Unmanned aerial vehicle thunderstorm monitoring device based on electric field and magnetic field detection and path planning method

Abstract

The invention discloses an unmanned aerial vehicle thunderstorm monitoring device based on electric field and magnetic field detection and a path planning method. The device comprises an orthogonal magnetic antenna array, a vertical differential electric field sensor, a signal processing module, a positioning algorithm module, a distance estimation module, an autonomous decision module, a GNSS module and an inertial measurement module. The orthogonal magnetic antenna array is composed of two orthogonally arranged magnetic antennas, forms an included angle of 45 degrees with the flight direction of the unmanned aerial vehicle, and is used for receiving horizontal magnetic field signals of lightning radiation. The vertical differential electric field sensor comprises an upper group of electrode plates and a lower group of electrode plates and is used for measuring an electric field intensity differential value in the vertical direction. A lightning azimuth angle (theta) is calculated through cooperative measurement of a magnetic antenna array and an electric field sensor in combination with a magnetic field intensity ratio, and a quadrant is corrected by using electric field polarity. And constructing a distance estimation model based on the magnetic field amplitude attenuation model (E / B = c) and the waveform rise time correction factor. The autonomous decision-making module generates an abutting path or an obstacle avoidance path according to the lightning orientation and the distance, and drives the flight control system to adjust the course.

Description

Technical Field

[0001] The present invention relates to the field of atmospheric science detection technology, and in particular to a UAV thunderstorm position monitoring device and path planning method based on electric and magnetic field detection, which enables the UAV to locate the thunderstorm position in real time, estimate the distance, and autonomously plan the flight path to collect atmospheric data. Background Art

[0002] Thunderstorms are severe weather phenomena often accompanied by dangerous conditions such as lightning, lightning strikes, short bursts of heavy rainfall, strong winds, and even hail. Thunderstorm monitoring can detect the occurrence and development of thunderstorms in advance, issue timely warnings to the public, and remind people to take protective measures, such as avoiding activities in open areas and staying away from tall trees and metal objects, thereby effectively reducing the threat to human life posed by thunderstorm-related events such as lightning strikes. Thunderstorm monitoring is of great significance to aviation, navigation, railways, and road transportation. Most existing thunderstorm monitoring technologies are based on methods such as atmospheric electric field instrument monitoring, ground-based lightning location, radar monitoring, satellite remote sensing monitoring, and manual observation.

[0003] With the continuous expansion of drone applications, the provision of an airborne carrier for thunderstorm detection has opened up a new avenue for aerial thunderstorm approach monitoring that tracks the movement of thunderstorm clouds. In complex meteorological environments, thunderstorm clouds are constantly moving, making drone approach monitoring difficult to achieve with manual control. In areas such as drone logistics and power line inspection, encountering thunderstorm clouds along planned routes is unavoidable. Lightning activity poses a serious threat to drones, and drone flight safety issues are becoming increasingly prominent. Some existing drones are equipped with simple lightning detectors, but these instruments have limited accuracy, making it difficult to accurately determine the direction and distance of lightning strikes, thus failing to provide effective support for drone path planning decisions. Furthermore, traditional detection methods are significantly affected by environmental noise and lack effective means to suppress electromagnetic interference from the drone itself, resulting in poor signal reliability. Regarding positioning accuracy, traditional methods fail to fully account for the varying propagation characteristics of electromagnetic waves under varying atmospheric conditions, as well as the impact of lightning waveform characteristics on distance estimation. This results in significant errors in positioning results and fails to meet the requirements for refined drone path planning. Furthermore, existing technologies fail to deeply integrate lightning detection with drone flight control, making it difficult to achieve real-time dynamic path planning and autonomous decision-making. Summary of the Invention

[0004] The present invention aims to provide a UAV thunderstorm location monitoring device and path planning method based on electric and magnetic field detection. The monitoring device includes an orthogonal magnetic antenna array, a vertical differential electric field sensor, a signal processing module, a positioning algorithm module, a distance estimation unit, an autonomous decision-making module, a GNSS module, and an inertial measurement module. The orthogonal magnetic antenna array consists of two orthogonal magnetic antennas arranged at a 45° angle to the UAV's flight direction, and is used to receive horizontal magnetic field signals radiated by lightning. The vertical differential electric field sensor comprises upper and lower sets of electrode plates, which measure the vertical electric field intensity differential. The signal processing module performs bandpass filtering, logarithmic amplification, and ratio calculation on the magnetic field signal, outputting the magnetic field intensity ratio of the two orthogonal magnetic field components. It then differentially amplifies the electric field signal, extracts the differential mode signal, and suppresses common-mode interference. The positioning algorithm module calculates the lightning incident azimuth based on the magnetic field intensity ratio and, combined with the electric field polarity, determines whether the lightning is located in front of or behind the UAV's heading. The distance estimation unit calculates the lightning distance based on a magnetic field amplitude attenuation model and waveform polarity correlation. The autonomous decision-making module generates an approach path or obstacle avoidance path based on the lightning direction and distance, and drives the flight control system to adjust the heading.

[0005] The present invention is achieved through the following technical solution: a UAV thunderstorm position monitoring device based on electric field and magnetic field detection includes an orthogonal magnetic antenna array 1, a vertical differential electric field sensor 2, a signal processing module 3, a positioning algorithm module 4, a distance estimation module 5, an autonomous decision-making module 6, a GNSS module 7, and an inertial measurement module 8.

[0006] The orthogonal magnetic antenna array 1 consists of two orthogonally arranged magnetic antennas 1a and 1b, arranged at a 45° angle to the drone's flight direction. It is used to receive horizontal magnetic field signals radiated by lightning. The vertical differential electric field sensor 2 comprises upper and lower sets of electrode plates 2a and 2b, which measure the vertical electric field intensity differential. The signal processing module 3 performs bandpass filtering, logarithmic amplification, and ratio calculation on the magnetic field signal, outputting the ratio of the magnetic field intensities of the two orthogonal magnetic field components (H1 / H2). The electric field signal is differentially amplified, extracting the differential-mode signal between the two electrodes and suppressing common-mode interference to achieve signal enhancement.

[0007] The positioning algorithm module 4 is based on the magnetic field intensity ratio (H1 / H2) and the lightning incident azimuth. i Calculated by the inverse tangent function: , If the lightning is a negative ground lightning (current direction is downward), the electric field signal E Negative polarity is required. i conduct π Angle correction: θ′ =θ+π , combined with the electric field polarity, it is determined whether the lightning is located in front of or behind the drone.

[0008] The distance estimation module 5: based on the magnetic field amplitude attenuation model ( E / B =c) and waveform polarity correlation to calculate lightning distance D : , in, K is the atmospheric propagation correction factor (0.8~1.2), E max is the peak value of the electric field, B peak is the peak value of the magnetic field, and c is the propagation speed of the electromagnetic wave.

[0009] The autonomous decision module 6 issues a proximity warning based on the lightning azimuth (±5°) and distance (±500m), generates a mission path plan, and drives the flight control system to adjust the course.

[0010] The monitoring device realizes the lightning azimuth ( i ) and distance ( D ) is solved by a single station. The magnetic antenna is arranged at a 45° angle to improve the directional resolution, and the electric field polarity is combined to determine the lightning azimuth quadrant. A distance estimation model is constructed based on the magnetic field amplitude attenuation model, and a waveform rise time correction is introduced to improve the close-range detection accuracy, forming a small UAV thunderstorm monitoring device.

[0011] Furthermore, the magnetic field signal is band-pass filtered using a Butterworth band-pass filter (1kHz-1MHz), whose transfer function is: , Where s is a complex frequency variable; oh 0 is the center angular frequency; Q is the quality factor, which measures the filter frequency selectivity, Q The higher the passband, the narrower the transition band, Q The lower the passband, the wider and gentler it is; K is the gain coefficient.

[0012] Furthermore, the magnetic field intensity ratio is calculated to compensate for temperature drift: , in, V H1 、 V H2 The voltage value output by the logarithmic amplifier circuit after the magnetic antennas 1a and 1b induce the magnetic field is proportional to the magnetic field strength. K cal1 ( T ),K cal2 ( T ) is the temperature calibration coefficient, which compensates for the temperature changes of the magnetic antenna and amplifier ( T The gain drift caused by temperature is ensured H 1 / H 2 precision.

[0013] Furthermore, the distance estimation module 5 uses the lightning waveform rise time ( t r <10μs) as the distance correction factor, when t r <2μs triggers a proximity alarm.

[0014] The absolute position information provided by the GNSS module 7 and the relative motion information provided by the inertial measurement module 8 are combined to calculate the heading angle of the UAV through a data fusion algorithm (such as Kalman filtering). β .

[0015] A path planning method for an unmanned aerial vehicle (UAV) thunderstorm monitoring device based on electric and magnetic field detection comprises the following steps: S1. Electromagnetic signal acquisition and preprocessing: The horizontal magnetic field component of lightning is collected in real time using an orthogonal magnetic antenna array (at a 45° angle to the drone's flight direction) H 1. H 2. Synchronously obtain the vertical electric field strength differential value through the vertical differential electric field sensor DE z ; The magnetic field signal is band-pass filtered (1kHz~1MHz) and logarithmically amplified to calculate the magnetic field intensity ratio. R H =H1 / H2; The signal processing module 3 automatically adjusts the filtering parameters according to the characteristics of the environmental noise through the adaptive noise suppression circuit, and uses wavelet transform to eliminate the 20~50kHz frequency band noise generated by the drone motor to improve the signal-to-noise ratio.

[0016] S2. Attitude calculation and position update: GNSS data acquisition acquires information such as the drone's latitude and longitude, altitude, speed, and timestamp. IMU inertial measurement data acquisition acquires information such as the drone's acceleration, angular velocity, and magnetic field strength. GNSS data undergoes preliminary processing, including outlier removal and smoothing filter preprocessing. IMU data also undergoes preliminary processing, including filtering, denoising, and temperature compensation for the accelerometer and gyroscope data. A Kalman filter data fusion algorithm is used to calculate the drone's attitude information (pitch, roll, and yaw) based on the IMU data. Based on the attitude information from the GNSS and IMU data, the drone's position is updated and the positioning result, including latitude and longitude, altitude, speed, and attitude, is output.

[0017] S3. Lightning direction calculation: based on magnetic field intensity ratio R H Calculate lightning incident azimuth i =arctan( R H ), if the lightning is a negative ground lightning, the electric field signal E Negative polarity is required. i conduct π Angle correction, θ′ =θ+π; Combined with the vertical electric field polarity, it is determined that the lightning is located in the direction of the UAV (Δ E z >0) or rear (Δ E z <0); GNSS module 7 provides the drone's geodetic coordinates and heading angle β, with an accuracy of ≤0.1°. The drone's heading angle β is 0° from due north and increases clockwise. The corrected lightning azimuth θ′ is based on the drone as the origin and in the same coordinate system as β, combined with the drone's heading angle β and the corrected lightning azimuth I will , calculate the angle of the lightning relative to the heading c , c =∣ β-θ' ∣,∣ β-θ' ∣≤90∘ and Δ E z >0 → Lightning is directly ahead (high-risk area); β-θ' ∣>90∘ and Δ E z >0 → Lightning is in front of the side (medium-risk zone), Δ E z <0 → Lightning is behind (low risk area).

[0018] S4. Based on the magnetic field amplitude attenuation model E / B = c Construct distance model, combined with waveform rise time tr Correct distance error: , in, K The atmospheric propagation correction coefficient is 0.8~1.2, k Take 0.5 as the empirical coefficient. t r Distance correction is performed when the time is less than 10μs. t r <2 m s triggers a close-range warning ( D' <1km).

[0019] S5. Dynamic flight path planning: The direction of lightning i ±5° and distance D' ±500m is input into the reinforcement learning model to generate an ellipsoidal thunderstorm area envelope centered on the lightning area; Based on the GNSS satellite navigation system, the global path A* algorithm is integrated with the local path requirements, and the dynamic window method (DWA) is used to calculate the next speed v of the UAV. next and three-dimensional position offsets Δx, Δy, Δz, satisfying dynamic constraints (maximum acceleration ≤ 5m / s², pitch angle ≤ 30°); Based on the real-time linkage mechanism of the rise time correction distance and the UAV flight path, D' When the distance is ≤1km, the emergency flight mode is triggered, and an avoidance path is generated based on the C² continuous B-spline curve. The avoidance radius is ≥300m, and the spiral path reaches the center of the thunderstorm.

[0020] S6. Flight control and feedback optimization: The attitude angle of the drone is adjusted by the PID controller ( α , f ) and thrust to execute the planned path; The flight is monitored in real time. If the lightning position is updated or the flight path is blocked, the system returns to step S1 to recalculate and update the reinforcement learning model weights. The autonomous decision module 6 integrates the reinforcement learning algorithm to optimize the path planning based on the historical lightning distribution data.

[0021] Furthermore, in step S5, the reinforcement learning model adopts a deep Q network (DQN), and the input is the modified lightning azimuth I will ,distance D' 、UAV current speed v current and remaining power, the output is the confidence score of the obstacle avoidance path (0.8~1.0 is a safe path); the direction of the major axis of the ellipsoid enveloping the thunderstorm area and the modified lightning azimuth I will Consistent, long axis length L =2D' , minor axis length W =0.5 L .

[0022] Furthermore, in step S6, the PID controller parameters are dynamically adjusted according to the mass m and the moment of inertia I of the UAV to meet the response time ≤ 200ms and the overshoot ≤ 10%.

[0023] The path planning method of the UAV thunderstorm monitoring device based on electric field and magnetic field detection also includes: presetting a lightning activity heat map database, and automatically switching to active detection mode when the UAV enters a historically high lightning frequency area, with a spiral radius R=0.2D ' , a spiral path with a climb rate ≥ 5 m / s approaches the lightning center (spacing ≥ 300 m) to collect atmospheric data.

[0024] The beneficial effects of the present invention are: using an orthogonal magnetic antenna array and a vertical differential electric field sensor for collaborative measurement, with the magnetic antennas arranged at a 45° angle to improve directional resolution, combining the electric field polarity to determine the lightning azimuth quadrant, improving the lightning azimuth positioning accuracy, and providing more accurate directional information for subsequent path planning decisions. The distance estimation unit constructs a distance estimation model based on the magnetic field amplitude attenuation model and introduces a waveform rise time correction factor to effectively improve close-range detection accuracy, enabling the drone to more accurately judge the lightning distance, take obstacle avoidance measures in a timely manner, and enhance flight safety. The signal processing module uses a Butterworth bandpass filter to filter the magnetic field signal and adopts temperature drift compensation technology to ensure the accuracy of the magnetic field intensity ratio calculation and improve the reliability of the system under different ambient temperatures. The autonomous decision-making module combines lightning azimuth and distance information to generate path planning, uses a reinforcement learning model to dynamically plan the path, integrates global path planning with local obstacle avoidance requirements, considers the dynamic constraints of the drone, and generates a safe and feasible path planning to meet the drone's real-time approach or obstacle avoidance requirements. The flight control and feedback optimization mechanism uses a PID controller to precisely adjust the drone's attitude and thrust, execute the planned path, and monitor the flight trajectory in real time. It optimizes the path strategy based on lightning position updates and environmental change feedback, improving drone flight safety in complex meteorological environments. A pre-installed lightning activity heat map database enables the drone to automatically switch to active detection mode when entering areas with historically high lightning frequencies. It then follows a specific spiral path into the center of a thunderstorm to collect atmospheric data. This enriches lightning detection data, provides more detailed information for meteorological research, and expands the drone's application capabilities in meteorology. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 System block diagram of the present invention; Figure 2 This is a bird's-eye view of the drone; Figure 3 This is the side view of the drone; Figure 4 This is a plane diagram of the relationship between the UAV heading and the lightning incident angle; Figure 5 Flowchart of the UAV positioning algorithm; Figure 6 This is a block diagram of the path planning method.

[0026] In the figure: 1-orthogonal magnetic antenna array, 1a-left magnetic antenna, 1b-right magnetic antenna, 2-vertical differential electric field sensor, 2a-upper electrode plate, 2b-lower electrode plate, 3-signal processing module, 4-positioning algorithm module, 5-distance estimation module, 6-autonomous decision module, 7- GNSS module, 8-inertial measurement module. DETAILED DESCRIPTION

[0027] In order for those skilled in the art to better understand the present invention, Figure 1~Figure 6 To further explain the present application, in the description of the present invention, the terms "up", "down", "left", "right", "inside", "outside", etc. indicate directions or positional relationships based on the directions or positional relationships shown in the accompanying drawings, or directions or positional relationships commonly understood by those skilled in the art. They are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the components or parts referred to must have a specific direction, be constructed and operated in a specific direction. The contents mentioned in the implementation manner are not limitations of the present invention.

[0028] The UAV thunderstorm monitoring device based on electric field and magnetic field detection of the present invention comprises an orthogonal magnetic antenna array 1, a vertical differential electric field sensor 2, a signal processing module 3, a positioning algorithm module 4, a distance estimation module 5, an autonomous decision module 6, a GNSS module 7, and an inertial measurement module 8. The system block diagram is shown in FIG. Figure 1 The orthogonal magnetic antenna array 1 consists of two orthogonally arranged magnetic antennas 1a and 1b, which are installed on both sides of the central axis of the drone body and form a 45° angle with the flight direction of the drone. Figure 2 Aerial view of the drone. The drone in the picture is a fixed-wing drone for illustration only. Rotary-wing drones can also be used. The arrow pointing forward from the center of the drone is the heading. The magnetic antennas 1a and 1b are set as shown in the figure. Figure 2 The orthogonal magnetic antenna array 1 is used to receive the horizontal magnetic field signal of lightning radiation. Its operating frequency range covers the typical frequency band of lightning radiation and is connected to the signal processing module 3 via a coaxial cable. The orthogonal magnetic antennas 1a and 1b use nanocrystalline alloy cores with a magnetic permeability of up to 10 4 The high magnetic permeability enables the antenna to more effectively convert the magnetic field energy into electrical signals when receiving weak magnetic field signals, thereby improving the sensitivity of the antenna and helping to detect farther or weaker signal sources.

[0029] The vertical differential electric field sensor 2 includes two sets of upper and lower electrode plates 2a and 2b. The upper electrode plate 2a is installed on the upper part of the drone fuselage, and the lower electrode plate 2b is installed on the lower part of the drone belly. The electrode plates are fixed with insulating brackets. Figure 3 Side view of the drone. Electrode plates 2a and 2b are metal plates with a diameter of 100 to 200 mm to improve electric field sensitivity. Their output signals are transmitted via differential cables to signal processing module 3. Electrode plates 2a and 2b measure the vertical electric field strength difference.

[0030] The signal processing module 3 is installed inside the UAV cabin and includes a bandpass filter circuit, an adaptive noise suppression circuit, a logarithmic amplifier circuit, a differential amplifier circuit, and an analog-to-digital converter (ADC). The signal processing module 3 performs bandpass filtering and logarithmic amplification on the magnetic field signal, calculates the magnetic field intensity ratio, and outputs the magnetic field intensity ratio of the two orthogonal magnetic field components R H =H1 / H2. The magnetic field signal is filtered using a Butterworth bandpass filter (1kHz~1MHz), and its transfer function is: , Where s is a complex frequency variable; oh 0 is the center angular frequency; Q is the quality factor, which measures the filter frequency selectivity, Q The higher the passband, the narrower the transition band, Q The lower the passband, the wider and gentler it is; K is the gain coefficient; Since temperature changes can directly affect the performance of magnetic materials, the output stability of sensors, and the overall reliability of the measurement system through various physical mechanisms, it is necessary to perform temperature drift compensation on the magnetic field intensity ratio calculation. The compensation formula is as follows: , in, V H1 、 V H2 The voltage value output by the logarithmic amplifier circuit after the magnetic antennas 1a and 1b induce the magnetic field is proportional to the magnetic field strength. K cal1 ( T ), K cal2 ( T ) is the temperature calibration coefficient, which compensates for the magnetic antenna and amplifier due to temperature changes ( T The gain drift caused by temperature is ensured H 1 / H 2. Temperature calibration coefficient K cal1 ( T ),K cal2 ( T ), by dynamic temperature linear regression or quadratic curve fitting R − T Relationship, extract slope K cal Temperature drift compensation is not only necessary to correct the inherent characteristics of materials, but also the core of ensuring system-level measurement reliability.

[0031] Signal processing module 3 performs differential amplification on the electric field signal, enhancing the signal by extracting the differential-mode signal between the two electrodes and suppressing common-mode interference. Electric field signal sources (such as electrodes) typically exhibit high output impedance (>1MΩ) and are susceptible to distributed capacitance and electromagnetic interference, necessitating a high-input impedance front end to prevent signal attenuation. Electric field interference in the environment is often in-phase noise, and differential amplification cancels out the common-mode signal to extract the effective differential-mode component. This differential amplification utilizes a dual operational amplifier, whose symmetrical design and high CMRR components achieve "differential amplification and common-mode suppression" while maintaining a balanced balance of bandwidth, noise, and stability.

[0032] The signal processed by the signal processing module 3 is transmitted to the positioning algorithm module 4. The positioning algorithm module 4 is based on an embedded processor (such as the ARM Cortex-A series) and stores and runs the lightning positioning algorithm. The positioning algorithm module 4 calculates the lightning incident azimuth angle based on the magnetic field intensity ratio (H1 / H2) using the inverse tangent function. i calculate: , The azimuth angle θ represents the horizontal direction angle of the lightning relative to the true north. The process of negative charge in the cloud discharging to the ground (the lower part of the cloud is usually negatively charged) causes the current to be downward, resulting in the electric field signal showing a negative polarity (i.e. the instantaneous value of the electric field strength is negative), which is called a negative ground lightning. If the lightning is a negative ground lightning (the current direction is downward), the electric field signal E Negative polarity will cause the actual direction to deviate by 180°. i conduct π Angle correction: Corrected lightning incident azimuth θ′ =θ+π , after correction, the azimuth I will Restore the real lightning position. Combined with the vertical electric field polarity, determine whether the lightning is in front of or behind the drone. At the same time, the GNSS module 7 provides the drone's geodetic coordinates and heading angle. β , accuracy ≤ 0.1°, UAV heading angle βThe latitude and longitude information provided by the GNSS module 7 can be used to determine the absolute position and general direction of movement of the UAV. The inertial measurement module 8 can provide the acceleration and angular velocity information of the UAV to reflect the short-term motion state of the UAV. The absolute position information provided by the GNSS module 7 and the relative motion information provided by the inertial measurement module 8 are combined to calculate the heading angle of the UAV through data fusion algorithms (such as Kalman filtering). β By fusing these two types of information, the UAV’s heading angle can be updated in real time, overcoming the limitations of a single sensor. I will Taking drone as the origin, β Same coordinate system, combined with the heading angle of the drone β and the corrected lightning azimuth I will , calculate the angle of the lightning relative to the heading c , c =∣ β-θ' ∣, and judge the danger level of the lightning area based on this. The relationship between the UAV heading and the lightning incident angle is shown in Figure 4 In the figure, N is north, E is east, and P is the lightning position.

[0033] The distance estimation module 5 is integrated into the embedded processor and is based on the magnetic field amplitude attenuation model ( E / B =c) and waveform polarity correlation to calculate lightning distance D : , in, K is the atmospheric propagation correction factor (0.8~1.2), E max is the peak value of the electric field, B peak is the peak value of the magnetic field, and c is the propagation speed of the electromagnetic wave.

[0034] Rise time of lightning waveform ( t r ) is a key parameter that describes the time required for the electric field or magnetic field intensity to rise from 10% to 90% of the pulse amplitude, reflecting the steepness of the lightning discharge process and the speed of energy release. t r ) has a significant positive correlation with the distance between the observation point and the lightning. This phenomenon is mainly caused by the dispersion effect and dielectric attenuation during the propagation of electromagnetic waves. t r) Correction of short-range calculations is mainly based on the dual requirements of electromagnetic wave propagation characteristics and sensor physical limitations. The electromagnetic waves generated by lightning contain abundant high-frequency components (MHz to GHz level), and high-frequency electromagnetic waves attenuate much more than low-frequency components when propagating in the atmosphere. t r ) directly reflects the intensity of the high-frequency components of the signal, t r The shorter the signal, the more high-frequency components it contains, and the more severe the attenuation is at close range. D <500m), strong electromagnetic fields will cause the magnetic field sensor to saturate, and the magnetic field peak will be compressed, resulting in B peak The distance calculation is invalid. At this time, the rising edge slope still maintains a linear response when it is saturated. The magnetic field sensor can still maintain dt r / dB The linear response of the rise time ( t r ) Correct the calculation of close distances and solve the problem of distance calculation distortion caused by magnetic field saturation at close distances through the coupling model of the high-frequency attenuation characteristics of electromagnetic waves and the physical response of the sensor.

[0035] Distance estimation module 5 is based on the magnetic field amplitude attenuation model E / B =c Construct the distance model using the lightning waveform rise time ( t r <10μs) as a distance correction factor, combined with the waveform rise time t r For distance error correction, the formula is: , in, K The atmospheric propagation correction coefficient is 0.8~1.2, k Take 0.5 as the empirical coefficient. t r Distance correction is performed when the time is less than 10μs. t r <2 m s triggers a close-range warning.

[0036] Autonomous decision-making module 6 integrates a reinforcement learning algorithm to optimize path planning based on historical lightning distribution data, issue close-range warnings based on lightning direction (±5°) and distance (±500m), generate mission path planning, and drive the flight control system to adjust its course.

[0037] The monitoring device realizes the lightning azimuth ( i ) and distance ( D ) single-station solution. The magnetic antenna is arranged at a 45° angle to improve directional resolution, and the electric field polarity is combined to determine the lightning azimuth quadrant. A distance estimation model is constructed based on the magnetic field amplitude attenuation model, and waveform rise time correction is introduced to improve close-range detection accuracy. This constitutes a small drone-based thunderstorm monitoring device.

[0038] The path planning method of the UAV thunderstorm monitoring device based on electric field and magnetic field detection of the present invention comprises the following steps: S1. Electromagnetic signal acquisition and preprocessing: The horizontal magnetic field component of lightning is collected in real time using an orthogonal magnetic antenna array (at a 45° angle to the drone's flight direction) H 1. H 2. Synchronously obtain the vertical electric field strength differential value through the vertical differential electric field sensor DE z ; The magnetic field signal is band-pass filtered (1kHz~1MHz) and logarithmically amplified to calculate the magnetic field intensity ratio. R H =H1 / H2; The signal processing module 3 automatically adjusts the filtering parameters according to the characteristics of the environmental noise through the adaptive noise suppression circuit, and uses wavelet transform to eliminate the 20~50kHz frequency band noise generated by the drone motor to improve the signal-to-noise ratio.

[0039] S2. Attitude calculation and position update: GNSS data collection obtains the UAV's latitude and longitude, altitude, speed, timestamp and other information, and IMU inertial measurement data collection obtains the UAV's acceleration, angular velocity, magnetic field strength and other information. GNSS and IMU data are collected in parallel, and hardware errors are eliminated through preprocessing. GNSS data is preliminarily processed, including removing outliers and smoothing filter data preprocessing; IMU data is preliminarily processed, including filtering and denoising the accelerometer and gyroscope data, and temperature compensation data preprocessing. The Kalman filter data fusion algorithm is used to calculate the UAV's attitude information (pitch angle, roll angle, yaw angle) based on the IMU data; the UAV's position information is updated based on the attitude information of the GNSS data and IMU data. The Kalman filter is used to fuse the dynamic response of the gyroscope and the static stability of the accelerometer to solve the problem of attitude angle drift. When the GNSS signal is temporarily lost or the accuracy is insufficient, the IMU data and the known initial position information are used to perform inertial navigation calculations to infer the current position of the UAV. The inertial navigation calculation formula is: , Among them, v k is the speed update, indicating the current speed v kThe velocity v at the previous moment k−1 Plus the acceleration after attitude change a b ( t ) in the time interval [ t k −1, t k ]; p k is the position update, indicating the current position p k From the position p at the previous moment k−1 Add the velocity v after the posture change b ( t ) in the time interval [ t k−1 , t k ] The integral within. The GNSS position information and the inertial navigation calculation results are integrated to obtain more accurate UAV position information, including latitude and longitude, altitude, speed, attitude and other information. See the UAV positioning flow chart Figure 5 .

[0040] S3. Lightning direction calculation: based on magnetic field intensity ratio R H Calculate lightning incident azimuth i =arctan( R H ), if the lightning is a negative ground lightning, the electric field signal E Negative polarity is required. i conduct π Angle correction, θ′ =θ+π , combined with the vertical electric field polarity, it is determined that the lightning is located in the direction of the UAV (Δ E z >0) or rear (Δ E z <0); GNSS module 7 provides the UAV's geodetic coordinates and heading angle β, with an accuracy of ≤0.1°. The UAV's heading angle β is 0° in the north and increases clockwise. The lightning azimuth θ′ is corrected with the UAV as the origin and in the same coordinate system as β, combined with the UAV's heading angle β and the corrected lightning azimuth I will , calculate the angle of the lightning relative to the heading c , c =∣ β-θ' ∣,∣ β-θ' ∣≤90∘ and Δ E z >0 → Lightning is directly ahead (high-risk area), ∣ β-θ' ∣>90∘ and Δ E z >0 → Lightning is in front of the side (medium-risk zone), ΔE z <0 → Lightning is behind (low risk area).

[0041] S4. Based on the magnetic field amplitude attenuation model E / B = c Construct distance model, combined with waveform rise time t r Correct distance error: , in, K The atmospheric propagation correction coefficient is 0.8~1.2, k Take 0.5 as the empirical coefficient. t r Distance correction is performed when the time is less than 10μs. t r <2 m s triggers a close-range warning ( D' <1km).

[0042] S5. Dynamic flight path planning: The direction of lightning i ±5° and distance D' ±500m input reinforcement learning model, generate ellipsoidal thunderstorm area envelope centered on the lightning area. The reinforcement learning model uses deep Q network (DQN), and the input is the corrected lightning azimuth. I will ,distance D' 、UAV current speed v current and remaining power, the output is the confidence score of the obstacle avoidance path (0.8~1.0 is a safe path); the major axis direction of the ellipsoid enveloping the thunderstorm area and the corrected lightning azimuth I will Consistent, long axis length L =2 D' , minor axis length W =0.5 L ; Based on the GNSS satellite navigation system, the global path A* algorithm is integrated with the local path requirements, and the dynamic window method (DWA) is used to calculate the next speed v of the UAV. next and three-dimensional position offsets Δx, Δy, Δz, satisfying dynamic constraints (maximum acceleration ≤ 5m / s², pitch angle ≤ 30°); Based on the real-time linkage mechanism of the rise time correction distance and the UAV flight path, D' When the distance is ≤1km, the emergency flight mode is triggered, and an avoidance path is generated based on the C² continuous B-spline curve. The avoidance radius is ≥300m, and the spiral path reaches the center of the thunderstorm.

[0043] S6. Flight control and feedback optimization: The attitude angle of the drone is adjusted by the PID controller ( α , f ) and thrust, executing the planned path. PID controller parameters are dynamically adjusted based on the drone's mass m and moment of inertia I, ensuring a response time of ≤200ms and an overshoot of ≤10%. The drone monitors its flight path in real time during flight. If the lightning position is updated or the flight path is obstructed, it returns to step S1 for recalculation and updates the reinforcement learning model weights. Autonomous decision-making module 6 integrates a reinforcement learning algorithm to optimize path planning based on historical lightning distribution data.

[0044] The path planning method of the UAV thunderstorm monitoring device based on electric field and magnetic field detection also includes: presetting a lightning activity heat map database, and automatically switching to active detection mode when the UAV enters a historically high lightning frequency area, with a spiral radius R=0.2D ' , a spiral path with a climb rate ≥ 5 m / s approaches the lightning center (spacing ≥ 300 m) to collect atmospheric data.

[0045] Example 1: Lightning incident azimuth i and the heading angle of the drone β Estimate. The voltage value output by the logarithmic amplifier circuit after the magnetic antenna 1a and 1b induces the magnetic field is V H1 =2.35V, V H2 =1.88V, temperature T=30℃; calibration coefficient K cal1 ( T )=1.02-0.003( T- 25), K cal2 ( T )=0.98+0.002( T- 25), (T in °C), temperature drift compensation: , , , substitute the parameters, ; Magnetic field ratio R H = H 1 / H 2=1.269→ i =arctan(1.269)≈51.83°, Electric field difference Δ E z =-1.8 kV / m (Negative polarity) → Negative ground-to-ground flash correction i ′=51.83°+180°=231.83°, GNSS heading angle β =120.5° (true north reference, clockwise), Relative heading angle c =∣120.5°−231.83°∣=111.33°>90°, and Δ E z <0 → Lightning is behind, low risk area.

[0046] Example 2: Distance estimation. Electric field peak value E max =42 V / m , peak magnetic field B peak =0.085 μT , atmospheric correction coefficient K = 1.05 (value when humidity is 70%), rise time t r =1.2 μs , calculate the distance: , , the actual lightning has approached the danger zone; Rise time t r =1.2 μs <10 μs → Trigger close range correction: , , after correction, it matches the real distance; Rise time t r =1.2 μs <2 μs →Trigger emergency flight mode.

[0047] The present invention is only an embodiment of a drone as an airborne carrier carrying an electric field electromagnetic detection device and autonomously planning a flight path according to mission requirements. In actual applications, it can carry atmospheric detection equipment to detect thunderstorm clouds as needed, and transmit the data back to the ground center by wireless means. In application, please note that D Electric field difference value Δ when <300m E z Nonlinearity leads to failure of azimuth angle solution. Therefore, when compiling the flight path planning algorithm, the aircraft should fly around the thunderstorm center at a distance of ≥300m to collect atmospheric data.

[0048] The present invention can also be used in fields such as drone logistics and power line inspection. By detecting the location of thunderstorms, the flight path can be autonomously planned to avoid thunderstorm areas.

[0049] The description and drawings of this application are only a specific implementation method and are not restrictive. Under the guidance of this application, those skilled in the art can make many forms without departing from the purpose of this application and the scope of protection of the claims, all of which are within the scope of protection of this application.

Claims

1. A UAV thunderstorm monitoring device based on electric and magnetic field detection, characterized by: It includes an orthogonal magnetic antenna array (1), a vertical differential electric field sensor (2), a signal processing module (3), a positioning algorithm module (4), a distance estimation module (5), an autonomous decision module (6), a GNSS module (7), and an IMU inertial measurement module (8); The orthogonal magnetic antenna array (1) is composed of two orthogonally arranged magnetic antennas (1a, 1b), which are at an angle of 45° to the flight direction of the UAV and are used to receive horizontal magnetic field signals radiated by lightning; The vertical differential electric field sensor (2) comprises two sets of upper and lower electrode plates (2a, 2b) for measuring the vertical electric field intensity differential value; The signal processing module (3) performs bandpass filtering, logarithmic amplification, and ratio calculation on the magnetic field signal, outputting the ratio of the magnetic field strength of two orthogonal magnetic field components (H1 / H2); differentially amplifies the electric field signal, extracts the differential mode signal between the two electrodes, and suppresses common mode interference to achieve signal enhancement; The positioning algorithm module (4) is based on the magnetic field intensity ratio (H1 / H2), the lightning incident azimuth θ Calculated by the inverse tangent function: ， If the lightning is a negative ground lightning (current direction is downward), the electric field signal E Negative polarity is required. θ conduct π Angle correction: θ′ =θ+π ,Combined with the electric field polarity, it is determined whether the lightning is located in front of or behind the UAV; The distance estimation module (5): based on the magnetic field amplitude attenuation model ( E / B =c) and waveform polarity correlation to calculate lightning distance D : ， in, K is the atmospheric propagation correction factor (0.8~1.2), E max is the peak value of the electric field, B peak is the peak value of the magnetic field, c is the propagation speed of the electromagnetic wave; The autonomous decision-making module (6) issues a close-range warning based on the lightning azimuth (±5°) and distance (±500m), generates a mission path plan, and drives the flight control system to adjust the course; The monitoring device realizes the lightning azimuth ( θ ) and distance ( D ) is solved by a single station. The magnetic antenna is arranged at a 45° angle to improve the directional resolution, and the electric field polarity is combined to determine the lightning azimuth quadrant. A distance estimation model is constructed based on the magnetic field amplitude attenuation model, and a waveform rise time correction is introduced to improve the close-range detection accuracy, forming a small UAV thunderstorm monitoring device.

2. The UAV thunderstorm monitoring device based on electric and magnetic field detection according to claim 1 is characterized in that: The magnetic field signal is band-pass filtered using a Butterworth band-pass filter (1kHz-1MHz), whose transfer function is: ， Where s is a complex frequency variable; ω 0 is the center angular frequency; Q is the quality factor, which measures the filter frequency selectivity, Q The higher the passband, the narrower the transition band, Q The lower the passband, the wider it is. K is the gain coefficient.

3. The UAV thunderstorm monitoring device based on electric and magnetic field detection according to claim 1 is characterized in that: The magnetic field intensity ratio calculation is used to compensate for temperature drift: ， in, V H1 、 V H2 is the voltage value output by the logarithmic amplifier circuit after the magnetic antenna (1a, 1b) induces the magnetic field, which is proportional to the magnetic field strength. K cal1 ( T ), K cal2 ( T ) is the temperature calibration coefficient, which compensates for the magnetic antenna and amplifier due to temperature changes ( T The gain drift caused by temperature is ensured H 1 / H 2 precision.

4. The UAV thunderstorm monitoring device based on electric and magnetic field detection according to claim 1 is characterized in that: The distance estimation module (5) uses the lightning waveform rise time ( t r <10μs) as the distance correction factor, when t r <2μs triggers a proximity alarm.

5. The UAV thunderstorm monitoring device based on electric and magnetic field detection according to claim 1 is characterized in that: The absolute position information provided by the GNSS module (7) and the relative motion information provided by the IMU inertial measurement module (8) are combined to calculate the heading angle of the UAV through a data fusion algorithm (such as Kalman filtering). β .

6. A path planning method for an unmanned aerial vehicle thunderstorm monitoring device based on electric and magnetic field detection according to claim 1, characterized in that: The following steps are involved: S1. Electromagnetic signal acquisition and preprocessing: Real-time acquisition of the horizontal magnetic field component of lightning using an orthogonal magnetic antenna array H 1. H 2. Synchronously obtain the vertical electric field strength differential value through the vertical differential electric field sensor ΔE z ; The magnetic field signal is band-pass filtered (1kHz~1MHz) and logarithmically amplified to calculate the magnetic field intensity ratio. R H =H1 / H2; The signal processing module (3) automatically adjusts the filter parameters according to the characteristics of the environmental noise through the adaptive noise suppression circuit, and uses wavelet transform to eliminate the 20~50kHz frequency band noise generated by the drone motor to improve the signal-to-noise ratio; S2. Attitude calculation and position update: GNSS data acquisition obtains the drone's latitude and longitude, altitude, speed, timestamp and other information, while IMU data acquisition obtains the drone's acceleration, angular velocity, magnetic field strength and other information; Perform preliminary processing on GNSS data, remove outliers, and perform smoothing and filtering data preprocessing; Perform preliminary processing on IMU data, filter and denoise the accelerometer and gyroscope data, and perform temperature compensation data preprocessing; The Kalman filter data fusion algorithm is used to calculate the attitude information (pitch, roll, and yaw) of the drone based on the IMU data. The position information of the drone is updated based on the attitude information of the GNSS data and the IMU data. S3. Lightning direction calculation: According to the magnetic field strength ratio R H Calculate lightning incident azimuth θ =arctan( R H ), if the lightning is a negative ground lightning, the electric field signal E Negative polarity is required. θ conduct π Angle correction, θ'=θ+π ; Combined with the vertical electric field polarity, it is determined that the lightning is located in the direction of the UAV (Δ E z >0) or rear (Δ E z <0); GNSS module (7) provides the UAV’s geodetic coordinates and heading angle β, with an accuracy of ≤0.1°. The UAV heading angle β is 0° in the north and increases clockwise to correct the lightning azimuth. θ′ Taking the drone as the origin, in the same coordinate system as β, combined with the heading angle of the drone β and the corrected lightning azimuth θ' , calculate the angle of the lightning relative to the heading γ , γ =∣ β-θ' ∣,∣ β-θ' ∣≤90∘ and Δ E z >0 → Lightning is directly ahead (high-risk area), ∣ β-θ' ∣>90∘ and Δ E z >0 → Lightning is in front of the side (medium-risk zone), Δ E z <0 → Lightning is behind (low risk area); S4. Based on the magnetic field amplitude attenuation model E / B = c Construct distance model, combined with waveform rise time t r Correct distance error: ， in, K The atmospheric propagation correction coefficient is 0.8~1.2, k Take 0.5 as the empirical coefficient. t r Distance correction is performed when the time is less than 10μs. t r <2 μ s triggers a close-range warning ( D' <1km); S5. Dynamic flight path planning: The direction of lightning θ ±5° and distance D' ±500m is input into the reinforcement learning model to generate an ellipsoidal thunderstorm area envelope centered on the lightning area; Based on the GNSS satellite navigation system, the global path A* algorithm is integrated with the local path requirements, and the dynamic window method (DWA) is used to calculate the next speed v of the UAV. next and three-dimensional position offsets Δx, Δy, Δz, satisfying dynamic constraints (maximum acceleration ≤ 5m / s², pitch angle ≤ 30°); Based on the real-time linkage mechanism of the rise time correction distance and the UAV flight path, D' When the distance is less than 1km, the emergency flight mode is triggered, and an avoidance path is generated based on the C² continuous B-spline curve. The avoidance radius is ≥300m, and the spiral path reaches the center of the thunderstorm. S6. Flight control and feedback optimization: The attitude angle of the drone is adjusted by the PID controller ( α , φ ) and thrust to execute the planned path; The flight is monitored in real time. If the lightning position is updated or the flight path is blocked, the algorithm returns to step S1 to recalculate and update the weights of the reinforcement learning model. The autonomous decision module (6) integrates the reinforcement learning algorithm to optimize the path planning based on the historical lightning distribution data.

7. A path planning method for an unmanned aerial vehicle thunderstorm monitoring device based on electric and magnetic field detection according to claim 6, characterized in that: In step S5, the reinforcement learning model uses a deep Q network (DQN), and the input is the modified lightning azimuth θ' ,distance D' 、UAV current speed v current and remaining battery power, and the output is the confidence score of the flight path (0.8 to 1.0 is a safe path); The major axis direction of the thunderstorm area envelope ellipsoid and the modified lightning azimuth θ' Consistent, long axis length L =2 D' , minor axis length W =0.5 L .

8. A path planning method for an unmanned aerial vehicle thunderstorm monitoring device based on electric and magnetic field detection according to claim 6, characterized in that: In step S6, the PID controller parameters are dynamically adjusted according to the mass m and the moment of inertia I of the UAV to meet the response time ≤ 200ms and the overshoot ≤ 10%.

9. A path planning method for an unmanned aerial vehicle thunderstorm monitoring device based on electric and magnetic field detection according to any one of claims 6 to 8, characterized in that: Also includes: The lightning activity heat map database is preset. When the drone enters the area with historically high lightning frequency, it automatically switches to active detection mode with a spiral radius of R=0.2D. ' , a spiral path with a climb rate of ≥5m / s reaches the center of the thunderstorm (spacing ≥300m) to collect atmospheric data.

Images