Method, device and equipment for applying improved nonlinear filtering algorithm to unmanned aerial vehicle

Abstract

The application discloses a method, device and equipment for applying an improved nonlinear filtering algorithm to a UAV, and relates to the technical field of UAV application. The method comprises the following steps: based on each sensor arranged on the upper portion of the UAV, collecting GNSS data and IMU data in real time and synchronously according to a set sampling frequency and inputting the GNSS data and the IMU data into the improved nonlinear filtering algorithm; performing preliminary fusion on the GNSS data and the IMU data through an unscented Kalman filtering algorithm, and combining UT transformation and an adaptive mechanism to obtain a preliminary state estimation value of the UAV; inputting the preliminary state estimation value into a cubature Kalman filtering algorithm, updating and predicting the state of the UAV based on a cubature rule, and obtaining final positioning and attitude estimation results and applying the final positioning and attitude estimation results to the UAV. The application can greatly improve the positioning and attitude estimation precision of the UAV.

Description

Technical Field

[0001] This application relates to the field of unmanned aerial vehicle (UAV) application technology, specifically to a method, apparatus, and equipment for applying an improved nonlinear filtering algorithm to UAVs. Background Technology

[0002] With the rapid development of drone technology, its applications in various fields are becoming increasingly widespread, and the accuracy and stability of navigation and data processing are of paramount importance. Traditional linear filtering algorithms are proving inadequate when faced with the complex nonlinear environment of drone flight. For example, the classic Kalman filter algorithm is only applicable to linear systems, and it struggles to accurately estimate the drone system state when dealing with the nonlinear characteristics exhibited by drones due to attitude changes, airflow interference, and other factors during flight.

[0003] To address this issue, the Extended Kalman Filter (EKF) was proposed. It extends Kalman filtering theory to the nonlinear domain by performing a Taylor expansion of the nonlinear function and omitting higher-order terms, retaining only first-order terms to achieve a linear approximation. However, this approximation introduces significant errors in complex environments such as high-speed UAV maneuvers and strong airflow interference, leading to inaccurate positioning and attitude estimation data. For example, in mountainous areas and other terrain-complex regions, where UAV flight states are highly variable, the linear approximation of the EKF cannot accurately track the actual movement of the UAV, significantly reducing navigation accuracy.

[0004] Meanwhile, in the field of UAV mapping, data fusion technology is of great significance for obtaining high-precision mapping results. The technology system combining multi-satellite joint positioning with inertial measurement units (IMUs) is gradually becoming mainstream. However, traditional data fusion algorithms have limitations. On the one hand, differences in time synchronization and accuracy matching between data from different sensors make it easy for errors to accumulate during data fusion. For example, the sampling frequencies of GNSS (Global Navigation Satellite System) receivers and IMUs are not synchronized, resulting in the inability to accurately correspond to position and attitude information at the same moment when fusing data. On the other hand, facing sudden changes in sensor data in complex environments, such as satellite signal obstruction causing jumps in positioning data in urban canyons, traditional algorithms struggle to process these changes quickly and effectively, affecting the accuracy and reliability of UAV mapping.

[0005] Therefore, it is evident that in the field of UAV technology, existing technologies have numerous problems in areas such as data processing and fusion, navigation and positioning, and coping with interference in complex environments. Specifically: (1) Traditional filtering algorithms are not accurate enough when dealing with nonlinear systems: Classical Kalman filtering is only applicable to linear systems and is inadequate when dealing with the nonlinear characteristics of UAVs in flight. Although Extended Kalman Filter (EKF) extends it to the nonlinear domain and approximates it through Taylor expansion, the linearization introduces a large error in complex scenarios such as high-speed maneuvering of UAVs and strong airflow interference. For example, when flying in mountainous areas, the attitude of UAVs changes frequently and EKF is difficult to track accurately, resulting in large deviations in positioning, attitude estimation and other data, which seriously affects the navigation accuracy and mission execution accuracy of UAVs. (2) Defects in multi-sensor data fusion: In the technical system of multi-satellite joint positioning and inertial measurement unit (IMU), there are problems such as time synchronization and accuracy matching of different sensor data. The sampling frequencies of GNSS receiver and IMU are not synchronized, making it difficult to correspond to the position and attitude information at the same time during fusion, resulting in error accumulation. In complex environments such as urban canyons, satellite signals are blocked, causing positioning data to jump. Traditional algorithms cannot process this quickly and effectively, reducing the accuracy and reliability of UAV mapping data. (3) Hardware equipment performance limitations and installation problems: Some GNSS receivers have poor positioning accuracy and reliability. When the signal is weak or interfered with, the positioning error increases significantly. Some low-cost IMUs have problems with high noise and poor stability, which affect the attitude measurement accuracy. In addition, if the sensor is installed in an unreasonable position, it will be too close to the center of gravity of the UAV and the vibration of the body will easily interfere with data acquisition, thereby reducing the overall performance of the UAV system. (4) Software system functionality and compatibility need to be improved: Some UAV flight control and data processing software platforms have limited functionality and lack advanced data processing and analysis functions, making it difficult to meet the needs of complex tasks. The compatibility between different software systems or modules is poor. When integrating improved nonlinear filtering algorithms, problems such as poor data transmission and interface mismatch may occur, hindering the effective application of the algorithm. Summary of the Invention

[0006] This application provides a method, apparatus, and device for applying an improved nonlinear filtering algorithm to a UAV, which can significantly improve the accuracy of UAV positioning and attitude estimation.

[0007] In a first aspect, embodiments of this application provide a method for applying an improved nonlinear filtering algorithm to a drone, the method comprising: Based on the various sensors deployed on the UAV, GNSS data and IMU data are collected in real time and synchronously according to the set sampling frequency and input into the improved nonlinear filtering algorithm; The GNSS and IMU data are initially fused using the unscented Kalman filter algorithm, and combined with UT transform and adaptive mechanism to obtain the initial state estimate of the UAV. The preliminary state estimate is input into the capacitive Kalman filter algorithm, and the UAV state is updated and predicted based on the capacitive rule calculation to obtain the final positioning and attitude estimation results, which are then applied to the UAV.

[0008] In conjunction with the first aspect, in one implementation, the real-time synchronous acquisition of GNSS data and IMU data based on the sensors deployed on the UAV according to a set sampling frequency specifically includes: Before the drone takes off, start the GNSS receiver and IMU deployed on the drone and ensure that the GNSS receiver and IMU are working properly; During the flight of the UAV, GNSS data and IMU data are collected synchronously in real time according to the set sampling frequency, and the collected GNSS data and IMU data are preprocessed. The preprocessing includes removing outliers and filling in missing values. The GNSS receiver and IMU are installed near the center of gravity of the UAV.

[0009] In conjunction with the first aspect, in one implementation method, The improved nonlinear filtering algorithms include the unscented Kalman filtering algorithm and the capacitive Kalman filtering algorithm; The unscented Kalman filter algorithm includes the definition of state variables and observation variables, UKF parameter initialization, UT transformation, and adaptive mechanism; The volumetric Kalman filter algorithm includes nonlinear model construction, CKF parameter initialization, volume rule calculation, and numerical stability optimization.

[0010] In conjunction with the first aspect, in one implementation method, In the definition of state variables and observation variables, state variables are determined according to UAV mapping requirements, and observation variables include position and velocity information output by GNSS receiver, as well as acceleration and angular velocity information measured by IMU. UAV system state variables include position, velocity, and attitude. The UKF parameters are initialized by selecting the number and weight of Sigma points, determining the distribution of Sigma points based on the dimension and characteristics of the UAV system state variables, and setting the initial state estimate and covariance matrix. The UT transformation is based on the UT transformation algorithm to calculate the predicted and updated values ​​of the Sigma point, and the state variables of the UAV system are transferred through a nonlinear function to obtain the state of the Sigma point after the prediction and update steps; The adaptive mechanism is based on an adaptive algorithm that adjusts the distribution of Sigma points in real time according to changes in the UAV system state.

[0011] In conjunction with the first aspect, in one implementation method, The nonlinear model is constructed based on the kinematics and dynamics principles of the UAV, establishing a nonlinear model that includes position, velocity, and attitude state variables; The CKF parameters are initialized by setting the initial state estimate, covariance matrix, and volume rule; The volume rule calculation is to use the volume rule and the spherical radial rule to update and predict the nonlinear model. At each time step, based on the state estimate and observation data of the previous moment, the predicted and updated values ​​of the UAV system state are calculated through the volume rule to obtain the position, velocity and attitude estimates of the current moment. The numerical stability optimization involves using a numerical stability algorithm to normalize and scale intermediate variables during the calculation process in the implementation of the capacitive Kalman filter algorithm.

[0012] In conjunction with the first aspect, in one implementation, obtaining the final positioning and attitude estimation results and applying them to the UAV specifically includes: Based on the final positioning and attitude estimation results, output data is obtained, which includes the UAV's position, velocity, attitude information, and corresponding state estimation covariance parameters. The output data is applied to mapping tasks performed by UAVs.

[0013] In conjunction with the first aspect, in one embodiment, the method of applying the improved nonlinear filtering algorithm to a UAV further includes: The performance of the improved nonlinear filtering algorithm is periodically evaluated. By comparing it with known reference data, the positioning error and attitude error are calculated to assess the accuracy and stability of the improved nonlinear filtering algorithm. The performance evaluation results are then used to optimize and improve the improved nonlinear filtering algorithm.

[0014] Secondly, embodiments of this application provide an apparatus for applying an improved nonlinear filtering algorithm to a drone, the apparatus comprising: The acquisition module is used to synchronously acquire GNSS data and IMU data in real time according to the set sampling frequency based on the various sensors deployed on the UAV and input them into the improved nonlinear filtering algorithm. The first execution module is used to perform preliminary fusion of GNSS data and IMU data through the unscented Kalman filter algorithm, and combine UT transformation and adaptive mechanism to obtain the preliminary state estimate of the UAV. The second execution module is used to input the preliminary state estimate into the capacitive Kalman filter algorithm, update and predict the UAV state based on the capacitive rule calculation, obtain the final positioning and attitude estimation results, and apply them to the UAV.

[0015] In conjunction with the second aspect, in one implementation method, The improved nonlinear filtering algorithms include the unscented Kalman filtering algorithm and the capacitive Kalman filtering algorithm; The unscented Kalman filter algorithm includes the definition of state variables and observation variables, UKF parameter initialization, UT transformation, and adaptive mechanism; The volumetric Kalman filter algorithm includes nonlinear model construction, CKF parameter initialization, volume rule calculation, and numerical stability optimization.

[0016] Thirdly, embodiments of this application provide a device for applying an improved nonlinear filtering algorithm to a drone. The device for applying the improved nonlinear filtering algorithm to a drone includes a processor, a memory, and a program for applying the improved nonlinear filtering algorithm to a drone stored in the memory and executable by the processor. When the program for applying the improved nonlinear filtering algorithm to a drone is executed by the processor, it implements the steps of the method for applying the improved nonlinear filtering algorithm to a drone as described above.

[0017] The beneficial effects of the technical solutions provided in this application include: (1) High-precision data processing: Through the improved nonlinear filtering algorithm, nonlinear systems and multi-sensor data can be processed more accurately, greatly improving the accuracy of UAV positioning and attitude estimation. In UAV mapping tasks, high-precision terrain models, orthophotos and other results can be generated to meet the needs of application scenarios with extremely high accuracy requirements such as urban planning and engineering surveying, and provide more reliable data support for related fields. (2) Strong environmental adaptability: The adaptive mechanism and advanced filtering principle introduced in the improved nonlinear filtering algorithm enable UAVs to quickly adapt to satellite signal blockage, multipath effect, airflow interference and other situations in complex environments such as urban canyons, mountains and dense forests. Even in harsh environments, it can ensure stable navigation and mapping performance, greatly expand the application range of UAVs and enhance their ability to perform tasks in complex environments. (3) High stability and reliability: Through reasonable selection and installation of hardware equipment, collaborative optimization of algorithms and continuous maintenance and upgrades, the risk of mission interruption due to data problems or positioning system failures has been reduced. The algorithm's effective suppression of sensor noise and rapid processing of data mutations have further improved reliability, reduced data loss and error accumulation, and ensured that the UAV can reliably complete various tasks. (4) Good cost-effectiveness: Improving data quality from the source reduces the workload of post-processing caused by inaccurate data, and reduces labor costs and project cycle; the efficiency of the algorithm and the stability of the positioning system reduce the rework costs caused by task failure or data error, and improve the economic benefits and efficiency of the entire surveying and mapping project; in addition, reasonable selection and optimization of the algorithm balances hardware costs and performance requirements while ensuring high performance, making the technology have good cost-effectiveness advantages. Attached Figure Description

[0018] Figure 1 This is a flowchart illustrating the method of applying the improved nonlinear filtering algorithm of this application to a UAV. Figure 2 This is a schematic diagram of the implementation process stages; Figure 3 This is a schematic diagram of the functional modules of the device for applying the improved nonlinear filtering algorithm of this application to a UAV; Figure 4 This is a schematic diagram of the hardware structure of the device for applying the improved nonlinear filtering algorithm of this application to a drone. Detailed Implementation

[0019] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present application.

[0020] To make the objectives, technical solutions, and advantages of this application clearer, the embodiments of this application will be described in further detail below with reference to the accompanying drawings.

[0021] In a first aspect, embodiments of this application provide a method for applying an improved nonlinear filtering algorithm to a drone, which is used to overcome the limitations of traditional algorithms and provide a solution for drones to achieve more accurate navigation, positioning and data processing in complex environments.

[0022] In one embodiment, reference is made to Figure 1 , Figure 1 This is a flowchart illustrating the method for applying the improved nonlinear filtering algorithm of this application to a UAV. Figure 1 As shown, the methods for applying the improved nonlinear filtering algorithm to UAVs include: S1: Based on the sensors deployed on the UAV, GNSS data and IMU data are collected in real time and synchronously according to the set sampling frequency and input into the improved nonlinear filtering algorithm; S2: The GNSS and IMU data are initially fused using the unscented Kalman filter algorithm, and combined with the UT transform and adaptive mechanism to obtain the preliminary state estimate of the UAV. The UT transform is a mathematical tool for processing the state estimate of nonlinear systems. It approximates the probability distribution of the original state by selecting Sigma points, and uses these points to propagate through a nonlinear function to finally update the predicted mean and covariance. S3: Input the preliminary state estimate into the capacitive Kalman filter algorithm, update and predict the UAV state based on the capacitive rule calculation, obtain the final positioning and attitude estimation results, and apply them to the UAV.

[0023] Furthermore, in one embodiment, based on the sensors deployed on the UAV, GNSS data and IMU data are collected synchronously in real time according to a set sampling frequency, specifically including: S101: Before the drone takes off, start the GNSS receiver and IMU deployed on the drone and ensure that the GNSS receiver and IMU are working properly; S102: During the flight of the UAV, GNSS data and IMU data are collected synchronously in real time according to the set sampling frequency, and the collected GNSS data and IMU data are preprocessed. The preprocessing includes removing outliers and filling in missing values. The GNSS receiver and IMU are installed near the center of gravity of the UAV.

[0024] For details, see Figure 2 As shown, in the implementation process, before the UAV takes off, the GNSS receiver and IMU are started to ensure that the sensors are working properly and begin collecting data. During the data collection process, GNSS data and IMU data are acquired synchronously according to the set sampling frequency. Then, the collected data is preprocessed, including removing outliers and filling in missing values. For GNSS data, valid data is selected based on satellite signal quality indicators; for IMU data, high-frequency noise is removed through filtering and other methods to improve data quality.

[0025] Preprocessed GNSS and IMU data are input into a deployed improved nonlinear filtering algorithm. First, the UKF algorithm (Unscented Kalman Filter) is used for initial data fusion. Based on the UT transform and adaptive mechanism, preliminary position, velocity, and attitude estimates are obtained. Then, the output of the UKF algorithm is used as input to the CKF algorithm (Vacuum-Capacitive Kalman Filter). The CKF algorithm further fuses the data, and through the volume rule calculation of the CKF algorithm, the state of the UAV system (i.e., the overall UAV) is updated and predicted more accurately, resulting in the final high-precision positioning and attitude estimation results. During the data fusion process, the algorithm's operating status and data processing results are monitored in real time. If data anomalies or slow algorithm convergence are detected, algorithm parameters are adjusted promptly or corresponding processing measures are taken.

[0026] The following describes the preliminary preparation stage and algorithm deployment stage of the method for applying the improved nonlinear filtering algorithm of this application to UAVs.

[0027] For the hardware selection and installation in the preliminary preparation stage: Select a GNSS receiver that supports multi-satellite joint positioning, ensuring it can simultaneously receive signals from multiple satellite navigation systems such as GPS, BeiDou, GLONASS, and GALILEO, improving positioning reliability and accuracy. Simultaneously, select a receiver with an appropriate accuracy level based on the UAV's payload capacity and mapping requirements. Equip the UAV with a high-performance inertial measurement unit (IMU), prioritizing products based on MEMS technology that feature high precision, low noise, and high stability. Ensure the IMU's sampling frequency matches the GNSS receiver's data output frequency to guarantee data synchronization. Correctly install the selected GNSS receiver and IMU on the UAV, ensuring the installation location is as close as possible to the UAV's center of gravity to minimize the impact of airframe vibrations on sensor data acquisition. Also, take proper measures to secure and protect the sensors to prevent loosening or damage during flight.

[0028] For the initial preparation phase, the software environment setup involves: building a UAV flight control and data processing software platform, selecting open-source or commercial software that supports secondary development, and ensuring it has data acquisition, transmission, storage, and basic data processing functions; installing relevant development tools and libraries, such as the Python language environment and scientific computing libraries like NumPy and SciPy, to implement improved nonlinear filtering algorithms; if using other programming languages ​​such as C++, installing the corresponding development environment and related libraries to support algorithm writing and debugging; and establishing a data storage system to store the raw GNSS and IMU data collected during UAV flight, as well as the filtered data, selecting appropriate data storage formats such as CSV and binary files to ensure data storage efficiency and readability.

[0029] It should be noted that the improved nonlinear filtering algorithms include the unscented Kalman filter algorithm and the capacitive Kalman filter algorithm; the unscented Kalman filter algorithm includes the definition of state variables and observation variables, UKF parameter initialization, UT transformation, and adaptive mechanism; the capacitive Kalman filter algorithm includes nonlinear model construction, CKF parameter initialization, capacitive rule calculation, and numerical stability optimization.

[0030] Specifically, the deployment of the unscented Kalman filter algorithm in the algorithm deployment phase includes defining state and observation variables, initializing UKF parameters, UT transformation, and adaptive mechanisms. The deployment of the volumetric Kalman filter algorithm in the algorithm deployment phase includes nonlinear model construction, CKF parameter initialization, volume rule calculation, and numerical stability optimization.

[0031] In the definition of state variables and observation variables, state variables are determined according to the UAV mapping requirements. Observation variables include position and velocity information output by the GNSS receiver, as well as acceleration and angular velocity information measured by the IMU. UAV system state variables include position, velocity, attitude (pitch angle, roll angle, yaw angle), etc.

[0032] The UKF parameters are initialized by selecting the number and weight of Sigma points, determining the distribution of Sigma points based on the dimension and characteristics of the UAV system state variables, and setting the initial state estimates and covariance matrix.

[0033] Specifically, the number and weight of Sigma points should be reasonably selected, and the distribution of Sigma points should be determined according to the dimension and characteristics of the UAV system state variables. The initial state estimate and covariance matrix should be set. The initial state estimate can be set according to the initial position and attitude of the UAV, and the covariance matrix reflects the uncertainty of the initial state estimate.

[0034] The UT transform is based on the UT transform algorithm to calculate the predicted and updated values ​​of the Sigma point. The state variables of the UAV system are transferred through a nonlinear function to obtain the state of the Sigma point after the prediction and update steps.

[0035] Specifically, the predicted and updated values ​​of the Sigma point are calculated using the UT transform. The UAV system state variables are transferred through a nonlinear function to obtain the state of the Sigma point after the prediction and update steps. During the calculation process, the mean and covariance of the Sigma point are accurately calculated based on the UAV's motion model and observation model.

[0036] The adaptive mechanism is based on an adaptive algorithm that adjusts the distribution of Sigma points in real time according to changes in the UAV system state.

[0037] Specifically, an adaptive algorithm is designed to adjust the distribution of Sigma points in real time according to changes in the UAV system state. By monitoring the UAV's flight state parameters, such as the magnitude of changes in acceleration and angular velocity, as well as GNSS signal quality indicators (such as the number of satellites and signal strength), the weight and number of Sigma points are dynamically adjusted to improve the algorithm's adaptability to dynamic environments.

[0038] The nonlinear model is constructed based on the kinematics and dynamics principles of the UAV, establishing a nonlinear model that includes position, velocity, and attitude state variables.

[0039] Specifically, based on the kinematics and dynamics principles of UAVs, a nonlinear model incorporating state variables such as position, velocity, and attitude is established. Various factors during UAV flight, such as air resistance and wind influence, are considered to ensure that the nonlinear model accurately describes the actual motion state of the UAV.

[0040] The CKF parameters are initialized to set the initial state estimate, covariance matrix, and volume rule.

[0041] Specifically, the initial state estimate, covariance matrix, and volume rule parameters are set. The initial state estimate and covariance matrix are set similarly to those in the UKF algorithm, while the volume rule parameters are configured appropriately according to the algorithm requirements to ensure the stability and accuracy of the algorithm.

[0042] The volume rule calculation utilizes the volume rule and the spherical radial rule to update and predict the nonlinear model. At each time step, based on the state estimate and observation data of the previous moment, the predicted and updated values ​​of the UAV system state are calculated through the volume rule to obtain the position, velocity and attitude estimates of the current moment.

[0043] Numerical stability optimization involves using a numerical stabilization algorithm to normalize and scale intermediate variables during the computation process in the implementation of the capacitive Kalman filter algorithm.

[0044] Specifically, in the algorithm implementation process, numerically stable calculation methods are adopted to avoid numerical overflow or instability during the calculation process. Intermediate variables in the calculation process are reasonably normalized and scaled to improve the reliability of the algorithm.

[0045] Furthermore, in one embodiment, obtaining the final positioning and attitude estimation results and applying them to the UAV specifically includes: S301: Based on the final positioning and attitude estimation results, output data is obtained, which includes the UAV's position, velocity, attitude information, and corresponding state estimation covariance parameters; S302: Apply the output data to the mapping task of the UAV.

[0046] Specifically, in the data output and application stage, the data, after being fused and processed by an improved nonlinear filtering algorithm, is output. The output data includes the UAV's precise position, velocity, attitude, and related parameters such as state estimation covariance. The output data is then applied to UAV mapping tasks, such as generating high-precision terrain models and orthophotos. Depending on different application requirements, the output data undergoes further processing and analysis to provide reliable data support for fields such as urban planning and engineering surveying.

[0047] Furthermore, in one embodiment, the method of applying the improved nonlinear filtering algorithm to a UAV further includes: The performance of the improved nonlinear filtering algorithm is periodically evaluated. By comparing it with known reference data, the positioning error and attitude error are calculated to assess the accuracy and stability of the improved nonlinear filtering algorithm. The performance evaluation results are then used to optimize and improve the improved nonlinear filtering algorithm.

[0048] Specifically, this also includes the subsequent optimization phase. For the performance evaluation of the algorithm in this phase, the performance of the improved nonlinear filtering algorithm is periodically assessed. By comparing it with known high-precision reference data, indicators such as positioning error and attitude error are calculated to evaluate the algorithm's accuracy and stability. The algorithm's performance under different environmental conditions (such as urban canyons, mountainous areas, and open areas) and different flight states (such as uniform flight, accelerated flight, and turning flight) is analyzed to identify its strengths and weaknesses.

[0049] For subsequent optimization and improvement of the algorithm, based on the performance evaluation results, targeted optimizations are performed on the improved nonlinear filtering algorithm. For cases with significant errors in certain environments, algorithm parameters are adjusted or the algorithm structure is improved, such as further optimizing the adaptive mechanism of the UKF algorithm or improving the volume rule calculation method of the CKF algorithm. We pay close attention to the technological developments in related fields and promptly apply new research results to the improved nonlinear filtering algorithm to continuously improve its performance and adaptability. For example, we can introduce new nonlinear filtering theories or combine machine learning algorithms to further optimize the data fusion process.

[0050] For subsequent optimization phases of system maintenance and upgrades, regular inspections and maintenance of the UAV's hardware will be conducted to ensure the normal operation of sensors such as the GNSS receiver and IMU. Aging or damaged equipment will be replaced promptly to guarantee the accuracy and reliability of data acquisition. The software system will be upgraded and updated, software vulnerabilities will be fixed, data processing flows will be optimized, and system stability and operational efficiency will be improved. Simultaneously, based on actual application needs, software functionality will be continuously improved to provide better support for the application of improved nonlinear filtering algorithms.

[0051] Specifically, for the method of applying the improved nonlinear filtering algorithm of this application to UAVs, the complete implementation process includes the preliminary preparation stage, the algorithm deployment stage, the implementation process stage, and the subsequent optimization stage.

[0052] This application effectively solves existing problems in the navigation, positioning, and surveying fields of UAVs by focusing on multiple dimensions such as algorithms, data processing, and system optimization, and has significant advantages, specifically: (1) Overcoming the accuracy problem of traditional filtering algorithms: In view of the problem that traditional filtering algorithms are not accurate enough when dealing with nonlinear systems of UAVs, the unscented Kalman filter algorithm and the capacitive Kalman filter algorithm are adopted. The UKF algorithm directly processes nonlinear functions through UT transformation, avoiding the error caused by the linearization approximation of the extended Kalman filter; the CKF algorithm is based on unique capacitive rules and spherical radial rules, which have better numerical stability and accuracy when dealing with nonlinear systems; in complex nonlinear environments such as high-speed maneuvering of UAVs and encountering strong airflow, the two work together to accurately describe the flight state of UAVs and realize accurate estimation of their position, speed and attitude, effectively improving the accuracy of positioning and attitude estimation. (2) Optimize the defects of multi-sensor data fusion: In order to address the problems of time asynchrony and accuracy mismatch in multi-sensor data fusion, when selecting hardware, ensure that the sampling frequency of IMU and GNSS receiver is consistent; in the data acquisition and preprocessing stage, strictly screen the raw data, remove outliers and fill in missing values; in the data fusion processing stage, first use the UKF algorithm to perform preliminary data fusion, and then use the CKF algorithm for further precise processing, give full play to the advantages of the two algorithms, reduce error accumulation, improve the accuracy and reliability of data fusion, and effectively cope with the situation of sensor data mutation in complex environment; (3) Make up for the shortcomings in hardware performance and installation: By carefully selecting high-precision GNSS receivers and high-performance MEMS-IMUs that support multi-satellite joint positioning, the accuracy and reliability of data acquisition can be improved from the source; at the same time, the sensors are installed reasonably, placed as close as possible to the center of gravity of the UAV, and fixed and protected measures are taken to reduce the interference of the body vibration on data acquisition, avoid data errors caused by poor hardware performance or improper installation, and ensure that the UAV can collect data stably and accurately. (4) Improve the functionality and compatibility of the software system: Build a UAV flight control and data processing software platform that supports secondary development, ensuring that it has data acquisition, transmission, storage and basic processing functions; install suitable development tools and library functions to support the writing and debugging of improved nonlinear filtering algorithms; in practical applications, the software system can be upgraded and updated according to needs, vulnerabilities can be fixed, processes can be optimized, system stability and operating efficiency can be enhanced, and the problems of insufficient software functionality and poor compatibility can be effectively solved.

[0053] Secondly, embodiments of this application also provide a device for applying an improved nonlinear filtering algorithm to a drone.

[0054] In one embodiment, reference is made to Figure 3 , Figure 3 This is a schematic diagram of the functional modules of a device for applying the improved nonlinear filtering algorithm of this application to a UAV. For example... Figure 3 As shown, the device for applying the improved nonlinear filtering algorithm to a UAV includes: a data acquisition module, a first execution module, and a second execution module.

[0055] The acquisition module is used to synchronously acquire GNSS data and IMU data in real time according to a set sampling frequency based on the various sensors deployed on the UAV and input them into an improved nonlinear filtering algorithm; the first execution module is used to perform preliminary fusion of GNSS data and IMU data through an unscented Kalman filter algorithm, and combine UT transform and adaptive mechanism to obtain a preliminary state estimate of the UAV; the second execution module is used to input the preliminary state estimate into a volumetric Kalman filter algorithm, and update and predict the UAV state based on the volume rule calculation to obtain the final positioning and attitude estimation results and apply them to the UAV.

[0056] In this application, the improved nonlinear filtering algorithm includes an unscented Kalman filter algorithm and a capacitive Kalman filter algorithm; the unscented Kalman filter algorithm includes the definition of state variables and observation variables, UKF parameter initialization, UT transformation, and adaptive mechanism; the capacitive Kalman filter algorithm includes nonlinear model construction, CKF parameter initialization, capacitive rule calculation, and numerical stability optimization.

[0057] Thirdly, this application provides an improved nonlinear filtering algorithm applied to a device for drones. The device for applying the improved nonlinear filtering algorithm to drones can be a personal computer (PC), laptop computer, server, or other device with data processing capabilities.

[0058] Reference Figure 4 , Figure 4 This is a schematic diagram of the hardware structure of a device for applying the improved nonlinear filtering algorithm involved in the embodiments of this application to a drone. In the embodiments of this application, the device for applying the improved nonlinear filtering algorithm to a drone may include a processor, memory, communication interface, and communication bus.

[0059] The communication bus can be of any type and is used to interconnect the processor, memory, and communication interface.

[0060] The communication interface includes input / output (I / O) interfaces, physical interfaces, and logical interfaces. These interfaces are used for interconnecting internal devices within the UAV to implement the improved nonlinear filtering algorithm, and for interconnecting the device with other devices (such as other computing devices or user equipment). Physical interfaces can be Ethernet interfaces, fiber optic interfaces, ATM interfaces, etc.; user equipment can be displays, keyboards, etc.

[0061] Memory can be various types of storage media, such as random access memory (RAM), read-only memory (ROM), non-volatile RAM (NVRAM), flash memory, optical storage, hard disk, programmable ROM (PROM), erasable PROM (EPROM), electrically erasable PROM (EEPROM), etc.

[0062] The processor can be a general-purpose processor, which can call the program for applying the improved nonlinear filtering algorithm to the UAV stored in memory and execute the method for applying the improved nonlinear filtering algorithm to the UAV provided in the embodiments of this application. For example, the general-purpose processor can be a central processing unit (CPU). The method executed when the program for applying the improved nonlinear filtering algorithm to the UAV is called can refer to the various embodiments of the method for applying the improved nonlinear filtering algorithm to the UAV in this application, and will not be repeated here.

[0063] Those skilled in the art will understand that Figure 4 The hardware structure shown does not constitute a limitation of this application and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0064] Fourthly, embodiments of this application also provide a computer-readable storage medium.

[0065] The present application stores a program on a computer-readable storage medium for applying an improved nonlinear filtering algorithm to a drone, wherein when the program for applying the improved nonlinear filtering algorithm to a drone is executed by a processor, the steps of the method for applying the improved nonlinear filtering algorithm to a drone as described above are implemented.

[0066] The method for implementing the improved nonlinear filtering algorithm when the program for the UAV is executed can be referred to in the various embodiments of the method for applying the improved nonlinear filtering algorithm to the UAV in this application, and will not be repeated here.

[0067] The terms "comprising" and "having," and any variations thereof, in the specification, claims, and accompanying drawings of this application are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to such process, method, product, or apparatus. The terms "first," "second," and "third," etc., are used to distinguish different objects, etc., and do not indicate a sequence, nor do they limit "first," "second," and "third" to different types.

[0068] In the description of the embodiments of this application, terms such as "exemplary," "for example," or "for instance" are used to indicate examples, illustrations, or explanations. Any embodiment or design described as "exemplary," "for example," or "for instance" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of terms such as "exemplary," "for example," or "for instance" is intended to present the relevant concepts in a concrete manner.

[0069] In the description of the embodiments of this application, unless otherwise stated, " / " means "or". For example, A / B can mean A or B. The "and / or" in the text is merely a description of the relationship between related objects, indicating that there can be three relationships. For example, A and / or B can mean: A exists alone, A and B exist simultaneously, and B exists alone. In addition, in the description of the embodiments of this application, "multiple" means two or more.

[0070] In some processes described in the embodiments of this application, multiple operations or steps are included in a specific order. However, it should be understood that these operations or steps may not be executed in the order they appear in the embodiments of this application, or they may be executed in parallel. The sequence number of the operation is only used to distinguish different operations, and the sequence number itself does not represent any execution order. In addition, these processes may include more or fewer operations, and these operations or steps may be executed sequentially or in parallel, and these operations or steps may be combined.

[0071] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes several instructions to cause a terminal device to execute the methods described in the various embodiments of this application.

[0072] The above are merely preferred embodiments of this application and do not limit the patent scope of this application. Any equivalent structural or procedural transformations made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

Claims

1. A method for applying an improved nonlinear filtering algorithm to a UAV, characterized in that, The improved nonlinear filtering algorithm is applied to UAVs in the following ways: Based on the various sensors deployed on the UAV, GNSS data and IMU data are collected in real time and synchronously according to the set sampling frequency and input into the improved nonlinear filtering algorithm; The GNSS and IMU data are initially fused using the unscented Kalman filter algorithm, and combined with UT transform and adaptive mechanism to obtain the initial state estimate of the UAV. The preliminary state estimate is input into the capacitive Kalman filter algorithm, and the UAV state is updated and predicted based on the capacitive rule calculation to obtain the final positioning and attitude estimation results and apply them to the UAV. The improved nonlinear filtering algorithm includes the unscented Kalman filtering algorithm and the commutative Kalman filtering algorithm; The unscented Kalman filter algorithm includes the definition of state variables and observation variables, UKF parameter initialization, UT transformation, and adaptive mechanism; The capacitive Kalman filter algorithm includes nonlinear model construction, CKF parameter initialization, capacitive rule calculation, and numerical stability optimization. The nonlinear model is constructed based on the kinematics and dynamics principles of the UAV, establishing a nonlinear model that includes position, velocity, and attitude state variables. The CKF parameters are initialized by setting the initial state estimate, covariance matrix, and volume rule; The volume rule calculation is to use the volume rule and the spherical radial rule to update and predict the nonlinear model. At each time step, based on the state estimate and observation data of the previous moment, the predicted and updated values ​​of the UAV system state are calculated through the volume rule to obtain the position, velocity and attitude estimates of the current moment. The numerical stability optimization involves using a numerical stability algorithm to normalize and scale intermediate variables during the calculation process in the implementation of the capacitive Kalman filter algorithm.

2. The method for applying the improved nonlinear filtering algorithm to a UAV as described in claim 1, characterized in that, The aforementioned method, based on sensors deployed on the UAV, synchronously collects GNSS and IMU data in real time according to a set sampling frequency, specifically including: Before the drone takes off, start the GNSS receiver and IMU deployed on the drone and ensure that the GNSS receiver and IMU are working properly; During the flight of the UAV, GNSS data and IMU data are collected synchronously in real time according to the set sampling frequency, and the collected GNSS data and IMU data are preprocessed. The preprocessing includes removing outliers and filling in missing values. The GNSS receiver and IMU are installed near the center of gravity of the UAV.

3. The method for applying the improved nonlinear filtering algorithm to a UAV as described in claim 1, characterized in that: In the definition of state variables and observation variables, state variables are determined according to UAV mapping requirements, and observation variables include position and velocity information output by GNSS receiver, as well as acceleration and angular velocity information measured by IMU. UAV system state variables include position, velocity, and attitude. The UKF parameters are initialized by selecting the number and weight of Sigma points, determining the distribution of Sigma points based on the dimension and characteristics of the UAV system state variables, and setting the initial state estimate and covariance matrix. The UT transformation is based on the UT transformation algorithm to calculate the predicted and updated values ​​of the Sigma point, and the state variables of the UAV system are transferred through a nonlinear function to obtain the state of the Sigma point after the prediction and update steps; The adaptive mechanism is based on an adaptive algorithm that adjusts the distribution of Sigma points in real time according to changes in the UAV system state.

4. The method for applying the improved nonlinear filtering algorithm to a UAV as described in claim 1, characterized in that, The process of obtaining the final positioning and attitude estimation results and applying them to the UAV specifically includes: Based on the final positioning and attitude estimation results, output data is obtained, which includes the UAV's position, velocity, attitude information, and corresponding state estimation covariance parameters. The output data is applied to mapping tasks performed by UAVs.

5. The method for applying the improved nonlinear filtering algorithm as described in claim 1 to a UAV, characterized in that, The method of applying the improved nonlinear filtering algorithm to UAVs also includes: The performance of the improved nonlinear filtering algorithm is periodically evaluated. By comparing it with known reference data, the positioning error and attitude error are calculated to assess the accuracy and stability of the improved nonlinear filtering algorithm. The performance evaluation results are then used to optimize and improve the improved nonlinear filtering algorithm.

6. A device for applying an improved nonlinear filtering algorithm to a UAV, characterized in that, The device for applying the improved nonlinear filtering algorithm to a UAV includes: The acquisition module is used to synchronously acquire GNSS data and IMU data in real time according to the set sampling frequency based on the various sensors deployed on the UAV and input them into the improved nonlinear filtering algorithm. The first execution module is used to perform preliminary fusion of GNSS data and IMU data through the unscented Kalman filter algorithm, and combine UT transformation and adaptive mechanism to obtain the preliminary state estimate of the UAV. The second execution module is used to input the preliminary state estimate into the capacitive Kalman filter algorithm, update and predict the UAV state based on the capacitive rule calculation, obtain the final positioning and attitude estimation results, and apply them to the UAV. The improved nonlinear filtering algorithm includes the unscented Kalman filtering algorithm and the commutative Kalman filtering algorithm; The unscented Kalman filter algorithm includes the definition of state variables and observation variables, UKF parameter initialization, UT transformation, and adaptive mechanism; The capacitive Kalman filter algorithm includes nonlinear model construction, CKF parameter initialization, capacitive rule calculation, and numerical stability optimization. The nonlinear model is constructed based on the kinematics and dynamics principles of the UAV, establishing a nonlinear model that includes position, velocity, and attitude state variables. The CKF parameters are initialized by setting the initial state estimate, covariance matrix, and volume rule; The volume rule calculation is to use the volume rule and the spherical radial rule to update and predict the nonlinear model. At each time step, based on the state estimate and observation data of the previous moment, the predicted and updated values ​​of the UAV system state are calculated through the volume rule to obtain the position, velocity and attitude estimates of the current moment. The numerical stability optimization involves using a numerical stability algorithm to normalize and scale intermediate variables during the calculation process in the implementation of the capacitive Kalman filter algorithm.

7. An improved nonlinear filtering algorithm applied to a UAV, characterized in that, The device for applying the improved nonlinear filtering algorithm to a drone includes a processor, a memory, and a program for applying the improved nonlinear filtering algorithm to the drone stored in the memory and executable by the processor. When the program for applying the improved nonlinear filtering algorithm to the drone is executed by the processor, it implements the steps of the method for applying the improved nonlinear filtering algorithm to a drone as described in any one of claims 1 to 5.

Images