model prediction error increases, and the trace difference triggers the dynamic adjustment of the Q matrix, enhancing the adaptability of the algorithm to wind field disturbances.

[0071] The adjustment of the observation noise covariance matrix (R) is achieved through the ratio of the trace of the new information covariance matrix to the trace of the R matrix. It should be noted that the new information vector represents the residual between the observed value and the predicted value, and the increase of its covariance trace indicates that the reliability of the sensor data has decreased. If the ratio exceeds the threshold, the system will scale down the R matrix and reduce the weight of the abnormal data. For example, when the magnetometer is subjected to strong electromagnetic interference and the yaw angle data is distorted, the new information covariance trace increases significantly, triggering the adjustment of the R matrix and suppressing the impact of abnormal data on attitude estimation.

[0072] Furthermore, a smoothing factor is introduced to balance the weight of historical data with the current adjustment value. For example, in the event of a brief, sudden change in sensor noise, such as a momentary loss of the GPS signal, the smoothing factor proportionally blends the historical Q / R matrix with the current adjustment value, for example, at 90% historical weight. This prevents oscillations in the estimation result caused by parameter changes and ensures algorithm stability.

[0073] The dynamics model provides accurate state predictions for the UKF, preprocessed multi-source data serves as highly reliable observation input, and adaptive Q / R adjustment dynamically optimizes filter parameters based on the real-time environment. The synergistic effect of these three factors significantly reduces the error in the UAV's roll, pitch, and yaw angle estimates in complex airflow and electromagnetic interference scenarios, providing reliable support for precise control and mission execution.

[0074] The above is an embodiment of the method proposed in this application. Based on the same inventive concept, this application embodiment also provides a quadrotor drone attitude estimation device based on improved UKF, whose structure is as follows: Figure 2 shown.

[0075] Figure 2 This is a schematic diagram of the internal structure of the quadrotor drone attitude estimation device based on the improved UKF provided in the embodiment of the present application. Figure 2 As shown, the equipment includes:

[0076] at least one processor;

[0077] and, a memory communicatively coupled to the at least one processor;

[0078] The memory stores instructions that can be executed by at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to:

[0079] Construct a dynamic model of a quadrotor drone, derive the attitude angle dynamic equations using the Newton-Euler method, and superimpose the wind shear model and turbulent wind field model;

[0080] Collect multi-source sensor data and perform hierarchical filtering on the multi-source sensor data through multi-source data fusion and dynamic noise suppression technology;

[0081] Based on the derived superimposed dynamic model and processed multi-source sensor data, the attitude estimation of the quadrotor UAV is performed using the UKF algorithm with improved adaptive factor.

[0082] The present application also provides a non-volatile computer storage medium storing computer-executable instructions. When the computer-executable instructions are executed, they can:

[0083] Construct a dynamic model of a quadrotor drone, derive the attitude angle dynamic equations using the Newton-Euler method, and superimpose the wind shear model and turbulent wind field model;

[0084] Collect multi-source sensor data and perform hierarchical filtering on the multi-source sensor data through multi-source data fusion and dynamic noise suppression technology;

[0085] Based on the derived superimposed dynamic model and processed multi-source sensor data, the attitude estimation of the quadrotor UAV is performed using the UKF algorithm with improved adaptive factor.

[0086] The various embodiments in this application are described in a progressive manner. Similar portions between the various embodiments can be referred to in conjunction with each other. Each embodiment focuses on the differences between the other embodiments. In particular, the device and medium embodiments are generally similar to the method embodiments, so their descriptions are relatively simple. For relevant portions, refer to the descriptions of the method embodiments.

[0087] The devices and media provided in the embodiments of the present application correspond one-to-one to the methods. Therefore, the devices and media also have similar beneficial technical effects to their corresponding methods. Since the beneficial technical effects of the methods have been described in detail above, the beneficial technical effects of the devices and media will not be repeated here.

[0088] Those skilled in the art will appreciate that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0089] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the steps in the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0090] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0091] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0092] In a typical configuration, a computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory.

[0093] Memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. Memory is an example of a computer-readable medium.

[0094] Computer-readable media includes permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. The information can be computer-readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassettes, magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include transitory computer-readable media (transitory media), such as modulated data signals and carrier waves.

[0095] It should also be noted that the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, commodity, or apparatus that includes a series of elements includes not only those elements but also other elements not explicitly listed, or includes elements inherent to such process, method, commodity, or apparatus. In the absence of further limitations, an element defined by the phrase "comprises a ..." does not exclude the presence of other identical elements in the process, method, commodity, or apparatus that includes the element.

[0096] The above are merely embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application should all be included within the scope of the claims of the present application.

Claims

1. A quadrotor UAV attitude estimation method based on improved UKF is characterized by: The method comprises: Construct a dynamic model of a quadrotor drone, derive the attitude angle dynamic equations using the Newton-Euler method, and superimpose the wind shear model and turbulent wind field model; Collecting multi-source sensor data and performing hierarchical filtering on the multi-source sensor data through multi-source data fusion and dynamic noise suppression technology; Based on the derived superimposed dynamic model and the processed multi-source sensor data, the attitude of the quadrotor UAV is estimated by adopting the UKF algorithm with improved adaptive factor.

2. The quadrotor UAV attitude estimation method based on improved UKF according to claim 1 is characterized in that: The multi-source sensor data is subjected to hierarchical filtering through multi-source data fusion and dynamic noise suppression technology, specifically including: Determining low-frequency signals and high-frequency signals in the multi-source sensor data; Based on the geomagnetic model, drift compensation is performed on the low-frequency signal, and sliding average filtering is used to suppress vibration noise in the high-frequency signal; The noise level of each sensor is dynamically evaluated to assign a weight coefficient to each sensor in real time.

3. The quadrotor UAV attitude estimation method based on improved UKF according to claim 2 is characterized in that: Dynamically evaluate the noise level of each sensor to assign a weight coefficient to each sensor in real time, including: Set an independent sliding time window for each sensor and calculate the variance and residual of the data within the sliding time window; Based on the variance and the residual, the instantaneous noise intensity of the quadrotor drone is evaluated, and the sensor weight is adjusted in inverse proportion to the noise intensity.

4. The quadrotor UAV attitude estimation method based on improved UKF according to claim 2 is characterized in that: Based on the geomagnetic model, the low-frequency signal is drift compensated, and the vibration noise in the high-frequency signal is suppressed by using a sliding average filter, specifically including: In view of the magnetometer's electromagnetic interference characteristics, a magnetometer calibration equation is constructed based on the geomagnetic field model, and the least squares method is used to correct the magnetometer's yaw angle data in real time to achieve drift compensation. For the gyroscope and accelerometer, the original data is divided into windows according to the time series, and the data mean within the window is calculated. The current sampling value is replaced by the data mean to suppress the vibration noise in the high-frequency signal.

5. The quadrotor UAV attitude estimation method based on improved UKF according to claim 1 is characterized in that: Based on the derived superimposed dynamic model and the processed multi-source sensor data, and using the UKF algorithm with an improved adaptive factor, the attitude estimation of the quadrotor UAV is performed, specifically including: According to the trace difference between the predicted covariance matrix and the updated covariance matrix, the process noise covariance matrix is dynamically adjusted; Dynamically adjusting the observation noise covariance matrix according to a ratio between a trace of the innovation covariance matrix and a trace of the observation noise covariance matrix; The weight ratio of historical data and current data is controlled by the smoothing factor.

6. The method for estimating the attitude of a quadrotor drone based on an improved UKF according to claim 5, wherein: According to the trace difference between the predicted covariance matrix and the updated covariance matrix, the process noise covariance matrix is dynamically adjusted, specifically including: Calculating the trace difference between the predicted covariance matrix of the current sampling step and the updated covariance matrix of the previous sampling step, and comparing the trace difference with a preset difference threshold; When the trace difference exceeds the preset difference threshold, scaling the process noise covariance matrix according to the ratio, and limiting the minimum scaling ratio of the process noise covariance matrix by a max function; A smoothing factor is added to blend the original process noise covariance matrix and the scaled process noise covariance matrix proportionally.

7. The quadrotor UAV attitude estimation method based on improved UKF according to claim 5 is characterized in that: According to the ratio between the trace of the innovation covariance matrix and the trace of the observation noise covariance matrix, the observation noise covariance matrix is dynamically adjusted, specifically including: defining an innovation vector, and calculating a ratio between a trace of an innovation covariance matrix corresponding to the innovation vector and a trace of an observation noise covariance matrix, and comparing the ratio with a preset ratio threshold; If the ratio exceeds the threshold ratio threshold, the observation noise covariance matrix is scaled proportionally, and the minimum scaling ratio of the observation noise covariance matrix is limited by a max function; A smoothing factor is added to fuse the original over-observation noise covariance matrix and the scaled observation noise covariance matrix proportionally.

8. The quadrotor UAV attitude estimation method based on improved UKF according to claim 1 is characterized in that: The attitude angle dynamic equations are derived using the Newton-Euler method, and the wind shear model and turbulent wind field model are superimposed, specifically including: According to the Newton-Euler method, the attitude angle dynamic equation of the quadrotor drone is expressed as follows: Wherein, U1 represents the input value of the roll angle channel, U2 represents the input value of the pitch angle channel, U3 represents the input value of the yaw angle channel, b represents the anti-torque coefficient of the propeller motor, and Ω1 to Ω4 represent the rotation speeds of the four rotors; Among them, l represents the farthest distance between the arm and the center of mass of the quadrotor drone, I r The moment of inertia of the rotor rotating around the axis, k i (i=1,2,3) represent the air resistance coefficient of the rotation around the corresponding coordinate axis in the carrier coordinate system, Represents the attitude angle component of the quadrotor drone in the inertial coordinate system, V=[p,q,r] T Represents the angular velocity component of the quadrotor drone in the carrier coordinate system, P = [x, y, z] T Represents the position component of the quadrotor UAV in the carrier coordinate system, I x , I y , I z represent the moment of inertia about the x, y, and z axes respectively.

9. A quadrotor UAV attitude estimation device based on improved UKF, characterized in that: The device comprises: at least one processor; and, a memory communicatively coupled to the at least one processor; The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the quadrotor drone attitude estimation method based on the improved UKF as described in any one of claims 1 to 8.

10. A non-volatile computer storage medium storing computer executable instructions, characterized in that: When the computer-executable instructions are executed, the quadrotor drone attitude estimation method based on the improved UKF as described in any one of claims 1 to 8 is implemented.

Images