Optimal opportunity adaptive decision method for lightning triggering based on unmanned aerial vehicle platform

Abstract

The application discloses a thunder and lightning trigger optimal opportunity adaptive decision method based on a UAV platform, and comprises the following steps: multi-source electric field data real-time acquisition and preprocessing; multi-dimensional feature extraction; multi-level fusion criterion calculation; adaptive weight updating and threshold adjustment; trigger decision output and safety interlocking. The application realizes deep fusion of multi-source electric field information and quantitative adaptive determination of a trigger opportunity.

Description

Technical Field

[0001] This invention belongs to the field of atmospheric electricity and lightning protection technology, and more specifically, relates to an adaptive decision-making method for the optimal timing of lightning triggering based on an unmanned aerial vehicle (UAV) platform. Background Technology

[0002] Artificial lightning induction technology is a core experimental method for studying the physical mechanisms of natural lightning, testing the performance of lightning protection devices, and obtaining electromagnetic parameters of lightning. Traditional rocket-guided lightning induction is currently the most mature method of artificially triggering lightning internationally. Its basic principle is to launch a small rocket with a thin metal wire attached to it towards a charged cloud during thunderstorm activity. When the electric field strength at the tip of the wire exceeds the air breakdown threshold, it induces an upward pilot to develop into the cloud, eventually forming a complete artificial lightning discharge channel.

[0003] However, traditional rocket-based lightning induction methods have the following prominent problems:

[0004] First, the timing of the triggering process heavily relies on human experience and judgment. Operators judge the timing of the launch based on limited information such as ground electric field readings, visual cloud conditions, and the direction of thunder. This is highly subjective and inconsistent, with different operators making significantly different judgments under the same conditions. The success rate of triggering is usually only 40% to 60%.

[0005] Second, the available information dimensions are limited. Traditional methods mainly rely on single-point ground electric field measurements, which cannot obtain information on the vertical distribution of the electric field in the cloud-ground space or the fine time-varying characteristics of the electric field, resulting in extremely limited understanding of the charge structure within the cloud.

[0006] Third, there is a lack of a systematic quantitative decision-making framework. Triggering criteria are mostly based on empirical thresholds, failing to fully consider multi-dimensional factors such as electric field change trends, synergistic effects of meteorological environment, and feedback from historical triggering data.

[0007] Fourth, key issues such as eliminating rotor electromagnetic interference, accurately measuring high-altitude electric fields, and coupling platform safety constraints with triggering decisions when using unmanned aerial vehicle (UAV) platforms for mine attraction have not yet been systematically resolved.

[0008] Therefore, there is an urgent need for an intelligent decision-making method that integrates multi-source electric field data, has adaptive learning capabilities, and can output quantified trigger confidence scores to overcome the above shortcomings. Summary of the Invention

[0009] To address the aforementioned issues, this invention proposes an adaptive decision-making method for optimal lightning triggering timing based on an unmanned aerial vehicle (UAV) platform. This method overcomes the shortcomings of traditional manual lightning triggering, such as reliance on experience for triggering timing, limited information dimensions, and a lack of a quantitative decision-making framework. It achieves deep fusion of multi-source electric field information and intelligent adaptive determination of triggering timing.

[0010] To address at least one of the aforementioned technical problems, according to one aspect of the present invention, an adaptive decision-making method for optimal lightning triggering timing based on an unmanned aerial vehicle (UAV) platform is provided, comprising the following steps: real-time acquisition and preprocessing of multi-source electric field data: synchronously acquiring the high-altitude electric field E using an airborne electric field sensor mounted on the UAV and a ground-based electric field mill. a (t) and near-ground electric field Eg(t), simultaneously collect conductor induced voltage Vw(t), meteorological parameters and regional lightning location data; the collected signals are sequentially subjected to notch filtering based on rotor speed adaptive tracking to eliminate rotor interference, wavelet decomposition and soft threshold denoising to suppress random noise, and multi-channel timestamp alignment according to GPS timing reference.

[0011] Multidimensional feature extraction: Within a sliding time window of width Tw, extract the mean ground electric field Ēg and the mean airborne electric field Ē. a Ground electric field change rate (kg), airborne electric field standard deviation (σ) a Nine characteristic quantities are included: vertical electric field gradient Gᵥ, peak value of conductor induced voltage Vmax, lightning activity density ρL, nearest lightning distance Dmin, and meteorological synergy index Mc.

[0012] Multi-level fusion criterion calculation: combining nine feature quantities with real-time wind speed W s The platform state parameters, namely the battery remaining capacity (Brem), are divided into three layers: the electric field core layer, the environmental coordination layer, and the safety constraint layer. Within each layer, the features are normalized and then weighted to obtain the layer scores SC, SE, and SS. Finally, the inter-layer weights are used to weight and fuse the results to obtain the comprehensive trigger confidence score.

[0013] ;

[0014] C trig To synthesize the trigger confidence level, it is compared with the dynamic trigger threshold θ to determine whether to trigger a discharge; β C β E β S These are the interlayer weights for the electric field core layer, the environmental coordination layer, and the safety constraint layer, respectively, satisfying β. C +β E +β S = 1; S C S E S S This is a weighted score for each feature within the three levels;

[0015] Adaptive weight update and threshold adjustment: After each trigger attempt, the feature weights within each layer and the weights between layers are updated according to the gradient direction rule based on the actual trigger result y.

[0016] w j(m) w represents the current weight of the j-th feature in the m-th iteration. j (m+1) The updated weights; η is the learning rate, η > 0, controlling the step size of each weight adjustment; y (m) For the m-th trigger, the binary actual result is 1 = success, 0 = failure; C trig (m) The overall trigger confidence level at the m-th trigger; (y (m) - C trig (m) That is, prediction error; The dimensionless value of the j-th feature after direction-consistent normalization is given when the m-th trigger occurs. After the weight is updated, non-negative truncation (set to ε when wj is lower than ε, and ε > 0 is the lower limit protection value of the weight to prevent any feature weight from being completely pushed to zero) and same-layer normalization (the sum of the weights in the same layer is 1) are also required to ensure that the physical meaning of the weight is reasonable.

[0017] It performs non-negative truncation and normalization, and adaptively adjusts the dynamic trigger threshold θ according to the asymmetric rule of lowering it on success and raising it on failure; when the system fails to meet the trigger conditions for multiple consecutive windows, the threshold is automatically lowered through the idle period threshold decay mechanism to prevent cold start lock-up.

[0018] Trigger decision output and safety interlock: Compare Ctrig with the threshold θ, and output a trigger command under the premise that the five safety interlock conditions of wind speed, battery, wire voltage, lightning safety distance and flight control status are met at the same time.

[0019] Furthermore, in the real-time acquisition and preprocessing of multi-source electric field data, the specific method for eliminating rotor interference is as follows: the fundamental frequency of interference fᵣ = nᵣ×Nb is calculated based on the rotor speed nᵣ and the number of blades Nb reported by the flight control system in real time. Notch filters are constructed for fᵣ, 2fᵣ and 3fᵣ respectively. The three notch filters are cascaded to form an adaptive filter group, and its center frequency is adjusted in real time with the change of rotational speed.

[0020] Furthermore, the formula for calculating the vertical electric field gradient is:

[0021] Where h is the flight altitude of the UAV, and Gᵥ reflects the steepness of the cloud-to-ground spatial potential descent.

[0022] Furthermore, the meteorological coordination index:

[0023] , where X̃ kThe normalized values ​​for each meteorological parameter are defined. The normalization direction for each parameter is determined according to the principle of mapping to larger values ​​that are more conducive to lightning triggering. The normalization results are truncated and restricted to the interval [0,1]. α k The coefficient of coordination and Σα k = 1.

[0024] Furthermore, the direction-consistent normalization process is as follows: For positive vectors, the following is applied: For the inverse vector, the following is applied:

[0025] F j F represents the original measured value of the j-th feature (with physical dimensions, such as kV / m, ℃, m / s, etc.); j ref The reference range value (F) for the j-th feature j ref > 0), which means the upper limit of the range or the reference value of this feature under typical working conditions; This represents the normalized dimensionless relative value, ranging from [0, 1]. The purpose of this is to eliminate differences in dimensions and numerical ranges between different features, ensuring a fair weighted summation of all features on the same scale.

[0026] Furthermore, after the weights are updated, non-negative truncation and normalization are performed:

[0027]

[0028] Where ε>0 is the lower limit protection value for weights, which prevents any feature weights from being completely pushed to zero.

[0029] Furthermore, the formula for adaptive adjustment of the trigger threshold is: Define the base value of the adjustment amount. Where γ > 0 is the bias adjustment term; when the trigger is successful, θ←θ−δ s ·Δ, when triggering fails, θ←θ +δf·Δ, where δf > δ s > 0 to reflect the principle of safety first, the threshold is limited to the range of [θmin, θmax].

[0030] The five safety interlock conditions are: wind speed below the wind speed limit, battery capacity above the minimum charge level, wire voltage below the voltage limit, the nearest lightning distance greater than the safe distance, and the flight control system in normal condition. These five conditions are applied to the trigger determination in a logical AND manner. If any condition is not met, the trigger output is automatically suppressed.

[0031] Maintain an idle window counter Nidle, which increments only when the overall trigger confidence is below a threshold. When the number of consecutively untriggered sliding windows exceeds the preset threshold Nidle,max, the trigger threshold is lowered by a fixed step size δidle, but not lower than the lower limit θmin. The counter is reset to zero and starts counting again after each decay.

[0032] According to another aspect of the present invention, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps in the adaptive decision-making method for optimal lightning triggering timing based on an unmanned aerial vehicle platform according to the present invention.

[0033] According to another aspect of the present invention, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of the adaptive decision-making method for optimal lightning triggering timing based on an unmanned aerial vehicle platform according to the present invention.

[0034] Compared with existing technologies, the beneficial effects of the above-described method of the present invention are as follows:

[0035] This invention elevates triggering decisions from empirical judgment based on a single threshold to quantitative decisions based on multi-dimensional feature fusion, significantly improving the triggering success rate. It introduces high-altitude electric field measurement from UAVs to acquire vertical electric field distribution information in the cloud-to-ground space, compensating for the information blind spots of traditional single-point ground measurements. An adaptive method for eliminating rotor electromagnetic interference is designed, ensuring reliable accuracy of the electric field data collected by the UAV platform. An adaptive weight update mechanism is constructed, allowing the system to continuously learn and optimize based on historical triggering data without requiring repeated manual parameter adjustments. A multi-dimensional safety interlock mechanism is implemented to automatically suppress triggering under extreme weather conditions or platform anomalies, ensuring operational safety. An idle-period threshold decay strategy is designed to solve the problem of excessively high thresholds preventing triggering during the system's cold start phase, ensuring the system can autonomously enter normal operating conditions under any initial conditions. Attached Figure Description

[0036] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings of the embodiments will be briefly described below. Obviously, the drawings described below only relate to some embodiments of the present invention and are not intended to limit the present invention.

[0037] Figure 1 This is a flowchart of a preferred embodiment of the present invention;

[0038] Figure 2 This is a schematic diagram of a multi-level fusion criterion and adaptive weight update mechanism according to a preferred embodiment of the present invention; Detailed Implementation

[0039] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention.

[0040] Unless otherwise defined, the technical or scientific terms used herein shall have the ordinary meaning as understood by one of ordinary skill in the art to which this invention pertains.

[0041] Example 1:

[0042] like Figure 1-2 As shown, this invention provides an adaptive decision-making method for optimal lightning triggering timing based on an unmanned aerial vehicle (UAV) platform, comprising the following steps:

[0043] Step 1: Real-time acquisition and preprocessing of multi-source electric field data;

[0044] 1.1 Data Acquisition:

[0045] After the drone hovers at a predetermined altitude h (in meters), the system simultaneously collects the following five types of sensor signals:

[0046] (a) Near-ground atmospheric electric field Eg(t) output by the ground electric field mill, in kV / m;

[0047] (b) The high-altitude electric field E output by the airborne electric field sensor a (t), unit kV / m;

[0048] (c) The potential difference between the conductor and ground output by the conductor induced voltage sensor, Vw(t), in kV;

[0049] (d) Temperature T(t), relative humidity RH(t), air pressure P(t), and wind speed W output by the meteorological sensor group s (t);

[0050] (e) The spatiotemporal information of lightning events provided by the regional lightning location network, including the time of lightning occurrence, latitude and longitude, and peak current, is received in real time via a data link.

[0051] Among them, electric field signals (Eg, E) a The sampling frequency of Vw should be no less than 100 Hz, and the sampling frequency of meteorological parameters should be no less than 1 Hz.

[0052] 1.2 Rotor interference elimination:

[0053] The rotation of the drone's rotor generates periodic electromagnetic interference, the fundamental frequency of which is fᵣ:

[0054]

[0055] Where nᵣ is the real-time rotor speed in r / s, provided in real time by the flight control system; Nb is the number of rotor blades in a single group.

[0056] The system constructs an adaptive notch filter bank for the fundamental frequency fᵣ and its second harmonic 2fᵣ and third harmonic 3fᵣ. The transfer function of the k-th order notch filter is:

[0057]

[0058] Where f s denoted by , where is the sampling frequency; r is the pole radius, 0.95 ≤ r ≤ 0.99. The closer r is to 1, the narrower the notch bandwidth and the higher the frequency selectivity. The center frequency of the notch filter is adjusted in real time according to the rotational speed nᵣ reported by the flight control, ensuring accurate suppression of rotor interference under variable rotational speed conditions.

[0059] 1.3 Wavelet Denoising

[0060] The signal after rotor interference cancellation is decomposed into three levels using the db4 wavelet basis. Let the high-frequency detail coefficients of the j-th level be dⱼ(k), and apply soft thresholding to them:

[0061]

[0062] The threshold λⱼ is estimated using a general thresholding method: the median absolute deviation of the j-th layer detail coefficients divided by 0.6745 is used as a robust estimate of the noise standard deviation σ̂ⱼ, and then λⱼ = σ̂ⱼ √(2 ln Nⱼ), where Nⱼ is the number of coefficients in the j-th layer. After processing, wavelet reconstruction is performed using the corrected detail coefficients and the low-frequency approximation coefficients of the third layer to obtain the denoised signal.

[0063] 1.4 Timestamp alignment and outlier handling;

[0064] All sensor channels are timestamped using a unified GPS timing reference to eliminate micro-delays between channels. Outlier sampling points outside the physically reasonable range are replaced with linear interpolation of the nearest valid sampling point.

[0065] Step 2: Multidimensional feature extraction;

[0066] Let the width of the sliding time window be Tw, and the window contain N sampling points, with sampling times of t1, t2, ..., tN. Extract the following nine features within each window:

[0067] Feature 1 and 2: Mean ground electric field Ēg and mean airborne electric field Ē a

[0068] Both take the arithmetic mean within the window, and the formulas are identical:

[0069]

[0070] Where E(tᵢ) is substituted into Eg(tᵢ) or E a (tᵢ), unit kV / m. Ēg reflects the average intensity of the ground electric field. a It contains direct information about the charge distribution near the cloud base.

[0071] Feature 3: Rate of change of ground electric field (kg)

[0072] The time series of the ground electric field within the window is fitted with a least-squares linear model, and its slope is taken:

[0073]

[0074] Unit: kV / (m·s). kg > 0 indicates that the electric field is continuously increasing, and the larger the kg value, the more mature the triggering conditions are.

[0075] Feature 4: Standard deviation σ of airborne electric field a

[0076]

[0077] The unit is kV / m. The population standard deviation formula is used here, and it has no substantial difference from the sample standard deviation when the number of sampling points N in the window is sufficiently large. A larger value indicates more severe fluctuations in the upper-level electric field, which is usually an indirect indicator of enhanced discharge leader activity or rapid charge reorganization within clouds.

[0078] Feature 5: Vertical electric field gradient Gᵥ

[0079]

[0080] Where h is the flight altitude of the UAV, in meters. Gᵥ is in kV / m², reflecting the average rate of change of electric field strength with altitude between the cloud base and the ground. The larger Gᵥ is, the stronger the non-uniformity of the electric field in the vertical direction, which usually means that there is significant charge accumulation in the cloud-ground space, which is more conducive to leader initiation.

[0081] Feature 6: Peak value of induced voltage in conductor Vmax

[0082]

[0083] The unit kV directly represents the level of electrical stress that the lightning conductor experiences in the current electric field environment.

[0084] Feature 7: Lightning activity density ρL

[0085] Within a circular region centered at the launch site and with radius R, the total number of lightning events nL within the statistical window [t1, tN] is calculated as follows:

[0086]

[0087] The unit is times / (s·km²), and the larger the value of ρL, the higher the maturity of the thunderstorm.

[0088] Feature 8: Nearest lightning distance Dmin

[0089]

[0090] Where d l denoted as , representing the horizontal distance from the l-th lightning event to the launch site, in km. If there are no lightning events within the window (nL = 0), then Dmin is taken as the search radius R.

[0091] Feature Nine: Meteorological Coordination Index Mc

[0092]

[0093] Where K is the number of meteorological parameters involved in the calculation; α k Let Σα be the coordination coefficient of each meteorological parameter. k =1;X̃ k Let be the dimensionless value of the normalized mean of the k-th meteorological parameter within the window. Normalization uses a linear mapping to the interval [0, 1].

[0094]

[0095] Where X k min and X k `max` represents the reasonable range boundary for the k-th meteorological parameter. The normalization direction of each parameter is determined according to the principle of "mapping the more favorable triggering conditions to larger values," for example, relative humidity is mapped in the positive direction, and wind speed is mapped in the negative direction. All normalization results are truncated and limited to the interval [0, 1], and `Mc` takes the value range [0, 1], with the value closer to 1 indicating that the meteorological conditions are more favorable for triggering.

[0096] Step 3: Calculation of multi-level fusion criteria;

[0097] 3.1 Hierarchical Division

[0098] The nine features extracted in step two, along with the two real-time platform state parameters, are divided into three levels:

[0099] Electric field core layer (Layer C): includes Ēg, Ēa kg, σ a Both Gᵥ and Gᵥ are positive vectors. This layer represents the electric field intensity and its changing trend in the cloud-ground space, and is the core information source for the triggering criterion.

[0100] The Environmental Coordination Layer (Layer E) includes ρL (forward), Dmin (reverse), and Mc (forward). This layer assesses the degree of coordination between thunderstorm activity and meteorological conditions from a macro-environmental perspective.

[0101] Security constraint layer (Layer S): includes Vmax (reverse) and W s (Reverse), Brem (Forward). Vmax, being directly related to the conductor insulation safety margin, is included in this layer. W s Brem is provided directly by the flight control system and weather sensors.

[0102] 3.2 Directional Consistency Normalization

[0103] All 11 quantities are subjected to directional consistency normalization processing, that is, all features are uniformly processed into the direction of "the larger the value, the more conducive to triggering", and mapped to the interval [0, 1].

[0104] For positive vectors (such as Ēg, Ē) a kg, σ a , Gᵥ, ρL, Mc, Brem):

[0105]

[0106] For inverse vectors (such as Dmin, Vmax, W) s ):

[0107]

[0108] Where Fⱼ,ref is the reference range value of the j-th quantity (Fⱼ,ref > 0), which is determined by statistical analysis of historical experimental data or physical experience, and is used to map the characteristics of different physical dimensions to a unified dimensionless interval.

[0109] After the above processing, all F̂ⱼ ∈ [0, 1], and the closer F̂ⱼ is to 1, the more favorable or safer the conditions are in that dimension. It should be noted that kg can take negative values. After substituting into the positive formula, it is truncated to 0 by max(·, 0), indicating that the dimension is completely unfavorable for triggering under the weakening trend of the electric field, which is consistent with the physical meaning.

[0110] 3.3 Weighted Summation within a Layer

[0111] Electric core layer score SC:

[0112]

[0113] F̂1 to F̂5 correspond to Ēg, Ē, and Ē respectively. a kg, σ a The normalized values ​​of Gᵥ; w1 to w5 are the feature weights of each element in the electric field core layer, satisfying wⱼ > 0 and Σwⱼ = 1.

[0114] The environmental coordination layer score (SE) and the safety constraint layer score (SS) use the same formula:

[0115]

[0116] Where uⱼ and vⱼ are the feature weights of the corresponding layers, and each of them satisfies that the sum of the weights is 1 and both are greater than 0. Since F̂ⱼ ∈ [0, 1] and the sum of the weights within the layer is 1, the score of each layer belongs to [0, 1].

[0117] 3.4 Inter-layer weighted fusion

[0118] The scores from the three levels are combined into a comprehensive trigger confidence score, Ctrig.

[0119]

[0120] Where βC, βE, and βS are the inter-layer weights of the three levels, satisfying βC + βE + βS = 1 and each weight being greater than 0. Ctrig ∈ [0, 1], the closer to 1, the more suitable the comprehensive conditions are for triggering. The initial weight setting principle is that the electric field core layer is the largest, followed by the environmental coordination layer, and the safety constraint layer is the bottom line guarantee. The specific initial values ​​are determined based on historical experimental statistics.

[0121] Step 4: Adaptive weight update and threshold adjustment;

[0122] 4.1 Triggering result feedback:

[0123] After each attempt is triggered, the system records the binary result variable y:

[0124]

[0125] Simultaneously, record all normalized feature vectors F̂⁽ᵐ⁾, three-level scores, and comprehensive trigger confidence Ctrig⁽ᵐ⁾ at the trigger moment, where the superscript (m) indicates the m-th trigger attempt.

[0126] 4.2 Intra-layer weight update rules:

[0127] Taking the electric field core layer as an example, after the m-th trigger, the feature weights are updated according to the following gradient direction rules:

[0128]

[0129] Where η > 0 is the learning rate. This rule increases the weight of high-value features upon successful triggering and decreases the weight of high-value features upon failed triggering. After updating, non-negative truncation and normalization are performed to ensure that the sum of the weights is 1 and that each weight is not lower than the lower limit protection value ε (ε > 0).

[0130]

[0131] The environmental coordination layer weight uⱼ and the security constraint layer weight vⱼ use the exact same update rule.

[0132] 4.3 Inter-layer weight update rules:

[0133] Inter-layer weights are updated using the scores of each layer as "feature values" in a structured manner, taking βC as an example:

[0134]

[0135] Similarly, for βE and βS, simply replace SC with SE and SS respectively. After the update, perform non-negative truncation and normalization to ensure that βC + βE + βS = 1 and that each weight is not lower than ε.

[0136] 4.4 Adaptive adjustment of trigger threshold:

[0137] The system maintains a dynamic trigger threshold θ (initial value denoted as θ⁽). 0 The threshold is adjusted only when the system actually performs a trigger attempt (i.e., Ctrig⁽ᵐ⁾≥ θ⁽ᵐ⁾ and passes the safety interlock check).

[0138] Define the adjustment base as:

[0139]

[0140] Where γ > 0 is the bias adjustment term, ensuring a non-zero adjustment even when Ctrig is exactly equal to θ. The threshold is lowered upon successful triggering and raised upon failed triggering.

[0141]

[0142] Where δ s δf > 0 represents the successful contraction step size factor, and δf > 0 represents the failed expansion step size factor. We take δf > δf. sTo reflect the principle of prioritizing safety, the threshold is limited to the range [θmin, θmax].

[0143]

[0144] 4.5 Idle period threshold decay mechanism:

[0145] To address the issue of prolonged trigger failures due to cold starts or excessively high thresholds, the system maintains an idle window counter, Nidle, to record the number of consecutively untriggered sliding windows. Nidle increments only when Ctrig < θ (i.e., confidence level is not met); if Ctrig ≥ θ but triggering is not executed due to failure to meet safety interlock conditions, Nidle does not increment to avoid unnecessary threshold decay during periods of unmet safety constraints. When Nidle exceeds a preset threshold, Nidle,max, the trigger threshold is automatically lowered.

[0146]

[0147] Where δidle > 0 is the idle decay step size, which is less than δ s To prevent the threshold from decreasing too quickly, the counter is reset to zero and starts counting again after each decay, ensuring that the system can autonomously enter the normal trigger-feedback-learning loop.

[0148] Step 5: Trigger decision output and safety interlock;

[0149] 5.1 Trigger Detection:

[0150] Ctrig is updated once per sliding window and compared with the current dynamic threshold θ:

[0151] ;

[0152] 5.2 Safety interlock conditions:

[0153] Even if the confidence level is passed, the system must still meet all of the following safety interlock conditions before it can output the final trigger signal:

[0154] Condition 1 (Wind Speed ​​Interlock): Real-time wind speed W s < W s Limits are set to prevent drones from becoming unstable or the power lines from swaying excessively in strong winds.

[0155] Condition 2 (Battery Interlock): The drone's remaining battery capacity Brem > Bmin, ensuring sufficient power for a safe return after triggering.

[0156] Condition 3 (Conductor Voltage Interlock): The peak value of the induced voltage in the conductor, Vmax, is less than Vlimit, to prevent equipment breakdown and damage under extreme high field conditions.

[0157] Condition 4 (Lightning Distance Interlock): The nearest lightning distance Dmin > Dsafe, to prevent natural lightning from directly hitting the drone or operator.

[0158] Condition 5 (Flight Control Status Interlock): No alarms in the flight control system and attitude angle deviations within the allowable range.

[0159] The five conditions are logically and formally linked to arrive at the final triggering decision:

[0160] If all conditions are met, a trigger command is output; if any condition is not met, a wait command is output, and the specific category of the unmet condition is displayed on the human-computer interaction interface.

[0161] Example 2:

[0162] The computer-readable storage medium of this embodiment stores a computer program that, when executed by a processor, implements the steps in the adaptive decision-making method for optimal lightning triggering timing based on an unmanned aerial vehicle platform in Embodiment 1.

[0163] The computer-readable storage medium in this embodiment can be an internal storage unit of the terminal, such as the terminal's hard disk or memory; the computer-readable storage medium in this embodiment can also be an external storage device of the terminal, such as a plug-in hard disk, smart memory card, secure digital card, flash memory card, etc. equipped on the terminal; furthermore, the computer-readable storage medium can include both the terminal's internal storage unit and external storage devices.

[0164] The computer-readable storage medium of this embodiment is used to store computer programs and other programs and data required by the terminal. The computer-readable storage medium can also be used to temporarily store data that has been output or will be output.

[0165] Example 3:

[0166] The computer device of this embodiment includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the steps in the adaptive decision-making method for optimal lightning triggering time based on a drone platform in Embodiment 1.

[0167] In this embodiment, the processor can be a central processing unit, or other general-purpose processors, digital signal processors, application-specific integrated circuits, off-the-shelf programmable gate arrays or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor, etc. The memory can include read-only memory and random access memory, and provides instructions and data to the processor. A portion of the memory can also include non-volatile random access memory. For example, the memory can also store device type information.

[0168] Those skilled in the art will understand that the content disclosed in the embodiments can be provided as a method, system, or computer program product. Therefore, this solution can take the form of a hardware embodiment, a software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this solution can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage and optical storage) containing computer-usable program code.

[0169] This solution is described with reference to flowchart illustrations and / or block diagrams of methods and computer program products according to embodiments of this solution. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing device, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0170] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0171] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0172] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0173] The examples described herein are merely preferred embodiments of the invention and are not intended to limit the concept and scope of the invention. Any modifications and improvements made by those skilled in the art to the technical solutions of the invention without departing from the design concept of the invention should fall within the protection scope of the invention.

[0174] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the specific embodiments described above. The specific embodiments and descriptions in the specification are merely for further illustrating the principles of the invention. Various changes and modifications can be made to the present invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the claims and their equivalents.

Claims

1. An adaptive decision-making method for optimal lightning triggering timing based on an unmanned aerial vehicle (UAV) platform, characterized in that, Includes the following steps: Real-time acquisition and preprocessing of multi-source electric field data: Simultaneous acquisition of high-altitude electric field E using an airborne electric field sensor on a UAV and a ground-based electric field mill. a (t) and near-ground electric field Eg(t), while simultaneously collecting conductor induced voltage Vw(t), meteorological parameters and regional lightning location data; The acquired signals are sequentially subjected to notch filtering based on rotor speed adaptive tracking to eliminate rotor interference, wavelet decomposition and soft threshold denoising to suppress random noise, and multi-channel timestamp alignment is performed according to GPS timing reference. Multidimensional feature extraction: Within a sliding time window of width Tw, extract the mean ground electric field Ēg and the mean airborne electric field Ē. a Ground electric field change rate (kg), airborne electric field standard deviation (σ) a Nine characteristic quantities are included: vertical electric field gradient Gᵥ, peak value of conductor induced voltage Vmax, lightning activity density ρL, nearest lightning distance Dmin, and meteorological synergy index Mc. Multi-level fusion criterion calculation: combining nine feature quantities with real-time wind speed W s The platform state parameters, namely the battery remaining capacity (Brem), are divided into three layers: the electric field core layer, the environmental coordination layer, and the safety constraint layer. Within each layer, the features are normalized and then weighted to obtain the layer scores SC, SE, and SS. Finally, the inter-layer weights are used to weight and fuse the results to obtain the comprehensive trigger confidence score. ； C trig To synthesize the trigger confidence level, it is compared with the dynamic trigger threshold θ to determine whether to trigger a discharge; β C β E β S These are the interlayer weights for the electric field core layer, the environmental coordination layer, and the safety constraint layer, respectively, satisfying β. C +β E +β S = 1; S C S E S S This is a weighted score for each feature within the three levels; Adaptive weight update and threshold adjustment: After each trigger attempt, the feature weights within each layer and the weights between layers are updated according to the gradient direction rule based on the actual trigger result y. w j (m) w represents the current weight of the j-th feature in the m-th iteration. j (m+1) The updated weights; η is the learning rate, η > 0, which controls the step size of each weight adjustment; y (m) For the m-th trigger, the binary actual result is 1 = success, 0 = failure; C trig (m) The overall trigger confidence level at the m-th trigger; (y (m) -C trig (m) That is, prediction error; The dimensionless value of the j-th feature after direction-consistent normalization at the m-th trigger; It performs non-negative truncation and normalization, and adaptively adjusts the dynamic trigger threshold θ according to the asymmetric rule of lowering it on success and raising it on failure; when the system fails to meet the trigger conditions for multiple consecutive windows, the threshold is automatically lowered through the idle period threshold decay mechanism to prevent cold start lock-up. Trigger decision output and safety interlock: Compare Ctrig with the threshold θ, and output a trigger command under the premise that the five safety interlock conditions of wind speed, battery, wire voltage, lightning safety distance and flight control status are met at the same time.

2. The method as described in claim 1, characterized in that, In the real-time acquisition and preprocessing of multi-source electric field data, the specific method for eliminating rotor interference is as follows: the fundamental frequency of interference fᵣ = nᵣ × Nb is calculated based on the rotor speed nᵣ and the number of blades Nb reported by the flight control system in real time. Notch filters are constructed for fᵣ, 2fᵣ and 3fᵣ respectively. The three notch filters are cascaded to form an adaptive filter group, and its center frequency is adjusted in real time with the change of rotational speed.

3. The method as described in claim 2, characterized in that, The formula for calculating the vertical electric field gradient is: Where h is the flight altitude of the UAV, and Gᵥ reflects the steepness of the cloud-to-ground spatial potential descent.

4. The method as described in claim 3, characterized in that, Meteorological Coordination Index: , where X̃ k The normalized values ​​for each meteorological parameter are defined. The normalization direction for each parameter is determined according to the principle of mapping to a larger value that is more conducive to lightning triggering. The normalization results are truncated and restricted to the interval [0, 1]. α k The coefficient of coordination and Σα k = 1.

5. The method as described in claim 4, characterized in that, The method for direction consistency normalization is as follows: For positive vectors, the following is applied: For the inverse vector, the following is applied: Where F j F represents the original measurement value of the j-th feature; j ref This is the reference range value for the j-th feature. This is the normalized dimensionless relative value, with a range of [0, 1].

6. The method as described in claim 5, characterized in that, After the weights are updated, non-negative truncation and normalization are performed: ； Where ε>0 is the lower limit protection value for weights, which prevents any feature weights from being completely pushed to zero.

7. The method as described in claim 6, characterized in that, The adaptive adjustment formula for the trigger threshold is: Define the base value of the adjustment amount. Where γ > 0 is the bias adjustment term; when the trigger is successful, θ←θ−δ s ·Δ, when triggering fails, θ←θ +δf·Δ, where δf > δ s > 0 to reflect the principle of safety first, the threshold is limited to the range of [θmin, θmax].

8. The method as described in claim 7, characterized in that, The five safety interlock conditions in the trigger decision output and safety interlock are: wind speed is lower than the wind speed limit, battery capacity is higher than the minimum charge, wire voltage is lower than the voltage limit, the distance to the nearest lightning is greater than the safe distance, and the flight control system is in normal condition. The five conditions are applied to the trigger decision in a logical AND manner. If any condition is not met, the trigger output is automatically suppressed.

9. The method as described in claim 8, characterized in that, Maintain an idle window counter Nidle, which increments only when the overall trigger confidence is below a threshold. When the number of consecutively untriggered sliding windows exceeds the preset threshold Nidle,max, the trigger threshold is lowered by a fixed step size δidle, but not lower than the lower limit θmin. The counter is reset to zero and starts counting again after each decay.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by the processor, it implements the steps in the adaptive decision-making method for optimal lightning triggering timing based on a drone platform as described in any one of claims 1 to 9.

Images