Fusion algorithm-based unmanned aerial vehicle focus drop point prediction and adjustment method and system

Abstract

The invention discloses an unmanned aerial vehicle focus drop point prediction and adjustment method and system based on a fusion algorithm, and belongs to the field of seismic exploration, and the method comprises the following steps: S1, collecting the environment data of a flight region, and the position, acceleration and attitude data of a focus, and forming multi-source data; s2, determining initial state information of the seismic source, and transmitting the multi-source data and the determined initial state information to a remote server; s3, the remote server constructs a seismic source motion model based on an adaptive Runge-Kutta algorithm in combination with deep learning, simulates and predicts the flight trajectory of the seismic source based on the constructed seismic source motion model, and generates flight trajectory and drop point prediction information; and S4, generating a control instruction based on the flight path and the drop point prediction information, and optimizing state parameters. According to the fusion algorithm-based unmanned aerial vehicle focus drop point prediction and adjustment method and system, based on multi-source data, the fusion algorithm is used to calculate and optimize the flight path, and accurate throwing of the unmanned aerial vehicle focus is realized.

Description

Technical Field

[0001] The present invention relates to the field of seismic exploration technology, and in particular to a method and system for predicting and adjusting the earthquake source drop point of an unmanned aerial vehicle based on a fusion algorithm. Background Art

[0002] With the vigorous development of science and technology, geophysical exploration technology is also continuously innovating. Among them, drone seismic source technology, as an emerging force in the field of geological exploration, has shown great application potential, thus bringing new ideas and methods to geological exploration work, and is expected to break through the limitations of traditional exploration methods and achieve more efficient and accurate exploration operations.

[0003] However, in actual application scenarios, especially in complex and changing environments, the UAV source throwing process is significantly affected by a variety of environmental factors, such as wind speed, wind direction, throwing angle, initial velocity, etc. These factors are intertwined and become the key bottleneck restricting the accuracy of UAV source technology, resulting in many challenges for the existing UAV source technology in its wide application.

[0004] At the same time, although the UAV seismic source system has the ability to perform basic geophysical exploration tasks, it has obvious deficiencies in responding to real-time changing environmental factors. Since it cannot fully consider the dynamic changes of the environment, the stability of the seismic source flight trajectory is difficult to guarantee, and the problem of throwing deviation occurs from time to time, which seriously affects the accuracy of the exploration results.

[0005] Moreover, when obtaining high-altitude environmental conditions, existing technologies mostly rely on indirect inferences based on the measurement data of ground wind speed and wind direction. However, the actual situation is that the wind speed on the ground and at high altitudes is quite different, and this indirect inference method cannot accurately reflect the actual flight environment of the earthquake source. When the high-altitude environment changes, the traditional UAV earthquake source system lacks the corresponding adjustment and compensation mechanism, making it difficult for the earthquake source to land accurately, which in turn causes deviations in the exploration results.

[0006] In addition, the reasonable selection of the throwing angle and initial velocity plays a vital role in the source trajectory. Different exploration tasks require flexible adjustment of the throwing angle and initial velocity due to the different geological characteristics and mission requirements of the target area. However, the traditional UAV source system has limited ability to dynamically adjust these parameters, and it is difficult to effectively optimize the flight trajectory in complex environments, which greatly restricts the effectiveness of the exploration mission. Summary of the invention

[0007] The purpose of the present invention is to provide a method and system for predicting and adjusting the earthquake source landing point of an unmanned aerial vehicle based on a fusion algorithm to solve the above-mentioned technical problems.

[0008] To achieve the above object, the present invention provides a method for predicting and adjusting the earthquake source drop point of a UAV based on a fusion algorithm, comprising the following steps: S1, collect environmental data of the flight area, as well as the location, acceleration and attitude data of the earthquake source to form multi-source data; S2, after preprocessing the multi-source data, determining the initial state information of the earthquake source, and transmitting the multi-source data and the determined initial state information to a remote server; S3. The remote server constructs a source motion model based on the adaptive Runge-Kutta algorithm combined with deep learning, and simulates and predicts the flight trajectory of the source based on the constructed source motion model to generate flight trajectory and landing point prediction information; S4. Generate control instructions based on the flight trajectory and landing point prediction information to optimize state parameters.

[0009] Preferably, the environmental data described in step S1 includes the wind speed in the flight area. ,wind direction ,temperature ,humidity and air pressure Data, acceleration includes acceleration data of the source on the X, Y, and Z axes , and , the attitude data includes the pitch angle of the source , Roll Angle and yaw angle ; Wind speed Including the wind speed in the flight area on the X, Y, and Z axes , and .

[0010] Preferably, the initial state information in step S2 includes the initial velocity, initial position, pitch angle, acceleration, flight angle and direction of the source, wherein the flight angle is the pitch angle collected. , direction is the collected yaw angle ; It specifically includes the following steps: S21, preprocessing multi-source data using Kalman filter algorithm; S22. Integrate the preprocessed acceleration data to obtain the initial velocity of the earthquake source: (1); In the formula, , , Respectively The initial velocity of the earthquake source on the X, Y, and Z axes at the moment; Indicates the initial time; The coordinates of the UAV when it drops the earthquake source are used as the initial position of the earthquake source ; At the same time, the posture data is used to determine the throwing angle : (2); In the formula, represents the initial velocity of the earthquake source, and ; represents the target direction, and , Indicates the target location coordinates; represents the modulus of the initial velocity, and ; represents the magnitude of the target direction, and ; S23, the preprocessed multi-source data and the information on the initial state of the determined earthquake source are transmitted to the communication chip through the UART interface or the SPI interface, and the communication chip transmits the data to the remote server using the NB-IoT protocol.

[0011] Preferably, step S3 specifically includes the following steps: S31. Set the initial step size of the Runge-Kutta algorithm based on the initial state information and environmental data of the earthquake source and preset tolerance error ,in, , Indicates the estimated flight time of the earthquake source. Indicates the expected number of calculation steps; S32. Construct a source motion model by considering air resistance, gravity and thrust factors: (3); (4); (5); In the formula, Respectively represent the position coordinates of the earthquake source in space along the X-axis, Y-axis, and Z-axis directions; , , They represent the components of the instantaneous velocity of the earthquake source in the X-axis, Y-axis, and Z-axis directions respectively; , and Respectively represent the thrust components in the X-axis, Y-axis, and Z-axis directions; Indicates the mass of the earthquake source; represents the air resistance coefficient; represents the air density based on the velocity, and , represents the dry air partial pressure, represents the molar mass of dry air, represents the saturated water vapor pressure, represents the molar mass of water vapor, represents the universal gas constant; It represents the projection area of ​​the earthquake source in the direction perpendicular to the instantaneous velocity; represents the acceleration due to gravity; S33. Use the fourth-order Runge-Kutta algorithm to numerically solve the earthquake source motion model and predict the state of the earthquake source: (6); in, (7); (8); (9); (10); In the formula, and Respectively represent the source status information at the next moment and the current moment; , , , All represent slope estimates; represents the step length; It means building a source motion model; S34. Collect historical data and train a deep learning model based on the historical data. Combine the trained deep learning model with the Runge-Kutta algorithm. Use the deep learning model to capture the nonlinear relationship between environmental data and earthquake source trajectory, and optimize flight trajectory prediction.

[0012] Preferably, in step S33, an adaptive strategy is used to dynamically adjust the step size; It specifically includes the following steps: S331, Calculate the error between the fourth-order Runge-Kutta algorithm and the second-order Runge-Kutta algorithm : (11); In the formula, and They represent the calculation results of the fourth-order Runge-Kutta algorithm and the second-order Runge-Kutta algorithm respectively; S332, based on error estimation Adjust step size: (12); In the formula, Represents the adjusted step size.

[0013] Preferably, step S34 specifically includes the following steps: S341. In the process of calculating the earthquake source motion trajectory by the Runge-Kutta algorithm, the environmental data and earthquake source state information at the current moment are input into the trained deep learning model to obtain a predicted correction value of the earthquake source state at the next moment; S342, the prediction correction value of the deep learning model is integrated with the calculation result of the Runge-Kutta algorithm to generate the final flight trajectory and landing point prediction information: (13); In the formula, It indicates the source state generated after fusion; Represents the source state correction value predicted by the deep learning model; Represents the weight of the correction value predicted by the deep learning model.

[0014] Preferably, step S4 specifically includes the following steps: S41, the remote server transmits the prediction result of the earthquake source flight trajectory, the landing point prediction information and the target information to the ground station, and the ground station optimizes the earthquake source state parameters through the optimization algorithm and generates a control instruction; S42, the ground station transmits the control command back to the remote server through an encrypted communication protocol, and the remote server then transmits the control command to the microcontroller of the drone through a cellular network; S43, after the microcontroller decrypts the control instruction, it adjusts the source state according to the control instruction.

[0015] Preferably, in step S41, a model predictive control algorithm is used to calculate the flight adjustment parameters of the UAV, and the flight adjustment parameters include a casting angle adjustment amount and an initial speed adjustment amount; It specifically includes the following steps: S411. Consider a discrete-time linear time-varying system to describe the motion of the earthquake source and the UAV. Its state space equation is expressed as: (14); In the formula, , , denote the system matrix, input matrix and noise vector respectively; represents the control input vector, and Includes throw angle adjustment and initial speed adjustment ; S412. Define a quadratic objective function To measure the deviation between the predicted state and the target state, as well as the change in the control input: (15); In the formula, Indicates that based on The information prediction at the moment The source state at the moment; Indicates The target state at the moment; represents the state weight matrix; Indicates that based on The information at the moment determines the Control input at all times; Indicates that based on The information prediction at the moment The source state at the moment; Indicates The target state at the moment; Also add the following constraints: (16); (17); In the formula, and Respectively represent the minimum and maximum values ​​of the throwing angle; and Respectively represent the minimum and maximum values ​​of the initial velocity; S413. Solving the quadratic objective function Determine the control input vector .

[0016] A UAV seismic source drop point prediction and adjustment system based on a fusion algorithm is used to execute a UAV seismic source drop point prediction and adjustment method based on a fusion algorithm, comprising: Data acquisition module, used to collect environmental data of the flight area, as well as acceleration and attitude data of the source; A microcontroller, used to determine the initial state information of the source after preprocessing the multi-source data, and adjust the flight parameters according to the control instructions fed back by the remote server; A communication chip, used to upload the initial status information to a remote server; A cellular module, used to configure network connection parameters and establish a cellular network connection between the communication chip and the remote server; A remote server, used to generate flight trajectory and landing point prediction information based on initial state information; The ground station is used to generate control instructions based on the flight trajectory and landing point prediction information, optimize the flight parameters of the drone and feed them back to the microcontroller; The data acquisition module is connected to the microcontroller, the microcontroller communicates with the communication chip via the UART interface or the SPI interface, the communication chip communicates bidirectionally with the remote server via the cellular module, and the remote server communicates bidirectionally with the ground station.

[0017] Preferably, the data acquisition module includes a wind speed sensor, a wind direction sensor, a temperature and humidity sensor, and a barometer installed on the drone, and a three-axis acceleration sensor, an attitude sensor, and a GPS integrated inside the source; The wind speed sensor and the wind direction sensor are connected to the microcontroller via a digital communication interface; the temperature and humidity sensor and the barometer are connected to the microcontroller via a digital interface.

[0018] Therefore, the present invention adopts the above-mentioned method and system for predicting and adjusting the earthquake source drop point of a UAV based on a fusion algorithm, which has the following beneficial effects: 1. Accuracy: Integrating real-time environmental data, flight data, status data and historical data can effectively reduce the prediction error caused by factors such as wind speed, wind direction, and throwing angle, significantly improve the accuracy of source delivery, and provide accurate landing point prediction for drone sources; 2. Real-time: Based on the constantly changing environmental data during flight, the prediction results can be updated in real time. The drone can be precisely controlled accordingly to ensure that the earthquake source falls in line with the target requirements, thus improving the flexibility and response speed of the system. 3. Adaptability: The system fully considers the complex changes of different environmental conditions such as wind speed, wind direction, temperature, humidity, etc., and can optimize the flight trajectory in different environments such as mountainous areas and coastal areas to ensure accurate placement of the earthquake source; 4. Multi-model fusion: Combining physical models, numerical integration methods and machine learning technology, the physical model provides a theoretical basis, and the machine learning algorithm handles complex nonlinear relationships, improving the overall performance of prediction and enabling the system to have high-precision theoretical analysis capabilities; 5. Real-time monitoring and feedback: Real-time monitoring of the UAV's flight trajectory, environmental changes, etc., and providing adjustment information for real-time adjustment of flight missions through feedback mechanisms to improve control accuracy and flight safety.

[0019] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 The present invention provides a flow chart of a method for predicting and adjusting the earthquake source landing point of an unmanned aerial vehicle based on a fusion algorithm. DETAILED DESCRIPTION

[0021] In order to make the purpose, technical scheme and advantages disclosed in the embodiments of the present invention clearer, the embodiments of the present invention are further described in detail in combination with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the embodiments of the present invention and are not used to limit the embodiments of the present invention. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application. Examples of the embodiments are shown in the accompanying drawings, where the same or similar reference numerals throughout represent the same or similar elements or elements with the same or similar functions.

[0022] It should be noted that the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units that are not explicitly listed or inherent to these processes, methods, products or devices.

[0023] The embodiments of the present invention are described in detail below in conjunction with the accompanying drawings.

[0024] like Figure 1 As shown, a method for predicting and adjusting the earthquake source drop point of a UAV based on a fusion algorithm includes the following steps: S1, collect environmental data of the flight area, as well as the location, acceleration and attitude data of the earthquake source to form multi-source data; The environmental data described in step S1 includes the wind speed in the flight area ,wind direction ,temperature ,humidity and air pressure Data, acceleration includes acceleration data of the source on the X, Y, and Z axes , and , the attitude data includes the elevation angle of the source , Roll Angle and yaw angle ; Wind speed Including the wind speed in the flight area on the X, Y, and Z axes , and .

[0025] S2, after preprocessing the multi-source data, determining the initial state information of the earthquake source, and transmitting the multi-source data and the determined initial state information to a remote server; The initial state information described in step S2 includes the initial velocity, initial position, pitch angle, acceleration, flight angle and direction of the source, where the flight angle is the pitch angle collected. , direction is the collected yaw angle ; It specifically includes the following steps: S21, preprocessing multi-source data using Kalman filter algorithm; S22. Integrate the preprocessed acceleration data to obtain the initial velocity of the earthquake source: (1); In the formula, , , Respectively The initial velocity of the earthquake source on the X, Y, and Z axes at the moment; Indicates the initial time; The coordinates of the UAV when it drops the earthquake source are used as the initial position of the earthquake source ; At the same time, the posture data is used to determine the throwing angle : (2); In the formula, represents the initial velocity of the earthquake source, and ; represents the target direction, and , Indicates the target location coordinates; represents the modulus of the initial velocity, and ; represents the magnitude of the target direction, and ; S23, the preprocessed multi-source data and the information on the initial state of the determined earthquake source are transmitted to the communication chip through the UART interface or the SPI interface, and the communication chip transmits the data to the remote server using the NB-IoT protocol.

[0026] S3. The remote server constructs a source motion model based on the adaptive Runge-Kutta algorithm combined with deep learning, and simulates and predicts the flight trajectory of the source based on the constructed source motion model to generate flight trajectory and landing point prediction information; Step S3 specifically includes the following steps: S31. Set the initial step size of the Runge-Kutta algorithm based on the initial state information and environmental data of the earthquake source and preset tolerance error ,in, , Indicates the estimated flight time of the earthquake source. Indicates the expected number of calculation steps; S32. Construct a source motion model by considering air resistance, gravity and thrust factors: (3); (4); (5); In the formula, Respectively represent the position coordinates of the earthquake source in space along the X-axis, Y-axis, and Z-axis directions; , , They represent the components of the instantaneous velocity of the earthquake source in the X-axis, Y-axis, and Z-axis directions respectively; , and Respectively represent the thrust components in the X-axis, Y-axis, and Z-axis directions; Indicates the mass of the earthquake source; represents the air resistance coefficient; represents the air density based on the velocity, and , represents the dry air partial pressure, represents the molar mass of dry air, represents the saturated water vapor pressure, represents the molar mass of water vapor, represents the universal gas constant; It represents the projection area of ​​the earthquake source in the direction perpendicular to the instantaneous velocity; represents the acceleration due to gravity; S33. Use the fourth-order Runge-Kutta algorithm to numerically solve the earthquake source motion model and predict the state of the earthquake source: (6); in, (7); (8); (9); (10); In the formula, and Respectively represent the source status information at the next moment and the current moment; , , , All represent slope estimates; represents the step length; Indicates the construction of earthquake source motion model; In step S33, the step size is dynamically adjusted using an adaptive strategy; It specifically includes the following steps: S331, Calculate the error between the fourth-order Runge-Kutta algorithm and the second-order Runge-Kutta algorithm : (11); In the formula, and They represent the calculation results of the fourth-order Runge-Kutta algorithm and the second-order Runge-Kutta algorithm respectively; S332, based on error estimation Adjust step size: (12); In the formula, Represents the adjusted step size.

[0027] S34. Collect historical data and train a deep learning model based on the historical data. Combine the trained deep learning model with the Runge-Kutta algorithm. Use the deep learning model to capture the nonlinear relationship between environmental data and earthquake source trajectory, and optimize flight trajectory prediction.

[0028] Step S34 specifically includes the following steps: S341. In the process of calculating the earthquake source motion trajectory by the Runge-Kutta algorithm, the environmental data and earthquake source state information at the current moment are input into the trained deep learning model to obtain a predicted correction value of the earthquake source state at the next moment; S342, the prediction correction value of the deep learning model is integrated with the calculation result of the Runge-Kutta algorithm to generate the final flight trajectory and landing point prediction information: (13); In the formula, It indicates the source state generated after fusion; Represents the source state correction value predicted by the deep learning model; Represents the weight of the correction value predicted by the deep learning model.

[0029] S4. Generate control instructions based on the flight trajectory and landing point prediction information to optimize state parameters.

[0030] Step S4 specifically includes the following steps: S41, the remote server transmits the prediction result of the earthquake source flight trajectory, the landing point prediction information and the target information to the ground station, and the ground station optimizes the earthquake source state parameters through the optimization algorithm and generates a control instruction; In step S41, a model predictive control algorithm is used to calculate the flight adjustment parameters of the UAV, and the flight adjustment parameters include a throwing angle adjustment amount and an initial speed adjustment amount; It specifically includes the following steps: S411. Consider a discrete-time linear time-varying system to describe the motion of the earthquake source and the UAV. Its state space equation is expressed as: (14); In the formula, , , denote the system matrix, input matrix and noise vector respectively; represents the control input vector, and Includes throw angle adjustment and initial speed adjustment ; S412. Define a quadratic objective function To measure the deviation between the predicted state and the target state, as well as the change in the control input: (15); In the formula, Indicates that based on The information prediction at the moment The source state at the moment; Indicates The target state at the moment; represents the state weight matrix; Indicates that based on The information at the moment determines the Control input at all times; Indicates that based on The information prediction at the moment The source state at the moment; Indicates The target state at the moment; Also add the following constraints: (16); (17); In the formula, and Respectively represent the minimum and maximum values ​​of the throwing angle; and Respectively represent the minimum and maximum values ​​of the initial velocity; S413. Solving the quadratic objective function Determine the control input vector .

[0031] S42, the ground station transmits the control command back to the remote server through an encrypted communication protocol, and the remote server transmits the control command to the microcontroller of the drone through a cellular network, wherein the remote server transmits the real-time calculated flight trajectory, landing point offset, control command and other information to the ground station through protocols such as HTTP, MQTT or WebSocket; S43, after the microcontroller decrypts the control instruction, it adjusts the source state according to the control instruction.

[0032] A UAV seismic source drop point prediction and adjustment system based on a fusion algorithm is used to execute a UAV seismic source drop point prediction and adjustment method based on a fusion algorithm, comprising: Data acquisition module, used to collect environmental data of the flight area, as well as acceleration and attitude data of the source; A microcontroller, used to determine the initial state information of the source after preprocessing the multi-source data, and adjust the flight parameters according to the control instructions fed back by the remote server; A communication chip, used to upload the initial status information to a remote server; Cellular module, used to configure network connection parameters (such as APN, authentication key, etc.) and establish a cellular network (such as 4G / 5G) connection between the communication chip and the remote server; A remote server, used to generate flight trajectory and landing point prediction information based on initial state information; The ground station is used to generate control instructions based on the flight trajectory and landing point prediction information, optimize the flight parameters of the drone and feed them back to the microcontroller; The data acquisition module is connected to the microcontroller, the microcontroller communicates with the communication chip via the UART interface or the SPI interface, the communication chip communicates bidirectionally with the remote server via the cellular module, and the remote server communicates bidirectionally with the ground station.

[0033] The data acquisition module includes a wind speed sensor, a wind direction sensor, a temperature and humidity sensor, and a barometer installed on the drone, as well as a three-axis acceleration sensor, an attitude sensor, and a GPS integrated inside the seismic source; the wind speed sensor and the wind direction sensor are connected to the microcontroller through a digital communication interface; the temperature and humidity sensor and the barometer are connected to the microcontroller via a digital interface.

[0034] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that they can still modify or replace the technical solution of the present invention with equivalents, and these modifications or equivalent replacements cannot cause the modified technical solution to deviate from the spirit and scope of the technical solution of the present invention.

Claims

1. A method for predicting and adjusting the earthquake source point of an unmanned aerial vehicle based on a fusion algorithm, characterized in that: The following steps are involved: S1, collect environmental data of the flight area, as well as the location, acceleration and attitude data of the earthquake source to form multi-source data; S2, after preprocessing the multi-source data, determining the initial state information of the earthquake source, and transmitting the multi-source data and the determined initial state information to a remote server; S3. The remote server constructs a source motion model based on the adaptive Runge-Kutta algorithm combined with deep learning, and simulates and predicts the flight trajectory of the source based on the constructed source motion model to generate flight trajectory and landing point prediction information; S4. Generate control instructions based on the flight trajectory and landing point prediction information to optimize state parameters.

2. The method for predicting and adjusting the earthquake source point of an unmanned aerial vehicle based on a fusion algorithm according to claim 1, characterized in that: The environmental data described in step S1 includes the wind speed in the flight area ,wind direction ,temperature ,humidity and air pressure Data, acceleration includes acceleration data of the source on the X, Y, and Z axes , and , the attitude data includes the elevation angle of the source , Roll Angle and yaw angle ; Wind speed Including the wind speed in the flight area on the X, Y, and Z axes , and .

3. The method for predicting and adjusting the earthquake source falling point of an unmanned aerial vehicle based on a fusion algorithm according to claim 2 is characterized in that: The initial state information described in step S2 includes the initial velocity, initial position, pitch angle, acceleration, flight angle and direction of the source, where the flight angle is the pitch angle collected. , direction is the collected yaw angle ; It specifically includes the following steps: S21, preprocessing multi-source data using Kalman filter algorithm; S22. Integrate the preprocessed acceleration data to obtain the initial velocity of the earthquake source: （1）； In the formula, , , Respectively The initial velocity of the earthquake source on the X, Y, and Z axes at the moment; Indicates the initial time; The coordinates of the UAV when it drops the earthquake source are used as the initial position of the earthquake source ; At the same time, the posture data is used to determine the throwing angle : （2）； In the formula, represents the initial velocity of the earthquake source, and ; represents the target direction, and , Indicates the target location coordinates; represents the modulus of the initial velocity, and ; represents the magnitude of the target direction, and ; S23, the preprocessed multi-source data and the information on the initial state of the determined earthquake source are transmitted to the communication chip through the UART interface or the SPI interface, and the communication chip transmits the data to the remote server using the NB-IoT protocol.

4. The method for predicting and adjusting the earthquake source falling point of an unmanned aerial vehicle based on a fusion algorithm according to claim 3 is characterized in that: Step S3 specifically includes the following steps: S31. Set the initial step size of the Runge-Kutta algorithm based on the initial state information and environmental data of the earthquake source and preset tolerance error ,in, , Indicates the estimated flight time of the earthquake source. Indicates the expected number of calculation steps; S32. Construct a source motion model by considering air resistance, gravity and thrust factors: （3）； （4）； （5）； In the formula, Respectively represent the position coordinates of the earthquake source in space along the X-axis, Y-axis, and Z-axis directions; , , They represent the components of the instantaneous velocity of the earthquake source in the X-axis, Y-axis, and Z-axis directions respectively; , and Respectively represent the thrust components in the X-axis, Y-axis, and Z-axis directions; Indicates the mass of the earthquake source; represents the air resistance coefficient; represents the air density based on the velocity, and , represents the dry air partial pressure, represents the molar mass of dry air, represents the saturated water vapor pressure, represents the molar mass of water vapor, represents the universal gas constant; It represents the projection area of ​​the earthquake source in the direction perpendicular to the instantaneous velocity; represents the acceleration due to gravity; S33. Use the fourth-order Runge-Kutta algorithm to numerically solve the earthquake source motion model and predict the state of the earthquake source: （6）； in, （7）； （8）； （9）； （10）； In the formula, and Respectively represent the source status information at the next moment and the current moment; , , , All represent slope estimates; represents the step length; Indicates the construction of earthquake source motion model; S34. Collect historical data and train a deep learning model based on the historical data. Combine the trained deep learning model with the Runge-Kutta algorithm. Use the deep learning model to capture the nonlinear relationship between environmental data and earthquake source trajectory, and optimize flight trajectory prediction.

5. The method for predicting and adjusting the earthquake source falling point of an unmanned aerial vehicle based on a fusion algorithm according to claim 4 is characterized in that: In step S33, the step size is dynamically adjusted using an adaptive strategy; It specifically includes the following steps: S331, Calculate the error between the fourth-order Runge-Kutta algorithm and the second-order Runge-Kutta algorithm : （11）； In the formula, and They represent the calculation results of the fourth-order Runge-Kutta algorithm and the second-order Runge-Kutta algorithm respectively; S332, based on error estimation Adjust step size: （12）； In the formula, Represents the adjusted step size.

6. The method for predicting and adjusting the earthquake source point of an unmanned aerial vehicle based on a fusion algorithm according to claim 5 is characterized in that: Step S34 specifically includes the following steps: S341. In the process of calculating the earthquake source motion trajectory by the Runge-Kutta algorithm, the environmental data and earthquake source state information at the current moment are input into the trained deep learning model to obtain a predicted correction value of the earthquake source state at the next moment; S342, the prediction correction value of the deep learning model is integrated with the calculation result of the Runge-Kutta algorithm to generate the final flight trajectory and landing point prediction information: （13）； In the formula, It indicates the source state generated after fusion; Represents the source state correction value predicted by the deep learning model; Represents the weight of the correction value predicted by the deep learning model.

7. The method for predicting and adjusting the earthquake source drop point of an unmanned aerial vehicle based on a fusion algorithm according to claim 6, characterized in that: Step S4 specifically includes the following steps: S41, the remote server transmits the prediction result of the earthquake source flight trajectory, the landing point prediction information and the target information to the ground station, and the ground station optimizes the earthquake source state parameters through the optimization algorithm and generates a control instruction; S42, the ground station transmits the control command back to the remote server through an encrypted communication protocol, and the remote server then transmits the control command to the microcontroller of the drone through a cellular network; S43, after the microcontroller decrypts the control instruction, it adjusts the source state according to the control instruction.

8. The method for predicting and adjusting the earthquake source drop point of an unmanned aerial vehicle based on a fusion algorithm according to claim 7, characterized in that: In step S41, a model predictive control algorithm is used to calculate the flight adjustment parameters of the UAV, and the flight adjustment parameters include a throwing angle adjustment amount and an initial speed adjustment amount; It specifically includes the following steps: S411. Consider a discrete-time linear time-varying system to describe the motion of the earthquake source and the UAV. Its state space equation is expressed as: （14）； In the formula, , , denote the system matrix, input matrix and noise vector respectively; represents the control input vector, and Includes throw angle adjustment and initial speed adjustment ; S412. Define a quadratic objective function To measure the deviation between the predicted state and the target state, as well as the change in the control input: （15）； In the formula, Indicates that based on The information prediction at the moment The source state at the moment; Indicates The target state at the moment; represents the state weight matrix; Indicates that based on The information at the moment determines the Control input at all times; Indicates that based on The information prediction at the moment The source state at the moment; Indicates The target state at the moment; Also add the following constraints: （16）； （17）； In the formula, and Respectively represent the minimum and maximum values ​​of the throwing angle; and Respectively represent the minimum and maximum values ​​of the initial velocity; S413. Solving the quadratic objective function Determine the control input vector .

9. A UAV seismic source fall point prediction and adjustment system based on a fusion algorithm, used to execute a UAV seismic source fall point prediction and adjustment method based on a fusion algorithm as claimed in claim 8, characterized in that: include: Data acquisition module, used to collect environmental data of the flight area, as well as acceleration and attitude data of the source; A microcontroller, used to determine the initial state information of the source after preprocessing the multi-source data, and adjust the flight parameters according to the control instructions fed back by the remote server; A communication chip, used to upload initial status information to a remote server; A cellular module, used to configure network connection parameters and establish a cellular network connection between the communication chip and the remote server; A remote server, used to generate flight trajectory and landing point prediction information based on initial state information; The ground station is used to generate control instructions based on the flight trajectory and landing point prediction information, optimize the flight parameters of the drone and feed them back to the microcontroller; The data acquisition module is connected to the microcontroller, the microcontroller communicates with the communication chip via the UART interface or the SPI interface, the communication chip communicates bidirectionally with the remote server via the cellular module, and the remote server communicates bidirectionally with the ground station.

10. The UAV earthquake source fall point prediction and adjustment system based on fusion algorithm according to claim 9, characterized in that: The data acquisition module includes a wind speed sensor, wind direction sensor, temperature and humidity sensor, and barometer installed on the drone, as well as a three-axis acceleration sensor, attitude sensor, and GPS integrated inside the seismic source; The wind speed sensor and the wind direction sensor are connected to the microcontroller via a digital communication interface; the temperature and humidity sensor and the barometer are connected to the microcontroller via a digital interface.

Images