Unmanned aerial vehicle aerodynamic deceleration control prediction method, electronic device and storage medium

Abstract

The unmanned aerial vehicle aerodynamic deceleration control prediction method, the electronic equipment and the storage medium belong to the unmanned aerial vehicle flight management technical field. In order to solve the reliability of increasing the unmanned aerial vehicle flight deceleration control strategy, the present application comprises a hidden layer of an unmanned aerial vehicle aerodynamic deceleration control prediction model, adopts a three-layer architecture of a dynamic perception layer, an aerodynamic characteristic layer and a control decision layer; the output index of the model is constructed, including flight state prediction index and control effect evaluation index. A loss function including speed loss term function, flight characteristic loss term function and stability loss term function is constructed. Based on the input feature parameter data of the normalized unmanned aerial vehicle aerodynamic deceleration control prediction model, a training set, a validation set and a test set are constructed, the parameter setting of the constructed unmanned aerial vehicle aerodynamic deceleration control prediction model is carried out, then the training set is obtained, the training set is used for training, the validation set is used for verification, the test set is used for testing and prediction.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of unmanned aerial vehicle flight management, and particularly relates to an unmanned aerial vehicle aerodynamic deceleration control prediction method, an electronic device and a storage medium. BACKGROUND

[0002] Unmanned aerial vehicle aerodynamic deceleration control is a key technology for safe and efficient landing of aircraft. By reasonably utilizing aerodynamic characteristics, coordinating control of ailerons, rudders, elevators, air brakes and deceleration parachutes and other devices, the unmanned aerial vehicle can smoothly transition from high-speed flight to low-speed flight while maintaining a stable attitude. This is of great significance for improving the success rate of unmanned aerial vehicle recovery, protecting on-board equipment, prolonging service life and ensuring mission integrity. However, the deep learning models widely used at present can establish a mapping relationship between input parameters and output results, but often regard the entire control process as a black box, lacking physical interpretation of the intermediate process. Such models cannot effectively describe the variation of key physical quantities such as aerodynamics and aerodynamic characteristics, and it is also difficult to depict the specific action mechanism between control parameters and flight states. The lack of interpretability makes it difficult to guarantee the reliability and robustness of the model, especially when faced with complex and variable flight environments, which may produce unpredictable control errors.

[0003] Traditional deep learning models often ignore the dynamic changes of aerodynamic characteristics with flight states when dealing with unmanned aerial vehicle aerodynamic deceleration control problems. These models rely solely on data-driven and cannot accurately capture the influence mechanism of environmental factors such as air density, temperature and wind speed on aerodynamic characteristics. At the same time, the model also cannot reflect the physical correlation between key parameters such as Mach number and dynamic pressure and deceleration performance. This neglect of physical processes may result in unreasonable control strategies and even output results that violate the basic principles of aerodynamics in actual application.

[0004] In the process of unmanned aerial vehicle deceleration control, accurate prediction of airspeed, angle of attack, altitude and other flight state parameters is of great significance. These parameters are directly related to flight safety and mission effectiveness, especially during the critical stage of high-speed flight to low-speed flight transition. However, existing models often lack accurate grasp of the trend of these parameters, and cannot fully consider the coupling effect of various control surfaces (such as ailerons, rudders, elevators, etc.) and their interaction with aerodynamic characteristics, which seriously restricts the accuracy and reliability of deceleration control.

[0005] The parameters of deceleration rate, aerodynamic drag coefficient, lift-drag ratio, etc. are crucial for achieving precise deceleration control. These indicators not only reflect the effect of the deceleration process, but also directly affect the energy management efficiency and flight safety. However, existing models often predict these indicators separately, ignoring their physical correlation, resulting in a lack of systematization and coordination in the prediction results. This fragmented prediction method cannot guarantee the overall optimization of the control strategy, and it is difficult to meet the needs of actual engineering applications. SUMMARY

[0006] The problem to be solved by the present application is to increase the reliability of unmanned aerial vehicle flight deceleration control strategy, and to propose an unmanned aerial vehicle aerodynamic deceleration control prediction method, an electronic device and a storage medium.

[0007] To achieve the above-mentioned purpose, the present application realizes the following technical scheme:

[0008] An unmanned aerial vehicle aerodynamic deceleration control prediction method, comprising the following steps:

[0009] S1. Collecting input feature parameters of the unmanned aerial vehicle aerodynamic deceleration control prediction model and performing normalization processing, including flight state parameters, control input parameters and environmental parameters;

[0010] S2. Constructing a hidden layer of the unmanned aerial vehicle aerodynamic deceleration control prediction model, adopting a three-layer architecture of dynamic perception layer, aerodynamic characteristic layer and control decision layer;

[0011] S3. Constructing output indicators of the model based on the hidden layer of the unmanned aerial vehicle aerodynamic deceleration control prediction model obtained in step S2, including flight state prediction indicators and control effect evaluation indicators; the flight state prediction indicators include predicted airspeed Y1, predicted angle of attack Y2, predicted height Y3, predicted Mach number Y4 and predicted dynamic pressure Y5; the control effect evaluation indicators include deceleration rate Y6, aerodynamic drag coefficient Y7, lift-drag ratio Y8, energy loss rate Y9 and stability index Y 10 ; establishing a mapping relationship between the model hidden layer and the output indicators;

[0012] S4. Constructing a loss function of the unmanned aerial vehicle aerodynamic deceleration control prediction model as a composite loss function of a basic loss function, a speed loss term function, a flight characteristic loss term function and a stability loss term function;

[0013] S5. Based on the input feature parameter data of the normalized UAV air dynamic deceleration control prediction model obtained in step S1 and the output index data of the normalized UAV air dynamic deceleration control prediction model obtained in step S3, a training set, a verification set and a test set are constructed, the parameter of the constructed UAV air dynamic deceleration control prediction model is set, then the training set is used to train the UAV air dynamic deceleration control prediction model, the verification set is used to verify the UAV air dynamic deceleration control prediction model, and the test set is used to test the UAV air dynamic deceleration control prediction model after verification.

[0014] Further, the input feature parameters of the UAV air dynamic deceleration control prediction model in step S1 are as follows:

[0015] The flight state parameters include airspeed X1, attack angle X2, sideslip angle X3, height X4, Mach number X5, dynamic pressure X6, roll angle X7 and pitch angle X8;

[0016] The control input parameters include aileron deflection angle X9, rudder deflection angle X 10 , elevator deflection angle X 11 , throttle opening X 12 , air brake deflection angle X 13 and deceleration parachute deployment area X 14 ;

[0017] The environmental parameters include air density X 15 , temperature X 16 , wind speed X 17 , wind direction X 18 , atmospheric pressure X 19 and humidity X 20 .

[0018] Further, the specific implementation method of step S2 includes the following steps:

[0019] S2.1. Based on considering the airspeed and height in the flight state parameters, and considering the influence of environmental parameters on the dynamic characteristics, a dynamic perception layer is constructed, and the expression is:

[0020]

[0021] wherein, is the output of the dynamic perception layer, is the activation function, is the weight coefficient of the input feature parameter of the i-th UAV air dynamic deceleration control prediction model, X i is the input feature parameter of the i-th UAV air dynamic deceleration control prediction model, is the dynamic characteristic coefficient, is the height influence factor, is the characteristic time scale factor, which characterizes the characteristic response speed of the UAV during deceleration; is the characteristic height scale factor, which represents the sensitivity of the UAV to altitude changes; is the bias term of the dynamic perception layer, i=1~20;

[0022] Dynamic characteristic coefficient The expression reflecting the dynamic characteristics of the UAV during deceleration is:

[0023]

[0024] in, For the quality of the drone, is the drag coefficient;

[0025] High impact factor The relationship between altitude and atmospheric pressure is expressed as:

[0026]

[0027] in, is the reference height, is standard atmospheric pressure;

[0028] S2.2. Based on the dynamic characteristics, the effects of angle of attack and dynamic pressure on aerodynamic forces are considered, and the Mach number effect is considered to construct the aerodynamic characteristic layer. The expression is:

[0029]

[0030] in, Output for the aerodynamic characteristics layer; is the bias term of the aerodynamic characteristic layer, is the lift characteristic coefficient; is the compressibility effect factor, is the characteristic angle of attack, which characterizes the response characteristics of the system to attitude changes during deceleration; is the characteristic dynamic pressure, which represents the adaptability of the UAV to aerodynamic changes;

[0031] lift characteristic coefficient The aerodynamic characteristics are reflected by the ratio of dynamic pressure, parachute deployment area and gravity, and the expression is:

[0032]

[0033] Where g is the acceleration due to gravity;

[0034] Compressibility effect factor Considering the influence of air density and Mach number on the compressibility effect, the expression is:

[0035] ;

[0036] S2.3. Constructing the control decision layer, realizing the rapid mapping of the initial dynamic state to the control decision through the establishment of a direct information channel from the dynamic perception layer to the control decision layer, the expression is:

[0037]

[0038] wherein, is the output of the control decision layer; is the residual connection coefficient, determined by expert experience; is the control efficiency coefficient; is the stability influence factor, is the bias term of the control decision layer; is the characteristic force, representing the influence of the control force change on the UAV; is the characteristic rudder angle, representing the influence of the rudder deflection on the UAV;

[0039] Control efficiency coefficient Reflect the control efficiency by considering the relationship between dynamic pressure and characteristic length, and the product of dynamic pressure and parachute deployment area represents the aerodynamic control force, the expression is:

[0040]

[0041] wherein, is the characteristic length, is the average aerodynamic chord length;

[0042] Stability influence factor Construct a stability evaluation index through the Mach number, the expression is:

[0043] .

[0044] Further, the specific implementation method of step S3 is to combine the outputs of the three hidden layers constructed in step S2, and introduce a nonlinear activation function and a characteristic scale factor of the corresponding index, to realize the accurate prediction of the flight state prediction index and the control effect evaluation index. The nonlinear activation function introduced for the flight state prediction index is the hyperbolic tangent function, and the nonlinear activation function introduced for the control effect evaluation index is the sigmoid function.

[0045] Further, the specific implementation method of step S4 includes the following steps:

[0046] S4.1. Designing a basic loss function in the form of mean square error, calculating all output indexes Y1~Y 10The mean square error between the predicted value and the measured value of the speed loss term is the basic loss ;

[0047] S4.2. The speed loss term function is designed to evaluate the difference between the predicted airspeed and the actual airspeed, and the corresponding state transition efficiency to quantify the deceleration effect, while considering the influence of air density and the deployment area of the parachute, and the expression is:

[0048]

[0049] wherein, is the speed loss term weight coefficient; is the state transition efficiency weight coefficient; , respectively represent the kinetic energy level of the predicted state and the current state; , respectively represent the work level of the predicted resistance and the actual resistance;

[0050] The speed loss term weight coefficient reflects the variation of the deceleration effect with the flight conditions, and the air density prediction speed square term represents the kinetic energy variation characteristics. Considering the air density, the deployment area of the parachute, and the mass factor, the expression is:

[0051]

[0052] The state transition efficiency weight coefficient evaluates the energy conversion efficiency of the deceleration process, considering the inertia characteristics of the unmanned aerial vehicle, the characteristics of the deceleration medium, the current motion state, and the characteristics of the deceleration device. The expression of the state transition efficiency weight coefficient is:

[0053] ;

[0054] S4.3. The flight characteristic loss term function is designed to consider the influence of Mach number effect on aerodynamic characteristics, and to evaluate the accuracy of lift, drag and moment prediction. By introducing characteristic length and aerodynamic parameters, a comprehensive evaluation index of flight quality is constructed, and the expression of the flight characteristic loss term function is:

[0055]

[0056] wherein, is the Mach number effect coefficient, is the moment weight coefficient, , respectively represent the predicted aerodynamic force intensity level and the actual aerodynamic force intensity level; , respectively represent the predicted moment level and the actual moment level;

[0057] The Mach number effect coefficient reflects the change of aerodynamic characteristics in the process of flight speed change through the Mach number, and directly reflects the strength of the Mach number effect. The expression is:

[0058]

[0059] The moment weight coefficient considers the influence of environmental characteristics, parachute area and characteristic length, and the characteristic length evaluates the influence of moment balance. The predicted speed square reflects the dynamic pressure effect. The expression is:

[0060]

[0061] S4.4. Design the stability loss function. By evaluating the height change, static stability and dynamic stability, a complete stability evaluation system is constructed. The expression of the stability loss function is:

[0062]

[0063] wherein, is the height influence coefficient, is the stability weight coefficient, and respectively represent the predicted state mechanical energy level and the current state mechanical energy level; and respectively represent the predicted stability moment level and the actual stability moment level, is the static stability coefficient; is the characteristic energy factor; is the characteristic moment factor;

[0064] The height influence coefficient represents the influence of height. The expression is:

[0065]

[0066] The stability weight coefficient evaluates the influence of stability. The predicted speed square reflects the dynamic pressure effect. The influence of environmental characteristics, parachute area, average aerodynamic chord and characteristic length is also considered. The expression is:

[0067]

[0068] S4.5. A loss function for predicting the model of the air dynamic deceleration control of the unmanned aerial vehicle is constructed as a composite loss function including the speed loss term, the flight characteristic loss term and the stability loss term. The expression is:

[0069]

[0070] wherein, , ,​​​ , respectively , , , weight coefficients.

[0071] Further, in step S5, the input feature parameter data of the normalized UAV air dynamic deceleration control prediction model and the output index data of the normalized UAV air dynamic deceleration control prediction model obtained in step S3 are divided into a training set, a validation set and a test set in a ratio of 8:1:1.

[0072] Further, in step S5, the model training adopts a back propagation algorithm, and in the training process, the model calculates a prediction value through forward propagation, substitutes the prediction value and a true value into a loss function to calculate a loss, and then updates network parameters through back propagation.

[0073] Further, in step S5, in order to prevent overfitting, an early stopping strategy is adopted, and when the loss of the validation set does not decrease for 10 consecutive epochs, the training is stopped; after the model training is completed, the relationship between the input and the output is established, and the output result is restored to an actual physical quantity through reverse normalization.

[0074] An electronic device comprises a memory and a processor, the memory stores a computer program, and the processor realizes the steps of the UAV air dynamic deceleration control prediction method when executing the computer program.

[0075] A computer readable storage medium stores a computer program, and the computer program is executed by a processor to realize the UAV air dynamic deceleration control prediction method.

[0076] Advantages of the present application:

[0077] The UAV air dynamic deceleration control prediction method provided by the present application establishes a deep learning prediction model based on a physical mechanism, which can accurately reflect the physical characteristics in the air dynamic deceleration process of the UAV and accurately predict key state parameters and performance indicators. The model needs to organically combine dynamic perception, aerodynamic characteristics and control decision to construct a multi-level prediction system, while ensuring that the model has good interpretability and generalization ability. Through reasonable loss function design, the model can accurately capture the coupling relationship between various physical parameters, and ensure that the prediction result conforms to the basic principles of aerodynamics.

[0078] The unmanned aerial vehicle aerodynamic deceleration control prediction method provided by the application realizes in-depth description of the physical mechanism of the deceleration process by constructing a three-layer architecture (a kinetic perception layer, an aerodynamic characteristic layer, and a control decision layer). Multi-dimensional and multi-level feature selection is adopted to ensure that the model can comprehensively perceive the flight state and effectively respond to environmental changes. Through a carefully designed loss function system, comprehensive evaluation of the speed change characteristics, flight characteristics, and stability is realized. The method not only improves the prediction accuracy, but also enhances the interpretability and practicality of the model, providing a reliable basis for the optimization of the unmanned aerial vehicle deceleration control strategy. BRIEF DESCRIPTION OF DRAWINGS

[0079] Figure 1 A flowchart of the unmanned aerial vehicle aerodynamic deceleration control prediction method provided by the application;

[0080] Figure 2 A loss curve diagram of the training process of the unmanned aerial vehicle aerodynamic deceleration control prediction model provided by the application;

[0081] Figure 3 A comparison diagram of the predicted value and the true value of the stability index of the unmanned aerial vehicle aerodynamic deceleration control prediction model provided by the application. DETAILED DESCRIPTION

[0082] In order to make the purpose, technical solutions and advantages of the application clearer, the application will be further described in detail below in combination with the drawings and specific embodiments. It should be understood that the specific embodiments described herein are only used to explain the application and do not limit the application, i.e., the described specific embodiments are only a part of the embodiments of the application, not all the specific embodiments. The components of the specific embodiments of the application described and shown in the drawings herein can be arranged and designed in various configurations, and the application can have other embodiments.

[0083] Therefore, the detailed description of the specific embodiments of the application provided in the drawings below is not intended to limit the scope of the claimed application, but only represents selected specific embodiments of the application. Based on the specific embodiments of the application, all other specific embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the application.

[0084] In order to further understand the invention content, characteristics and effects of the application, the following specific embodiments are exemplified, and the drawings are Figure 1 -APPENDIX Figure 3 The detailed description is as follows:

[0085] Example 1:

[0086] A unmanned aerial vehicle aerodynamic deceleration control prediction method, comprising the following steps:

[0087] S1. Collect the input feature parameters of the UAV aerodynamic deceleration control prediction model and perform normalization processing, including flight state parameters, control input parameters and environmental parameters;

[0088] Further, the input feature parameters of the UAV aerodynamic deceleration control prediction model in step S1 are as follows:

[0089] The flight state parameters include airspeed X1, angle of attack X2, sideslip angle X3, altitude X4, Mach number X5, dynamic pressure X6, roll angle X7 and pitch angle X8;

[0090] Further, airspeed X1, angle of attack X2, sideslip angle X3 and altitude X4 directly reflect the instantaneous motion state of the UAV and are the basis for deceleration control; Mach number X5 and dynamic pressure X6 reflect the aerodynamic characteristics and are crucial for evaluating transonic effects and structural loads; roll angle X7 and pitch angle X8 ensure attitude stability during deceleration;

[0091] The control input parameters include aileron deflection angle X9, rudder deflection angle X 10 , elevator deflection angle X 11 , throttle opening X 12 , air brake deflection angle X 13 and deceleration parachute deployment area X 14 ; these parameters constitute the active control variables during deceleration;

[0092] The environmental parameters include air density X 15 , temperature X 16 , wind speed X 17 , wind direction X 18 , atmospheric pressure X 19 and humidity X 20 . Real-time monitoring of these parameters is crucial for developing adaptive control strategies.

[0093] Further, this multi-dimensional and multi-level feature selection ensures that the model can comprehensively perceive flight states, accurately control actuators and effectively respond to environmental changes, thereby achieving safe and efficient deceleration control. As can be seen from the above parameters, during the training process of the deep learning model, different input parameters have different physical dimensions and numerical ranges, such as airspeed which may be in the order of hundreds of meters per second, while the angle of attack only varies within a few degrees. This numerical difference can cause uneven gradient updates during model training, affecting the convergence speed and accuracy of the model. Therefore, normalization processing is required for all input and output parameters to map them to the [0, 1] interval, making different features numerically comparable and avoiding certain features dominating the model training process due to their large numerical values. This also helps to reduce rounding errors in numerical calculations and improve the numerical stability of the model.

[0094] Specifically, based on the maximum and minimum value range of each parameter obtained by statistical history flight data, the maximum and minimum value normalization method is used for processing: in the flight state parameter, the airspeed adopts the ratio of the actual value to the historical maximum airspeed, the attack angle and the sideslip angle are mapped based on the historical data statistical range, the height adopts the ratio of the actual height to the historical maximum flight height, the Mach number and the dynamic pressure respectively adopt the ratio of the actual value to the historical maximum value; in the attitude angle parameter, the roll angle and the pitch angle are mapped based on the historical attitude data range; the control surface deflection angle (including aileron, rudder, elevator and brake plate) adopts the ratio of the actual deflection angle to the physical limit range; the throttle opening is normalized based on the actual control range, and the parachute deployment area adopts the ratio of the actual deployment area to the maximum design area; in terms of environmental parameters, the air density and atmospheric pressure adopt the ratio of the actual value in the historical data statistical range, the temperature is mapped based on the historical meteorological data range, the wind speed adopts the ratio of the actual wind speed to the historical maximum wind speed, the wind direction angle is mapped based on the whole circle angle, and the humidity is normalized based on the actual percentage range.

[0095] Table 1 is part of the input data of the unmanned aerial vehicle aerodynamic deceleration control prediction model;

[0096]

[0097] In addition, it should be noted that, in order to ensure the consistency of physical parameters in the model, all physical quantities involved in the embodiment are first normalized and then operated. This normalization processing based on historical data statistics not only ensures the numerical stability and convergence efficiency of the deep learning model training, but also enables the model to better capture the internal relationship between different physical quantities and improve the prediction accuracy. In the process of using the model, the input data and the output data are also normalized data.

[0098] S2. Construct a hidden layer of an unmanned aerial vehicle aerodynamic deceleration control prediction model, adopt a three-layer architecture of dynamic perception layer, aerodynamic characteristic layer and control decision layer;

[0099] The dynamic perception layer mainly processes the basic kinematics and dynamics. The aerodynamic characteristic layer focuses on the processing of aerodynamic effects. The control decision layer integrates the information of the previous two layers and focuses on the influence of control input on the system. This hierarchical structure not only conforms to the internal law of physical phenomena, but also can gradually process various aspects from basic physics to complex control, ensuring the interpretability of the model and the accuracy of the control. Each layer has a clear physical meaning and functional positioning, which greatly improves the training efficiency and generalization ability of the model.

[0100] Further, the specific implementation method of step S2 includes the following steps:

[0101] S2.1. Based on considering the airspeed and height in the flight state parameters, and considering the influence of environmental parameters on the dynamic characteristics, the dynamic perception layer is constructed, and the expression is:

[0102]

[0103] wherein, is the output of the dynamic perception layer, is the activation function, is the weight coefficient of the input feature parameter of the i-th unmanned aerial vehicle air dynamic deceleration control prediction model, X i is the input feature parameter of the i-th unmanned aerial vehicle air dynamic deceleration control prediction model, is the dynamic characteristic coefficient, is the height influence factor, is the characteristic time scale factor, which represents the characteristic response speed of the unmanned aerial vehicle in the deceleration process; is the characteristic height scale factor, which represents the sensitivity of the unmanned aerial vehicle to the change of height; is the bias term of the dynamic perception layer, i=1~20;

[0104] Dynamic characteristic coefficient reflects the dynamic characteristics of the unmanned aerial vehicle in the deceleration process, and the expression is:

[0105]

[0106] wherein, is the mass of the unmanned aerial vehicle, is the drag coefficient; air density affects the size of air dynamic, the square term of airspeed embodies the effect of dynamic pressure, the deployment area of deceleration parachute represents the effective action area, and the drag coefficient represents the air dynamic characteristics, while considering the influence of attack angle on the dynamic characteristics;

[0107] Height influence factor reflects the relationship between height and atmospheric pressure, and the expression is:

[0108]

[0109] wherein, is the reference height, is the standard atmospheric pressure;

[0110] Height influence factor reflects the relationship between height and atmospheric pressure. Here, the basic law of atmospheric physics is considered: with the increase of height, the atmospheric density decays exponentially, which directly affects the aerodynamic characteristics of the aircraft. At the same time, the ratio of the atmospheric pressure in the actual flight process to the standard atmospheric pressure can directly reflect the atmospheric environmental state at the current flight height, which has a significant influence on the deceleration performance of the aircraft;

[0111] S2.2. The influence of attack angle and dynamic pressure on aerodynamic force is considered based on the dynamic characteristics, and the Mach number effect is considered to construct the aerodynamic characteristic layer, which is expressed as:

[0112]

[0113] wherein, is the output of the aerodynamic characteristic layer; is the bias term of the aerodynamic characteristic layer, is the lift characteristic coefficient; is the compressibility effect factor, is the characteristic attack angle, representing the response characteristics of the system to attitude changes during deceleration; is the characteristic dynamic pressure, representing the adaptability of the UAV to changes in aerodynamic force;

[0114] Lift characteristic coefficient The aerodynamic force characteristics are reflected by the ratio of dynamic pressure, deceleration parachute deployment area and gravity, which is expressed as:

[0115]

[0116] wherein, g is the acceleration of gravity;

[0117] Compressibility effect factor The influence of air density and Mach number on compressibility effect is considered, which is expressed as:

[0118] ;

[0119] S2.3. The control decision layer is constructed, and the direct information channel from the dynamic perception layer to the control decision layer is established to realize the rapid mapping of the initial dynamic state to the control decision, which is expressed as:

[0120]

[0121] wherein, is the output of the control decision layer; is the residual connection coefficient, which is determined by expert experience; is the control efficiency coefficient; is the stability influence factor, is the bias term of the control decision layer; is the characteristic force, representing the influence of control force change on the UAV; is the characteristic rudder angle, representing the influence of rudder deflection on the UAV;

[0122] Control efficiency coefficient The control efficiency is reflected by considering the relationship between the dynamic pressure and the characteristic length, and the product of the dynamic pressure and the deployment area of the drag parachute represents the aerodynamic control force, and the expression is:

[0123]

[0124] wherein, is the characteristic length, is the average aerodynamic chord length;

[0125] Stability influence factor The stability evaluation index is constructed by the Mach number, when the Mach number is close to 1, a more conservative control strategy is needed, and when the Mach number is smaller, a more aggressive control behavior is allowed, and the expression is:

[0126] .

[0127] S3. Based on the output indicators of the hidden layer model of the unmanned aerial vehicle aerodynamic deceleration control prediction model obtained in step S2, flight state prediction indicators and control effect evaluation indicators are included; the flight state prediction indicators include predicted airspeed Y1, predicted attack angle Y2, predicted height Y3, predicted Mach number Y4 and predicted dynamic pressure Y5; the control effect evaluation indicators include deceleration rate Y6, aerodynamic drag coefficient Y7, lift-drag ratio Y8, energy loss rate Y9 and stability index Y 10 ; a mapping relationship between the model hidden layer and the output indicators is established;

[0128] Further, the specific implementation method of step S3 is to combine the outputs of the three hidden layers constructed in step S2 by weighting, and introduce a nonlinear activation function and a characteristic scale factor of the corresponding indicator, so as to realize accurate prediction of the flight state prediction indicators and the control effect evaluation indicators; the nonlinear activation function introduced for the flight state prediction indicators is the hyperbolic tangent function, and the nonlinear activation function introduced for the control effect evaluation indicators is the sigmoid function.

[0129] Further, for the output parameters, the predicted flight state quantities adopt the same normalization method as the input parameters, and the performance indicators are mapped according to the change range obtained by statistical data, and are mapped to [0, 1]. Based on the relationship between the outputs of the hidden layers (dynamic perception layer, aerodynamic characteristic layer, control decision layer) and the final control target, a mapping relationship between the output parameters and the hidden layers needs to be established. Considering the different characteristics of the flight state prediction (Y1-Y5) and the control effect evaluation (Y6-Y10) in the deceleration control process, the outputs of the three hidden layers are combined by weighting, and a nonlinear activation function is introduced, so as to realize accurate prediction of different types of output parameters.

[0130] Among them, the dynamic perception layer provides basic state information, the aerodynamic characteristic layer provides nonlinear aerodynamic characteristics, and the control decision layer provides unmanned aerial vehicle system response characteristics. The three layers of information are fused through weight coefficients. In specific implementation, for each output index, the information from the three hidden layers is processed in a weighted combination manner, and a nonlinear activation function is introduced for output mapping to ensure the physical rationality of the prediction and evaluation results.

[0131] Flight state prediction focuses on the prediction of key physical quantities such as airspeed Y1 (reflecting the direct effect of deceleration), predicted attack angle Y2 (monitoring stall risk), predicted height Y3 (ensuring safety margin), predicted Mach number Y4, and predicted dynamic pressure Y5 (evaluating aerodynamic characteristics);

[0132] The expression of the flight state prediction index is:

[0133]

[0134] Among them, , , are the weight coefficients corresponding to the dynamic perception layer, the aerodynamic characteristic layer, and the control decision layer in the flight state prediction process, respectively; is the characteristic scale factor of each index, determined by expert experience, i=1~5; tanh is the hyperbolic tangent activation function;

[0135] Control effect evaluation focuses on the evaluation of performance indicators such as deceleration rate Y6 (quantifying control effect), aerodynamic drag coefficient Y7 (evaluating deceleration performance), lift-drag ratio Y8 (measuring efficiency), energy loss rate Y9 (optimizing energy consumption), and stability index Y 10 (ensuring safety). By using a nonlinear activation function to process the weighted combination results, both the physical rationality of the output and the accurate prediction and evaluation of different types of indicators are achieved;

[0136] The expression of the control effect evaluation index is:

[0137]

[0138] Among them, , , are the weight coefficients corresponding to the dynamic perception layer, the aerodynamic characteristic layer, and the control decision layer in the control effect evaluation process, respectively; is the characteristic scale factor of each index, determined by expert experience, i=6~10; is the activation function.

[0139] S4. The loss function for building a prediction model of the UAV aerodynamic deceleration control is a composite loss function of a basic loss function, a speed loss term function, a flight characteristic loss term function, and a stability loss term function;

[0140] Further, the specific implementation method of step S4 includes the following steps:

[0141] S4.1. Design the basic loss function in the form of mean square error, calculate the mean square error between the predicted value and the measured value of all output indicators Y1~Y 10 , and the basic loss is ;

[0142] S4.2. Design the speed loss term function, evaluate the difference between the predicted airspeed and the actual airspeed, and the corresponding state transition efficiency to quantify the deceleration effect, considering the influence of air density and deceleration parachute deployment area, and the expression is:

[0143]

[0144] wherein, is the speed loss term weight coefficient; is the state transition efficiency weight coefficient; , respectively represent the kinetic energy level of the predicted state and the current state; , respectively represent the work level of the predicted resistance and the actual resistance;

[0145] The speed loss term weight coefficient reflects the change law of the deceleration effect with the flight condition, and the air density prediction speed square term represents the kinetic energy change characteristic, and considering the air density, the deceleration parachute deployment area, and the mass factor, the expression is:

[0146]

[0147] The state transition efficiency weight coefficient evaluates the energy conversion efficiency of the deceleration process, considering the inertia characteristics of the UAV, the characteristics of the deceleration medium, the current motion state, and the characteristics of the deceleration device, and the expression of the state transition efficiency weight coefficient is:

[0148] ;

[0149] S4.3. Design the flight characteristic loss term function, consider the influence of Mach number effect on aerodynamic characteristics, and evaluate the accuracy of lift, drag and moment prediction, by introducing characteristic length and aerodynamic parameters, build a comprehensive evaluation index of flight quality, and the expression of the flight characteristic loss term function is:

[0150]

[0151] wherein, is the Mach effect coefficient, is the moment weight coefficient, , respectively represent the predicted aerodynamic force intensity level and the actual aerodynamic force intensity level; , respectively represent the predicted moment level and the actual moment level;

[0152] The Mach effect coefficient reflects the change of aerodynamic characteristics in the process of flight speed change through the Mach number, and directly reflects the strength of the Mach effect, and the expression is:

[0153] ;

[0154] The moment weight coefficient considers the influence of environmental characteristics, parachute area and characteristic length, the characteristic length evaluates the influence of moment balance, and the predicted speed square reflects the dynamic pressure effect, and the expression is:

[0155] ;

[0156] S4.4. Design the stability loss term function, evaluate the height change, static stability and dynamic stability, and build a complete stability evaluation system, and the expression of the stability loss term function is:

[0157]

[0158] wherein, is the height influence coefficient, is the stability weight coefficient, and respectively represent the predicted state mechanical energy level and the current state mechanical energy level; and respectively represent the predicted stability moment level and the actual stability moment level, is the static stability coefficient; is the characteristic energy factor; is the characteristic moment factor;

[0159] The height influence coefficient represents the influence of height, and the expression is:

[0160]

[0161] The stability weight coefficient evaluates the influence of stability, the predicted speed square reflects the dynamic pressure effect, and the influence of environmental characteristics, parachute area, average aerodynamic chord length and characteristic length is considered, and the expression is:

[0162] ;

[0163] S4.5. The loss function of the unmanned aerial vehicle aerodynamic deceleration control prediction model is a composite loss function including a speed loss term, a flight characteristic loss term and a stability loss term, and the expression is:

[0164]

[0165] wherein, , , , are weight coefficients of , , , respectively.

[0166] S5. Based on the input feature parameter data of the normalized unmanned aerial vehicle aerodynamic deceleration control prediction model obtained in step S1 and the output index data of the normalized unmanned aerial vehicle aerodynamic deceleration control prediction model obtained in step S3, a training set, a validation set and a test set are constructed, the parameters of the constructed unmanned aerial vehicle aerodynamic deceleration control prediction model are set, then the training set is used to train the unmanned aerial vehicle aerodynamic deceleration control prediction model, the validation set is used for validation, and the test set is used to test the validated unmanned aerial vehicle aerodynamic deceleration control prediction model.

[0167] Further, the data sources are commercial flight data obtained from the Civil Aviation Administration, test data obtained from large unmanned aerial vehicle manufacturers, design parameters obtained from unmanned aerial vehicle design specifications, experimental data obtained from unmanned aerial vehicle laboratories, and monitoring data obtained from unmanned aerial vehicle monitoring management platforms.

[0168] Further, in step S5, the input feature parameter data of the normalized unmanned aerial vehicle aerodynamic deceleration control prediction model and the output index data of the normalized unmanned aerial vehicle aerodynamic deceleration control prediction model obtained in step S3 are divided into a training set, a validation set and a test set in a ratio of 8:1:1.

[0169] In step S5, the model training adopts the back propagation algorithm. During the training process, the model calculates the predicted value through forward propagation, substitutes the predicted value and the true value into the loss function to calculate the loss, and then updates the network parameters through back propagation.

[0170] In step S5, to prevent overfitting, the early stopping strategy is adopted, and the training is stopped when the validation set loss does not decrease for 10 consecutive epochs. After the model training is completed, the relationship between the input and the output is established, and the output result is restored to the actual physical quantity through inverse normalization.

[0171] Further, the parameter setting of the model needs manual setting of hyperparameters based on expert experience, including network structure parameters, the number of neurons in each layer; hyperparameters that need to be manually set in the model training process, including learning rate, batch size, training rounds, weight coefficients in the loss function; parameters that need to be determined by the model during the training process: weight coefficients and bias terms of each layer.

[0172] Further, by inputting X1~X 20 The corresponding parameters, that is, Y1~Y 10 ; the output result is recovered to the actual physical quantity through inverse normalization.

[0173] Table 2 is part of the input data of the unmanned aerial vehicle aerodynamic deceleration control prediction model; Figure 2 is the loss curve diagram of the training process of the unmanned aerial vehicle aerodynamic deceleration control prediction model of the application; Figure 3 is a comparison diagram of the predicted value and the true value of the stability index of the unmanned aerial vehicle aerodynamic deceleration control prediction model of the application.

[0174] Table 2

[0175]

[0176] Example 2:

[0177] An electronic device, comprising a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the steps of the unmanned aerial vehicle aerodynamic deceleration control prediction method of any one of the embodiments 1.

[0178] The computer device of the present application can be a device comprising a processor and a memory, such as a single-chip microcomputer comprising a central processing unit, etc. The processor is used to execute the computer program stored in the memory to realize the steps of the unmanned aerial vehicle air dynamic deceleration control prediction method. The processor can be a central processing unit (CPU), and can also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) 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 mainly comprise a program storage area and a data storage area, wherein the program storage area can store an operating system, at least one application program required for a function (such as a sound playing function, an image playing function, etc.), etc.; and the data storage area can store data created according to the use of the mobile phone (such as audio data, a phone book, etc.), etc. In addition, the memory can comprise a high-speed random access memory, and can also comprise a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, at least one disk storage device, a flash memory device, or other volatile solid-state storage device. Embodiment 3:

[0179] A computer readable storage medium, having stored thereon a computer program, the computer program being executed by a processor to realize the unmanned aerial vehicle air dynamic deceleration control prediction method of embodiment 1.

[0180] The computer readable storage medium of the present application can be any form of storage medium readable by the processor of the computer device, including but not limited to non-volatile memory, volatile memory, ferroelectric memory, etc., and the computer readable storage medium stores a computer program. When the processor of the computer device reads and executes the computer program stored in the memory, the steps of the above-mentioned unmanned aerial vehicle air dynamic speed control prediction method can be realized. The computer program includes computer program code, which can be in the form of source code, object code, executable file or some intermediate form, etc. The computer readable medium can include any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal and software distribution medium, etc.

[0181] It should be noted that the relational terms such as "first" and "second" and the like are used solely to distinguish one from another entity or action, without necessarily requiring or implying any actual relationship or order between such entities or actions. Moreover, the terms "comprises", "comprising", or any other variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without more limitations, an element defined by the phrase "comprising a" does not exclude the existence of additional identical elements in the process, method, article, or apparatus that includes the element.

[0182] Although the present application has been described above with reference to specific embodiments, various modifications can be made without departing from the scope of the present application, and equivalent components can be substituted therefor. In particular, each feature in the specific embodiments disclosed by the present application can be combined with any other feature in any manner, unless there is a structural conflict. The combinations of these features are not exhaustively described in the specification, which is merely for the purpose of omitting the description and saving resources. Therefore, the present application is not limited to the specific embodiments disclosed herein, but includes all technical solutions falling within the scope of the claims.

Claims

1. A method for predicting the aerodynamic deceleration control of a UAV, characterized in that, Comprising the following steps: S1. Collect the input characteristic parameters of the unmanned aerial vehicle aerodynamic deceleration control prediction model and perform normalization processing, including flight state parameters, control input parameters and environmental parameters; S2. Construct a hidden layer of the unmanned aerial vehicle aerodynamic deceleration control prediction model, using a three-layer architecture of a dynamics perception layer, an aerodynamic characteristic layer and a control decision layer; S3. Based on the hidden layer of the unmanned aerial vehicle aerodynamic deceleration control prediction model obtained in step S2, the output indicators of the model are constructed, including flight state prediction indicators and control effect evaluation indicators; the flight state prediction indicators include predicted airspeed Y1, predicted attack angle Y2, predicted altitude Y3, predicted Mach number Y4, and predicted dynamic pressure Y5; the control effect evaluation indicators include deceleration rate Y6, aerodynamic drag coefficient Y7, lift-drag ratio Y8, energy loss rate Y9, and stability index Y 10 ; a mapping relationship between the model hidden layer and the output indicators is established; S4. Construct a loss function of the unmanned aerial vehicle aerodynamic deceleration control prediction model as a composite loss function of a basic loss function, a speed loss term function, a flight characteristic loss term function and a stability loss term function; The specific implementation method of step S4 comprises the following steps: S4.

1. Design the basic loss function, in the form of mean square error, calculate the mean square error between the predicted value and the measured value of all output indicators Y1~Y 10 , that is, the basic loss ; S4.

2. Design a speed loss term function, which quantifies the deceleration effect by evaluating the difference between the predicted airspeed and the actual airspeed, as well as the corresponding state transition efficiency, while considering the influence of air density and the deployment area of the deceleration parachute, to obtain the speed loss term function L1; S4.

3. Design a flight characteristic loss term function, which considers the influence of Mach number effect on aerodynamic characteristics, while evaluating the accuracy of lift, drag and moment prediction, by introducing characteristic length and aerodynamic parameters to construct a comprehensive evaluation index of flight quality, the flight characteristic loss term function L2; S4.

4. Design a stability loss term function, which constructs a complete stability evaluation system by evaluating height change, static stability and dynamic stability, the stability loss term function L3; S4.

5. Construct a loss function of the unmanned aerial vehicle aerodynamic deceleration control prediction model as a composite loss function including the speed loss term, the flight characteristic loss term and the stability loss term; S5. Based on the normalized input characteristic parameter data of the unmanned aerial vehicle aerodynamic deceleration control prediction model obtained in step S1 and the normalized output index data of the unmanned aerial vehicle aerodynamic deceleration control prediction model obtained in step S3, construct a training set, a validation set and a test set, set the parameters of the constructed unmanned aerial vehicle aerodynamic deceleration control prediction model, then train the unmanned aerial vehicle aerodynamic deceleration control prediction model using the obtained training set, validate it using the validation set, and test the validated unmanned aerial vehicle aerodynamic deceleration control prediction model using the test set.

2. The UAV aerodynamic deceleration control prediction method of claim 1, wherein, The input characteristic parameters of the unmanned aerial vehicle aerodynamic deceleration control prediction model in step S1 are as follows: The flight state parameters include airspeed X1, angle of attack X2, sideslip angle X3, altitude X4, Mach number X5, dynamic pressure X6, roll angle X7 and pitch angle X8; Control input parameters include aileron deflection angle X9, rudder deflection angle X 10 , elevator deflection angle X 11 , throttle opening X 12 , air brake panel deflection angle X 13 , and parachute deployment area X 14 ; The environmental parameters include air density X 15 , temperature X 16 , wind speed X 17 , wind direction X 18 , atmospheric pressure X 19 and humidity X 20 .

3. The UAV aerodynamic deceleration control prediction method of claim 2, wherein, The specific implementation method of step S2 comprises the following steps: S2.

1. Based on considering the airspeed and altitude in the flight state parameters, and considering the influence of environmental parameters on the dynamics characteristics, construct a dynamics perception layer, the expression is: ； wherein, is the output of the dynamic perception layer, is the activation function, is the weight coefficient of the input feature parameter of the i-th UAV aerodynamic deceleration control prediction model, X i is the input feature parameter of the i-th UAV aerodynamic deceleration control prediction model, is the dynamic characteristic coefficient, is the height influence factor, is the characteristic time scale factor, representing the characteristic response speed in the UAV deceleration process; is the characteristic height scale factor, representing the sensitivity of the UAV to height changes; is the bias term of the dynamic perception layer, i = 1 ~ 20; coefficient of power characteristics Reflecting the dynamics characteristics of the UAV in the deceleration process, the expression is: ； wherein, is the mass of the drone, is the drag coefficient; Highly influential factors The relationship between altitude and atmospheric pressure is embodied in the expression ； wherein is the reference height, is the standard atmospheric pressure; S2.

2. Based on the influence of angle of attack and dynamic pressure on aerodynamic force, and considering the Mach number effect, construct an aerodynamic characteristic layer, the expression is: ； wherein, is the output of the aerodynamic characteristics layer; is the bias term of the aerodynamic characteristics layer, is the lift coefficient; is the compressibility effect influence factor, is the characteristic angle of attack, representing the response characteristics of the system to the change of attitude during deceleration; is the characteristic dynamic pressure, representing the adaptability of the UAV to the change of aerodynamic force; Lift coefficient The aerodynamic force characteristics are reflected by the ratio of the dynamic pressure, the deceleration parachute deployment area and the gravity, and the expression is: ； Where g is the acceleration of gravity; Compressibility effect factor Considering the effect of air density and Mach number on compressibility, the expression is: ； S2.

3. Construct a control decision layer, which realizes the rapid mapping of the initial dynamics state to the control decision by establishing a direct information channel from the dynamics perception layer to the control decision layer, the expression is: ； wherein, is a control decision layer output; is a residual connection coefficient, determined by expert experience; is a control efficiency coefficient; is a stability influence factor, is a bias term of the control decision layer; is a characteristic force, representing the influence of the control force change on the unmanned aerial vehicle; is a characteristic rudder deflection angle, representing the influence of the rudder deflection on the unmanned aerial vehicle; Control efficiency coefficient The control efficiency is reflected by considering the relationship between the dynamic pressure and the characteristic length. The product of the dynamic pressure and the deceleration parachute deployment area represents the aerodynamic control force. The expression is: ； wherein, is a characteristic length, is an average aerodynamic chord; Stability influencing factors The stability evaluation index is constructed by the Mach number, and the expression is: 。 4. The UAV aerodynamic deceleration control prediction method of claim 3, wherein, The specific implementation method of step S3 is to combine the outputs of the three hidden layers constructed in step S2 by weighting, and introduce a nonlinear activation function and a characteristic scale factor of the corresponding index, so as to realize accurate prediction of the flight state prediction index and the control effect evaluation index. The nonlinear activation function introduced for the flight state prediction index is a hyperbolic tangent function, and the nonlinear activation function introduced for the control effect evaluation index is a sigmoid function.

5. The UAV aerodynamic deceleration control prediction method of claim 4, wherein, The expression of the loss function of the unmanned aerial vehicle aerodynamic deceleration control prediction model constructed in step S4 is: ； wherein, , , , are weight coefficients of , , , respectively.

6. The UAV aerodynamic deceleration control prediction method of claim 5, wherein, In step S5, the normalized input feature parameter data of the unmanned aerial vehicle aerodynamic deceleration control prediction model and the normalized output index data of the unmanned aerial vehicle aerodynamic deceleration control prediction model obtained in step S3 are divided into a training set, a validation set and a test set according to a ratio of 8:1:

1.

7. The UAV aerodynamic deceleration control prediction method of claim 6, wherein, In step S5, the back propagation algorithm is used for model training. In the training process, the model calculates the predicted value by forward propagation, substitutes the predicted value and the true value into the loss function to calculate the loss, and then updates the network parameters by back propagation.

8. The UAV aerodynamic deceleration control prediction method of claim 7, wherein, In step S5, in order to prevent overfitting, the early stopping strategy is adopted, and the training is stopped when the validation set loss does not decrease for 10 consecutive epochs. After the model training is completed, the relationship between the input and the output is established, and the output result is restored to the actual physical quantity by inverse normalization.

9. An electronic device, comprising: The computer program is executed by the processor to realize the steps of the unmanned aerial vehicle aerodynamic deceleration control prediction method according to any one of claims 1-8.

10. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to realize the steps of the unmanned aerial vehicle aerodynamic deceleration control prediction method according to any one of claims 1-8.

Images