Unmanned aerial vehicle position prediction method and system based on trusted UKF-BiLSTM

Abstract

The application discloses a UAV position prediction method and system based on a credible UKF-BiLSTM, relates to the technical field of UAV position prediction, and the UAV position prediction method provided by the application utilizes BiLSTM to obtain a nonlinear transfer function when performing prediction, can better describe the nonlinear transfer process of UAV position change, combines UKF to track and predict the UAV position, sets a related loss function according to the credibility, makes the optimization direction of the BiLSTM always consistent with the increasing direction of the credibility, and can reflect the matching of the model in a relatively real-time and observable manner. Compared with traditional UAV prediction algorithms, the application has better precision and interpretability.

Description

Technical Field

[0001] This invention relates to the field of unmanned aerial vehicle (UAV) location prediction technology, specifically to a UAV location prediction method and system based on Trusted UKF-BiLSTM. Background Technology

[0002] When drones are in operation, they are inevitably affected by the external environment. Factors such as wind and cyclones can cause the drone's trajectory to deviate. Such deviations can have a certain impact on the normal flight and operation of the drone.

[0003] In the process of using Kalman filtering to track and predict UAVs, due to the influence of various external environments, it is impossible to accurately model the position changes of the UAV. At the same time, the motion of the UAV is nonlinear and requires approximation techniques. All of these factors can lead to a mismatch between the UAV position change model and the actual model, which seriously affects the performance of the filter.

[0004] Most existing algorithms addressing this phenomenon utilize adaptive techniques to estimate inaccurate model parameters. However, these estimation methods often only estimate a single inaccurate parameter. For example, Sage-Husa filtering inevitably exhibits divergence when simultaneously estimating the covariance of process noise and measurement noise, significantly limiting the effectiveness of model parameter estimation. Furthermore, most adaptive filters suffer from poor interpretability, failing to accurately explain how they improve model fit at the model level, and cannot accurately quantify model fit.

[0005] Nowadays, with the continuous deepening of sky exploration, the prediction algorithm for drones is also developing rapidly. Accurate tracking and prediction of drones is an important guarantee for the successful completion of drone missions. Therefore, the current research on drone position prediction continues, with the hope of further improving prediction accuracy. Summary of the Invention

[0006] To address the shortcomings mentioned in the background section, the present invention aims to provide a method and system for UAV location prediction based on trusted UKF-BiLSTM.

[0007] Firstly, the objective of this invention can be achieved through the following technical solution: a UAV location prediction method based on trusted UKF-BiLSTM, the method comprising the following steps:

[0008] Acquire UAV position data, input the UAV position data into a pre-established BiLSTM model for training, and output the UAV position change model.

[0009] The system acquires observations of the UAV's position, inputs these observations into the UAV position change model to predict the UAV's position change, and outputs the UAV position prediction result.

[0010] In conjunction with the first aspect, in some implementations of the first aspect, the method further includes: the process of acquiring the UAV location data includes:

[0011] It is equipped with a drone that supports RTK technology and a ground reference station, with the ground reference station serving as a fixed reference point for the RTK system.

[0012] Configure the RTK base station and connect the RTK receiver on the drone to the base station via radio or network to enable communication between the devices.

[0013] During the flight of the drone, position data and other sensor data are stored in the drone's internal memory or transmitted wirelessly to the ground station for storage in real time.

[0014] In conjunction with the first aspect, in some implementations of the first aspect, the method further includes: inputting the UAV position data into a pre-established BiLSTM model for training, and outputting the calculation formula of the UAV position change model as follows:

[0015] X i,k|k-1 =BiLSTM(X) i,k-1 )

[0016] In the formula, X i,k-1 Let X be the Sigma sampling point at time k-1. i,k|k-1 X after nonlinear transfer i,k-1 .

[0017] In conjunction with the first aspect, in some implementations of the first aspect, the method further includes: the process of inputting the observed values ​​of the UAV's position into the UAV position change model to predict the UAV's position change:

[0018] Nonlinear transfer:

[0019]

[0020] Depend on and X is obtained by performing Sigma sampling i,k-1 :

[0021]

[0022] The result of the Sigma point set after nonlinear propagation is as follows:

[0023]

[0024] Update the filter gain, estimate, and error covariance:

[0025]

[0026]

[0027] In the formula, the number of sampling points is 2n+1. Calculate the error covariance for the filter. To calculate the measurement error covariance, ω is the inertia weight. Here, λ represents the process noise covariance assumed by the filter; h(·) is the scaling factor; and K is the measurement function. The cross-variance of prior error and measurement error is denoted as .

[0028] In conjunction with the first aspect, in some implementations of the first aspect, the method also includes: the credibility calculation process and formula as follows:

[0029]

[0030] The credibility level is set to:

[0031]

[0032] Set the loss function to

[0033] Secondly, in order to achieve the above objectives, this invention discloses a UAV position prediction system based on Trusted UKF-BiLSTM, comprising:

[0034] The model training module is used to acquire UAV position data, input the UAV position data into a pre-built BiLSTM model for training, and output the UAV position change model.

[0035] The position prediction module is used to acquire the observed value of the UAV's position, input the observed value of the UAV's position into the UAV position change model to predict the change of the UAV's position, and output the UAV position prediction result.

[0036] In conjunction with the second aspect, in some implementations of the second aspect, the system further includes: the process of acquiring UAV position data for the model training module includes:

[0037] It is equipped with a drone that supports RTK technology and a ground reference station, with the ground reference station serving as a fixed reference point for the RTK system.

[0038] Configure the RTK base station and connect the RTK receiver on the drone to the base station via radio or network to enable communication between the devices.

[0039] During the flight of the drone, position data and other sensor data are stored in the drone's internal memory or transmitted wirelessly to the ground station for storage in real time.

[0040] The model training module inputs the UAV position data into a pre-built BiLSTM model for training, and outputs the calculation formula for the UAV position change model as follows:

[0041] X i,k|k-1 =BiLSTM(X) i,k-1 )

[0042] In the formula, X i,k-1 Let X be the Sigma sampling point at time k-1. i,k|k-1 X after nonlinear transfer i,k-1 ;

[0043] The position prediction module inputs the observed position values ​​of the UAV into the UAV position change model to predict the UAV position change:

[0044] Nonlinear transfer:

[0045]

[0046] Depend on and X is obtained by performing Sigma sampling i,k-1 :

[0047]

[0048] The result of the Sigma point set after nonlinear propagation is as follows:

[0049]

[0050] Update the filter gain, estimate, and error covariance:

[0051]

[0052] In the formula, the number of sampling points is 2n+1. Calculate the error covariance for the filter. To calculate the measurement error covariance, ω is the inertia weight. Here, λ represents the process noise covariance assumed by the filter; h(·) is the scaling factor; and K is the measurement function. The cross-covariance of prior error and measurement error;

[0053] The process and formula for solving the confidence level of the location prediction module are as follows:

[0054]

[0055] The credibility level is set to:

[0056]

[0057] Set the loss function to

[0058] In another aspect of the present invention, in order to achieve the above objectives, a terminal device is disclosed, including a memory, a processor, and a computer program stored in the memory and capable of running on the processor. The memory stores the computer program capable of running on the processor. When the processor loads and executes the computer program, it employs the UAV position prediction method based on Trusted UKF-BiLSTM as described above.

[0059] In another aspect of the present invention, in order to achieve the above objective, a computer-readable storage medium is disclosed, wherein a computer program is stored in the computer-readable storage medium, characterized in that, when the computer program is loaded and executed by a processor, it employs the UAV position prediction method based on Trusted UKF-BiLSTM as described above.

[0060] The beneficial effects of this invention are:

[0061] The UAV position prediction method proposed in this invention utilizes BiLSTM to obtain the nonlinear transfer function during prediction, which can better describe the nonlinear transfer process of UAV position changes. It combines UKF (Uniform Keyhole Function) for UAV position tracking and prediction, and sets a relevant loss function based on confidence level, ensuring that the optimization direction of BiLSTM always aligns with the direction of increasing confidence. Simultaneously, it can reflect the model's matching accuracy in a relatively real-time and observable manner. Compared with traditional UAV prediction algorithms, it has better accuracy and interpretability. Attached Figure Description

[0062] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0063] Figure 1 This is a schematic diagram of the method flow of the present invention;

[0064] Figure 2 This is the first BiLSTM mechanism diagram used in this invention;

[0065] Figure 3 This is the second BiLSTM mechanism diagram used in this invention;

[0066] Figure 4 This is a schematic diagram of the BiLSTM-UKF calculation process of the present invention;

[0067] Figure 5 This is a schematic diagram of the system structure of the present invention;

[0068] Figure 6 This is a schematic diagram of the average error of the filtering algorithm in the simulation experiment of this embodiment;

[0069] Figure 7 This is a schematic diagram illustrating the credibility of the filtering model in the simulation experiment of this embodiment. Detailed Implementation

[0070] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0071] Example 1:

[0072] The following is a description of the relevant terms used in the embodiments of this application:

[0073] Unmanned aerial vehicles (UAVs) are unmanned aircraft controlled by radio remote control equipment and their own program control devices, or operated autonomously by an onboard computer, either completely or intermittently.

[0074] Compared to manned aircraft, drones are often better suited for tasks that are too "dull, dirty, or dangerous." Drones can be categorized into military and civilian applications. In the military field, drones are divided into reconnaissance drones and target drones. In the civilian field, drones combined with industry applications represent the true necessity of drones; their applications in aerial photography, agriculture, plant protection, miniature selfies, express delivery, disaster relief, wildlife observation, infectious disease monitoring, surveying, news reporting, power line inspection, disaster relief, film and television production, and creating romantic moments, among others, have greatly expanded the uses of drones. Developed countries are also actively expanding industry applications and developing drone technology.

[0075] like Figure 1 As shown, the UAV position prediction method based on Trusted UKF-BiLSTM is characterized by the following steps:

[0076] S101: Acquire UAV position data, input the UAV position data into a pre-established BiLSTM model for training, and output the UAV position change model.

[0077] The process of collecting the UAV location data includes:

[0078] It is equipped with a drone that supports RTK technology and a ground reference station, with the ground reference station serving as a fixed reference point for the RTK system.

[0079] Configure the RTK base station and connect the RTK receiver on the drone to the base station via radio or network to enable communication between the devices.

[0080] During the flight of the drone, position data and other sensor data are stored in the drone's internal memory or transmitted wirelessly to the ground station for storage in real time.

[0081] like Figure 2 As shown, the UAV position data is input into a pre-built BiLSTM model for training, and the calculation formula for the UAV position change model is output as follows:

[0082] X i,k|k-1 =BiLSTM(X) i,k-1 )

[0083] In the formula, X i,k-1 Let X be the Sigma sampling point at time k-1. i,k|k-1 X after nonlinear transfer i,k-1 .

[0084] S102: Obtain the observed value of the UAV's position, input the observed value of the UAV's position into the UAV position change model to predict the UAV's position change, and output the UAV position prediction result.

[0085] The process of collecting observations of the UAV's location includes:

[0086] like Figure 4 As shown, the process of inputting the observed values ​​of the UAV's position into the UAV position change model to predict the UAV's position change is as follows:

[0087] Nonlinear transfer:

[0088]

[0089] Depend on and X is obtained by performing Sigma sampling i,k-1 :

[0090]

[0091] The result of the Sigma point set after nonlinear propagation is as follows:

[0092]

[0093] Update the filter gain, estimate, and error covariance:

[0094]

[0095] In the formula, the number of sampling points is 2n+1. Calculate the error covariance for the filter. To calculate the measurement error covariance, ω is the inertia weight. Here, λ represents the process noise covariance assumed by the filter; h(·) is the scaling factor; and K is the measurement function. The cross-variance of prior error and measurement error is denoted as .

[0096] The process and formula for determining the credibility are as follows:

[0097]

[0098] The credibility level is set to:

[0099]

[0100] To ensure that the optimization direction of the model is consistent with the direction of increasing credibility, the loss function is set to... The second application of BiLSTM, such as Figure 3 As shown.

[0101] Specifically, the present invention will be further illustrated below through embodiments:

[0102] like Figure 6 and Figure 7 As shown, the implementation process of the UAV position prediction method using Trusted UKF-BiLSTM is as follows:

[0103] 1) Set up an RTK base station to receive the drone's location data in real time during the drone's flight and store it in the base station. Then, use a computer to read the data.

[0104] 2) Use Python to write relevant code to train the first BiLSTM network using historical data, and fit a correlation function to describe the nonlinear propagation process.

[0105] 3) The second BiLSTM network is trained using historical estimates, the measured noise covariance calculated by filtering, and the prior error covariance as features. At the same time, a relevant loss function is set to obtain a more reliable actual measurement error covariance to participate in the filtering calculation, thereby achieving the purpose of adaptive adjustment of the filtering process.

[0106] 4) Combine real-time observations to track and predict the location of the UAV.

[0107] The experimental UAV position prediction system consists of three parts: a UAV with an RTK receiver, an RTK base station, and a computer. The algorithm was tested at various altitudes on the UAV and applied to scenarios such as UAV-Unmanned Aerial Vehicle (UAV) cooperative landing, all of which showed good results.

[0108] Example 2: Second aspect, such as Figure 5 As shown, in order to achieve the above objectives, this invention discloses a UAV position prediction system based on Trusted UKF-BiLSTM, comprising:

[0109] Model training module 11 is used to acquire UAV position data, input the UAV position data into a pre-established BiLSTM model for training, and output the UAV position change model.

[0110] The position prediction module 12 is used to acquire the observed value of the UAV's position, input the observed value of the UAV's position into the UAV position change model to predict the UAV's position change, and output the UAV position prediction result.

[0111] In conjunction with the second aspect, in some implementations of the second aspect, the system further includes: the process of acquiring UAV position data by the model training module 11 includes:

[0112] It is equipped with a drone that supports RTK technology and a ground reference station, with the ground reference station serving as a fixed reference point for the RTK system.

[0113] Configure the RTK base station and connect the RTK receiver on the drone to the base station via radio or network to enable communication between the devices.

[0114] During the flight of the drone, position data and other sensor data are stored in the drone's internal memory or transmitted wirelessly to the ground station for storage in real time.

[0115] The model training module 11 inputs the UAV position data into the pre-established BiLSTM model for training, and outputs the calculation formula of the UAV position change model as follows:

[0116] X i,k|k-1 =BiLSTM(X) i,k-1 )

[0117] In the formula, X i,k-1 Let X be the Sigma sampling point at time k-1. i,k|k-1 X after nonlinear transfer i,k-1 ;

[0118] The position prediction module 12 inputs the observed values ​​of the UAV's position into the UAV position change model to predict the UAV's position change:

[0119] Nonlinear transfer:

[0120]

[0121] Depend on and X is obtained by performing Sigma sampling i,k-1 :

[0122]

[0123] The result of the Sigma point set after nonlinear propagation is as follows:

[0124]

[0125] Update the filter gain, estimate, and error covariance:

[0126]

[0127] In the formula, the number of sampling points is 2n+1. Calculate the error covariance for the filter. To calculate the measurement error covariance, ω is the inertia weight. Here, λ represents the process noise covariance assumed by the filter; h(·) is the scaling factor; and K is the measurement function. The cross-covariance of prior error and measurement error;

[0128] The process and formula for solving the confidence level of the location prediction module 12 are as follows:

[0129]

[0130] The credibility level is set to:

[0131]

[0132] Set the loss function to

[0133] Based on the same inventive concept, this invention also provides a computer device, comprising: one or more processors, and a memory for storing one or more computer programs; the programs include program instructions, and the processor executes the program instructions stored in the memory. The processor may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing and control core of the terminal, used to implement one or more instructions, specifically for loading and executing one or more instructions stored in a computer storage medium to implement the above-described method.

[0134] It should be further explained that, based on the same inventive concept, the present invention also provides a computer storage medium storing a computer program, which, when executed by a processor, performs the above-described method. This storage medium can be any combination of one or more computer-readable media. The computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. The computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of computer-readable storage media (a non-exhaustive list) include: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In the present invention, the computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0135] In the description of this specification, references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this disclosure. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0136] The foregoing has shown and described the basic principles, main features, and advantages of this disclosure. Those skilled in the art should understand that this disclosure is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of this disclosure. Various changes and modifications can be made to this disclosure without departing from its spirit and scope, and all such changes and modifications fall within the scope of this disclosure as claimed.

Claims

1. A method for UAV position prediction based on trusted UKF-BiLSTM, characterized in that, The method includes the following steps: Acquire UAV position data, input the UAV position data into a pre-established BiLSTM model for training, and output the UAV position change model. The UAV position data is input into a pre-built BiLSTM model for training, and the calculation formula for the UAV position change model is as follows: In the formula, For the Sigma sampling point at time k-1, After nonlinear transfer ; The system acquires observations of the UAV's position, inputs these observations into the UAV position change model to predict the UAV's position change, and outputs the UAV position prediction result. The process of inputting observed drone positions into a drone position change model to predict drone position changes: Nonlinear transfer: Depend on and Sigma sampling was performed to obtain : The result of the Sigma point set after nonlinear propagation is as follows: Update the filter gain, estimate, and error covariance: In the formula, the number of sampling points is 2n+1. Calculate the error covariance for the filter. To calculate the measurement error covariance, For inertial weights, The process noise covariance is assumed for the filter; It is a scaling factor; For measurement functions; This is the gain matrix; The cross-covariance of prior error and measurement error; The process and formula for determining the credibility are as follows: The credibility level is set to: Set the loss function to .

2. The UAV position prediction method based on Trusted UKF-BiLSTM according to claim 1, characterized in that, The process of collecting the UAV location data includes: Equipped with a drone supporting RTK technology and a ground reference station, the ground reference station serves as a fixed reference point for the RTK system. Configure the RTK base station and connect the RTK receiver on the drone to the base station via radio or network; During drone flight, drone position data and other sensor data are stored in the drone's internal memory or transmitted wirelessly to a ground station for storage in real time.

3. A UAV position prediction system based on Trusted UKF-BiLSTM, characterized in that, include: The model training module is used to acquire UAV position data, input the UAV position data into a pre-built BiLSTM model for training, and output the UAV position change model. The model training module inputs the UAV position data into a pre-built BiLSTM model for training, and outputs the calculation formula for the UAV position change model as follows: In the formula, For the Sigma sampling point at time k-1, After nonlinear transfer ; The position prediction module is used to acquire the observed value of the UAV's position, input the observed value of the UAV's position into the UAV position change model to predict the change of the UAV's position, and output the UAV position prediction result. The position prediction module inputs the observed position values ​​of the UAV into the UAV position change model to predict the UAV position change: Nonlinear transfer: Depend on and Sigma sampling was performed to obtain : The result of the Sigma point set after nonlinear propagation is as follows: Update the filter gain, estimate, and error covariance: In the formula, the number of sampling points is 2n+1. Calculate the error covariance for the filter. To calculate the measurement error covariance, For inertial weights, The process noise covariance is assumed for the filter; It is a scaling factor; For measurement functions; This is the gain matrix; The cross-covariance of prior error and measurement error; The process and formula for solving the confidence level of the location prediction module are as follows: The credibility level is set to: Set the loss function to .

4. The UAV position prediction system based on Trusted UKF-BiLSTM according to claim 3, characterized in that, The process of collecting UAV location data in the model training module includes: Equipped with a drone supporting RTK technology and a ground reference station, the ground reference station serves as a fixed reference point for the RTK system. Configure the RTK base station and connect the RTK receiver on the drone to the base station via radio or network; During drone flight, drone position data and other sensor data are stored in the drone's internal memory or transmitted wirelessly to a ground station for storage in real time.

5. A terminal device, comprising a memory, a processor, and a computer program stored in the memory and capable of running on the processor, characterized in that, The memory stores a computer program that can run on a processor. When the processor loads and executes the computer program, it employs the UAV position prediction method based on Trusted UKF-BiLSTM as described in any one of claims 1 to 2.

6. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is loaded and executed by the processor, it employs the UAV position prediction method based on Trusted UKF-BiLSTM as described in any one of claims 1 to 2.

Images