Intelligent trajectory planning and cooperative control method for aircraft

Abstract

The invention discloses an aircraft intelligent trajectory planning and cooperative control method based on offline optimization-intelligent learning-online guidance, and the method comprises the steps: firstly building an aircraft motion model and an aircraft-target relative motion model, and constructing a full-airspace and full-feature-point optimal trajectory data set; secondly, parameterized representation is carried out on the three-dimensional optimal trajectory, and a proportional guidance coefficient data set is constructed; then neural network training fitting is carried out on the proportional guidance coefficient data set, and a multi-constraint guidance method based on a neural network is designed; and finally, designing a three-dimensional collaborative guidance law by combining a multi-constraint guidance method and a residual time accurate estimation technology. According to the invention, the problems of insufficient real-time performance, incapability of meeting index optimality, incapability of coping with multiple constraints, difficulty in realizing multiple flight modes and the like of the existing trajectory planning and collaborative guidance technology are solved; the multi-constraint and optimality of the trajectory are realized through offline optimization, the global adaptability and small calculation amount of the trajectory are realized through intelligent learning, and a multi-cooperative flight mode is realized through online cooperative guidance.

Description

Technical Field

[0001] The present invention belongs to the technical field of multi-aircraft collaborative control, and relates to a method for intelligent trajectory planning and collaborative control of aircraft in a three-dimensional scene, and in particular to a method for intelligent trajectory planning and collaborative control of aircraft based on offline optimization-intelligent learning-online guidance. Background Art

[0002] In recent years, the requirements for aircraft flight missions have become increasingly diverse, flight scenarios have become increasingly complex, and technical means such as interception and interference have become increasingly powerful. Collaboration and intelligence have become the hallmark technologies of the next generation of aircraft. In addition, the design process of aircraft needs to consider multiple constraints such as overload, range, terminal speed, and terminal angle. There are also multiple flight modes such as simultaneous arrival and sequential arrival, and the requirements for trajectory planning speed are becoming increasingly higher. Therefore, ensuring real-time performance and considering how to achieve multiple collaborative flight modes under multiple constraints have become key technologies to enhance aircraft capabilities.

[0003] At present, neural network learning methods have become an effective solution to the slow speed of trajectory planning. For example, the patent document with authorization announcement number CN115407664A discloses a non-programmed guidance method based on neural network training. This method first generates samples including optimal guidance instructions and optimal initial co-states, and then generates a neural network. During actual flight, the current disturbance state quantity and the deviation state quantity are used as neural network inputs, which solves the problem of insufficient real-time performance of existing trajectory planning methods. However, this method still has difficulty in coping with other challenges such as multiple constraints and multiple collaborative flight modes. Summary of the Invention

[0004] In response to the problems existing in the prior art, the present invention proposes an intelligent trajectory planning and collaborative control method for aircraft based on the combination of offline optimization, intelligent learning and online guidance, which realizes multiple collaborative flight modes under the premise of satisfying multiple constraints: first, an aircraft motion model and an aircraft-target relative motion model are established, and the entire mission airspace is discretized and modeled to obtain a feature point set. The HP adaptive pseudo-spectral method and sequential quadratic programming method are used to perform offline trajectory optimization on the feature points to construct an optimal trajectory data set; secondly, a three-dimensional optimal trajectory is fitted using time-varying proportional guidance, and the proportional guidance coefficients are parameter-optimized to construct a proportional guidance coefficient data set; then, a neural network training and fitting is performed on the proportional guidance coefficient data set, and a neural network-based multi-constraint guidance method is designed based on the proportional guidance coefficient network; finally, a three-dimensional collaborative guidance law for aircraft that satisfies multiple constraints and multiple flight modes is designed by combining the multi-constraint guidance method and accurate remaining time estimation.

[0005] To achieve the above objectives, the technical solutions provided by the present invention are:

[0006] A method for intelligent trajectory planning and coordinated control of an aircraft, characterized by comprising the following steps:

[0007] Step 1: Establish a six-degree-of-freedom motion model of the aircraft in three-dimensional space and a relative motion model between the aircraft and the target;

[0008] Step 2: Establish an optimal trajectory dataset for the entire airspace and all feature points of the aircraft in three-dimensional space;

[0009] Step 3: Parameterize the optimal trajectory of the aircraft obtained in step 2 to obtain a proportional guidance coefficient dataset;

[0010] Step 4: Perform neural network learning and fitting on the proportional guidance coefficient dataset obtained in step 3, and design a multi-constraint guidance method based on neural network;

[0011] Step 5: Based on the neural network-based multi-constraint guidance method obtained in step 4, further design a three-dimensional collaborative guidance law for the aircraft that meets multiple constraints and multiple flight modes to achieve three-dimensional collaborative control of the aircraft.

[0012] Furthermore, in step 1, the six-degree-of-freedom motion model of the aircraft includes:

[0013] The dynamic equation of the center of mass motion of the aircraft is:

[0014]

[0015] Where,

[0016] m is the mass of the aircraft, g is the acceleration of gravity, P is the engine thrust, V is the speed of the aircraft, X, Y, Z are the drag, lift and lateral force respectively, α, β are the angle of attack and sideslip angle respectively, θ, ψ v are the track inclination and track deviation, γ v is the velocity tilt angle;

[0017] Linearized model of aerodynamic forces and aerodynamic moments acting on the aircraft:

[0018]

[0019] Where,

[0020] C x 、C y 、C z are the drag coefficient, lift coefficient and lateral force coefficient respectively, M x 、M y 、M z are rolling moment, yaw moment and pitching moment respectively, m x 、m y 、m zare the rolling moment coefficient, yaw moment coefficient and pitching moment coefficient respectively, q is the dynamic pressure, S is the characteristic area, and L is the characteristic length;

[0021] Kinematic equations for the motion of the center of mass of an aircraft:

[0022]

[0023] Where,

[0024] x, y, and z are the coordinates of the aircraft's position in the launch coordinate system;

[0025] The dynamic equation of the aircraft rotating around the center of mass:

[0026]

[0027] Where,

[0028] J x 、J y 、J z are the moment of inertia of the aircraft on the three axes of the body coordinate system, ω x 、ω y 、ω z are the components of the angular velocity ω of the aircraft relative to the ground coordinate system on the three axes of the body coordinate system, M x 、M y 、M z are the components of the total external moment acting on the aircraft in the three axes of the body coordinate system, namely, the rolling moment, the yaw moment and the pitching moment;

[0029] The kinematic equations of the aircraft rotating around its center of mass are:

[0030]

[0031] Where,

[0032] ψ and γ are the pitch, yaw, and roll angles of the aircraft, respectively;

[0033] Angle geometric relationship equation:

[0034]

[0035] The relative motion model between the aircraft and the target includes:

[0036] The relative distance r between the aircraft and the target and its derivative

[0037]

[0038] Sight height angle q α , sight azimuth qβ and line of sight height and angular velocity Line of sight azimuth

[0039]

[0040] Where,

[0041] x r =x T -x,y r =y T -y,z r =z T -z, (x T ,y T ,z T )

[0042] is the component of the target position in the launch system, is the component of the target velocity in the launch frame.

[0043] Furthermore, the step 2 specifically includes:

[0044] Step 2.1: Perform discretization modeling of the entire airspace of the flight mission to obtain the feature point set of the entire airspace of the aircraft mission;

[0045] First, the launch altitude is discretized to obtain a launch altitude sequence. Then, the range of flight distance and lateral intercept at a typical launch altitude is determined and discretized to obtain a flight distance sequence and a lateral intercept sequence. Finally, the launch altitude sequence, flight distance sequence, and lateral intercept sequence are sequentially taken to form feature points and stored to obtain a feature point set for the entire airspace of the aircraft mission.

[0046] Step 2.2: Establish a mathematical description of the trajectory planning problem from the aircraft to the ground target point:

[0047] The aircraft state equation is the mathematical model of the aircraft. The aircraft's flight process satisfies the initial conditions, process constraints, and terminal constraints. Ultimately, the aircraft can hit the target. The trajectory planning ultimately obtains a set of control instructions that maximizes the aircraft's speed when it reaches the target point.

[0048] Step 2.3: Based on a feature point in the full airspace feature point set of the aircraft mission, the aircraft trajectory planning problem statement established in step 2.2 is solved using the HP adaptive pseudo-spectral method and the sequential quadratic programming method;

[0049] Step 2.4: Repeat step 2.3 for all feature points in the feature point set obtained in step 2.1 to obtain the optimal trajectory data set for the entire airspace and all feature points of the aircraft, which constitutes the optimal trajectory of the aircraft.

[0050] Further,

[0051] The aircraft state equation is:

[0052]

[0053] Where,

[0054] x=[x,y,z,V,θ,ψ v ], where x, y, and z are the coordinates of the aircraft position in the launch coordinate system, V is the velocity of the aircraft, θ, ψ v are the aircraft’s track inclination and track deviation respectively; u=[α,β], where α and β are the aircraft’s angle of attack and sideslip angle respectively; t0 is the initial time, t f End time

[0055] The initial conditions are:

[0056]

[0057] Where,

[0058] x0,y0,z0,V0,θ0,ψ v0 They are the three components of the aircraft's initial position, initial velocity, initial track inclination, and initial track deviation at the initial moment;

[0059] The process constraints are:

[0060]

[0061] The terminal constraints are:

[0062]

[0063] Where,

[0064] x f ,y f ,z f ,θ f They are the three components of the target point position and the terminal track inclination;

[0065] The trajectory planning is to maximize the terminal velocity of the optimization objective function, that is, the optimization objective is to make J = -V(t f ) takes the minimum value.

[0066] Furthermore, the step 3 specifically includes:

[0067] Step 3.1: Discretize and segment the optimal trajectory according to the discretization segmentation criterion;

[0068] Step 3.2: Fit the proportional guidance law on each segment;

[0069] Step 3.3: Use genetic algorithm to optimize the proportional guidance coefficient on each segment:

[0070] Step 3.4: Repeat steps 3.1, 3.2, and 3.3 for each optimal trajectory in the optimal trajectory data set obtained in step 2, and convert the optimal trajectory state x, y, z, V, θ, ψ v And the optimized proportional guidance coefficient K yi,j ,K zi,j Merge and construct the proportional guidance coefficient dataset.

[0071] Furthermore, in step 3.1, the discretization segmentation criterion is expressed as:

[0072]

[0073] Where,

[0074] s p is the distance between the current point and the starting point of the i-th segment; a i+1 is the starting point of the next segment; τ1 is a constant; b is the optimal trajectory i,j+1 Distance b i,j Flight distance, b is the optimal trajectory i,j+2 Distance b i,j+1 The flight distance is calculated by referring to the optimal trajectory b i,j-1 Distance b i,j Flight distance of a point

[0075]

[0076] In step 3.2, the fitting formula is as follows:

[0077]

[0078] Where,

[0079] The symbol subscript i represents the i-th segment, the symbol subscript j represents the j-th point, Δt i,j is the time interval between adjacent points in the fitted segment, t i,j is the moment of the jth trajectory point in the i-th segment, t i,j-1 is the moment of the j-1th trajectory point in the i-th segment, K yi,j ,K zi,j are the longitudinal proportional guidance coefficient and lateral proportional guidance coefficient to be optimized, are the estimated longitudinal overload and lateral overload, q αi,j ,q βi,jare the longitudinal target sight angular rate and the lateral target sight angular rate, V i,j is the flight speed, g0 is the acceleration due to gravity, are the estimated track inclination and track deviation, are the estimated positions of the aircraft in the launch coordinate system;

[0080] In step 3.3, the optimization process is as follows:

[0081] 1. Determine the parameter range and chromosome code length, chromosome code, establish the population size, determine the selection, crossover, and mutation probabilities, and establish the fitness function:

[0082]

[0083] Where,

[0084] m=(x,y,θ) is the state point on the optimal trajectory, The longitudinal proportional guidance coefficient is solved by using a discrete calculation method to obtain the state point estimate on the optimal trajectory;

[0085] 2. Initialize the population and randomly generate 200 individuals. Each individual carries chromosome information as the proportional guidance coefficient to obtain the first generation population.

[0086] 3. Use the fitness function to evaluate each individual in the population and obtain the fitness value of each individual;

[0087] 4. Perform selection, crossover, and mutation operations on the population;

[0088] 5. Perform 200 generations of iteration until the predetermined population iteration number is reached, stop the simulation and output the results.

[0089] Furthermore, the step 4 specifically includes:

[0090] Step 4.1: Preprocess the proportional guidance coefficient data set and map each data into the interval [-1, 1].

[0091] Step 4.2: Divide the entire data set into training and test sets in a ratio of 9:1, and perform neural network training and fitting to obtain the proportional guidance coefficient network;

[0092] Step 4.3: Design a neural network-based multi-constraint guidance method.

[0093] Furthermore, the mapping method in step 4.1 is:

[0094]

[0095] Where,

[0096] n=[x r ,y r ,z r ,V,ψ v ,θ] is the data vector input to the neural network, n min 、n max are the minimum and maximum vectors of the corresponding states in the proportional guidance coefficient data set; y is the output normalized data vector, y min =[-1,-1,-1,-1], y max =[1,1,1,1];

[0097] The multi-constraint guidance method in step 4.3 is:

[0098] The parameter time-varying proportional guidance law output by the neural network in the initial and intermediate guidance stages is:

[0099]

[0100] Where,

[0101] N y (t),N z (t) is the time-varying proportional guidance coefficient obtained by network training, n y1 ,n z1 Overload instructions for the initial and intermediate guidance segments;

[0102] At 1000m from the target, switch to the final guidance segment and use the proportional guidance law in the final guidance segment:

[0103]

[0104] Where,

[0105] N=3 is the proportional guidance constant, n y2 ,n z2 Overload instruction for the last leading segment.

[0106] Furthermore, the step 5 specifically includes:

[0107] Step 5.1: Define the remaining flight time t of the aircraft go The time t when the aircraft arrives at the target point end Subtract the flight time t of the aircraft and based on the optimal trajectory data set obtained in step 2, the trajectory state x=[x,y,z,V,θ,ψ v ] and the corresponding remaining flight time t goMerge and establish a vehicle state-remaining time dataset. The obtained vehicle state-remaining time dataset is divided into a training set and a test set. After data normalization, a neural network is used to fit the relationship between trajectory state and flight time to obtain a remaining time estimation network.

[0108] Step 5.2: Design a cooperative guidance law based on accurate estimation of the remaining time;

[0109] Step 5.3: Design relative remaining time error instructions for different flight modes.

[0110] Furthermore, the cooperative guidance law in step 5.2 is to put the remaining time error as feedback into the lateral guidance instruction, and its basic form is:

[0111] a M =a1+a2

[0112] Where,

[0113] Obtain the basic guidance law for the aircraft's flight target for step 4; is the time-constrained bias term containing the residual time error, where Ω(t) = k1rε(t) is a time-varying parameter, where r is the residual distance, k1>0 is a parameter, and ε(t) is the relative residual time error;

[0114] The relative remaining time error instructions designed in step 5.3 include:

[0115] The remaining time error command of the aircraft in fixed flight time arrival mode:

[0116] ε(t)=t d -tt go

[0117] Where,

[0118] t d is the set flight time, t is the aircraft flight time;

[0119] The remaining time error instruction of the i-th aircraft in the multi-aircraft sequential flight mode:

[0120] ε i (t) = min(t go,1 ,t go,2 ,…,t go,n )+Δt-t go,i

[0121] Where,

[0122] Δt is the time interval between the i-th aircraft and the first aircraft to reach the target;

[0123] The remaining time error command of the i-th aircraft in the multi-aircraft simultaneous flight mode:

[0124]

[0125] Where,

[0126] is the set of aircraft that can communicate with aircraft i, and the number of aircraft in this set is m.

[0127] The beneficial effects of the present invention are:

[0128] This paper proposes an intelligent trajectory planning and collaborative control method for aircraft based on offline optimization, intelligent learning, and online guidance. While ensuring real-time performance, it also implements a multi-cooperative flight mode under the premise of satisfying multiple constraints. During the execution of this method:

[0129] 1. Establish an optimal trajectory dataset for the entire airspace and all feature points of the aircraft. Using an offline trajectory optimization algorithm, we ensure that the trajectory satisfies multiple constraints and achieves optimal performance indicators.

[0130] 2. In step 3, the optimal trajectory dataset is fitted with proportional guidance and the proportional guidance coefficients are optimized to obtain a proportional guidance coefficient dataset. The proportional guidance coefficient network is then obtained through intelligent learning fitting in steps 4.2 and 4.3. This ensures that the guidance method is adaptable to the entire mission airspace, has low computational complexity, and has good real-time performance.

[0131] 3. Based on step 5.3, design the remaining time feedback instructions under different collaborative flight modes, and then design the collaborative bias items to achieve multiple flight modes such as sequential flight and simultaneous flight of aircraft. BRIEF DESCRIPTION OF THE DRAWINGS

[0132] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. It should be understood that the following drawings only illustrate certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other relevant drawings can be obtained based on these drawings without paying any creative work.

[0133] Figure 1 : Overall schematic diagram of the collaborative guidance and control scheme;

[0134] Figure 2 : Schematic diagram of the flight trajectory of Example 1;

[0135] Figure 3 : Track inclination curve diagram of Example 1;

[0136] Figure 4: Remaining flight time curve diagram of Example 1;

[0137] Figure 5 : Remaining distance curve diagram of Example 1;

[0138] Figure 6 : Schematic diagram of the multi-aircraft cooperative communication topology structure of Example 2;

[0139] Figure 7 : The flight trajectory diagram of Example 2;

[0140] Figure 8 : Track inclination curve diagram of Example 2;

[0141] Figure 9 : Remaining flight time curve diagram of Example 2;

[0142] Figure 10 : Remaining distance curve diagram of Example 2. DETAILED DESCRIPTION

[0143] The following describes embodiments of the present invention in detail, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, and are not to be construed as limiting the present invention.

[0144] This paper proposes an intelligent trajectory planning and collaborative control method for aircraft based on offline optimization-intelligent learning-online guidance, the overall schematic diagram of which is shown in the figure. Figure 1 As shown, the specific steps are:

[0145] Step 1: Establish a six-degree-of-freedom motion model of the aircraft in three-dimensional space and a relative motion model between the aircraft and the target;

[0146] Step 1.1: Establish a six-degree-of-freedom motion model of the aircraft;

[0147] 1. Dynamic equation of the center of mass motion of the aircraft:

[0148]

[0149] Where,

[0150] m is the mass of the aircraft, g is the acceleration of gravity, P is the engine thrust, V is the speed of the aircraft, X, Y, Z are the drag, lift and lateral force respectively, α, β are the angle of attack and sideslip angle respectively, θ, ψ v are the track inclination and track deviation, γ v is the velocity tilt angle;

[0151] 2. Linearized model of aerodynamic forces and aerodynamic moments acting on the aircraft:

[0152]

[0153] Where,

[0154] C x 、C y 、C z are the drag coefficient, lift coefficient and lateral force coefficient respectively, M x 、M y 、M z are rolling moment, yaw moment and pitching moment respectively, m x 、m y 、m z are the rolling moment coefficient, yaw moment coefficient and pitching moment coefficient respectively, q is the dynamic pressure, S is the characteristic area, and L is the characteristic length;

[0155] 3. Kinematic equations of the center of mass motion of the aircraft:

[0156]

[0157] Where,

[0158] x, y, and z are the coordinates of the aircraft's position in the launch coordinate system;

[0159] 4. Dynamic equations of the aircraft rotating around its center of mass:

[0160]

[0161] Where,

[0162] J x 、J y 、J z are the moment of inertia of the aircraft on the three axes of the body coordinate system, ω x 、ω y 、ω z are the components of the angular velocity ω of the aircraft relative to the ground coordinate system on the three axes of the body coordinate system, M x 、M y 、M z are the components of the total external moment acting on the aircraft in the three axes of the body coordinate system, namely, the rolling moment, the yaw moment and the pitching moment;

[0163] 5. Kinematic equations for the aircraft's rotation around its center of mass:

[0164]

[0165] Where,

[0166] ψ and γ are the pitch, yaw, and roll angles of the aircraft, respectively;

[0167] 6. Angle geometric relationship equation:

[0168]

[0169] Step 1.2: Establish the relative motion model between the aircraft and the target;

[0170] The component of the target position in the launch coordinate system is defined as (x T ,y T ,z T ), the component of the target velocity in the launch coordinate system is The relative distance r between the aircraft and the target and its derivative The calculation formula is:

[0171]

[0172] Where,

[0173] x r =x T -x,y r =y T -y,z r =z T -z,

[0174] Sight height angle q α , sight azimuth q β and line of sight height and angular velocity Line of sight azimuth The expressions are:

[0175]

[0176] Step 2: Establish an optimal trajectory dataset for the entire airspace and all feature points of the aircraft in three-dimensional space;

[0177] Step 2.1: Perform discretization modeling of the entire airspace of the flight mission to obtain the feature point set of the entire airspace of the aircraft mission;

[0178] First, the launch altitude is discretized to obtain a launch altitude sequence. Then, the range of flight distance and lateral intercept at a typical launch altitude is determined and discretized to obtain a flight distance sequence and a lateral intercept sequence. Finally, the launch altitude sequence, flight distance sequence, and lateral intercept sequence are sequentially taken to form feature points and stored to obtain a feature point set for the entire airspace of the aircraft mission.

[0179] Step 2.2: Establish a mathematical description of the trajectory planning problem from the aircraft to the ground target point;

[0180] The state equation of the aircraft is defined as:

[0181]

[0182] Where,

[0183] x=[x,y,z,V,θ,ψ v ], where x, y, and z are the coordinates of the aircraft position in the launch coordinate system, V is the velocity of the aircraft, θ, ψ v are the aircraft’s track inclination and track deviation respectively; u=[α,β], where α and β are the aircraft’s angle of attack and sideslip angle respectively; t0 is the initial time, t f is the end time;

[0184] The initial conditions of the aircraft are as follows:

[0185]

[0186] Where x0, y0, z0, V0, θ0, ψ v0 They are the three components of the aircraft's initial position, initial velocity, initial track inclination, and initial track deviation at the initial moment;

[0187] The flight process constraints of the aircraft are as follows:

[0188]

[0189] The aircraft flight terminal constraints are as follows:

[0190]

[0191] Where x f ,y f ,z f ,θ f They are the three components of the target point position and the terminal track inclination;

[0192] The optimization objective function is to maximize the terminal velocity, that is, the optimization goal is to minimize the following objective function:

[0193] J=-V(t f ) (13)

[0194] To summarize, the mathematical description of the trajectory planning problem from an aircraft to a ground target point is: the aircraft state equation is the aircraft mathematical model, the aircraft flight process satisfies the initial conditions, process constraints and terminal constraints, and the aircraft can eventually hit the target. The trajectory planning ultimately obtains a set of control instructions that maximizes the aircraft's speed when it reaches the target point.

[0195] Step 2.3: Based on a certain feature point in the full airspace feature point set of the aircraft mission, the aircraft trajectory planning problem description established in step 2.2 is solved using the HP adaptive pseudo-spectral method and the sequential quadratic programming method. Using the HP adaptive pseudo-spectral method and the sequential quadratic programming method for solution is a conventional method in this field and will not be repeated here.

[0196] Step 2.4: Repeat step 2.3 for all feature points in the feature point set obtained in step 2.1 to obtain the optimal trajectory data set for the entire airspace and all feature points of the aircraft, which constitutes the optimal trajectory of the aircraft.

[0197] Step 3: Parameterize the optimal trajectory of the aircraft obtained in step 2 to obtain a proportional guidance coefficient dataset;

[0198] Step 3.1: Discretize the optimal trajectory into segments, and the jth point on the optimal trajectory in the i-th segment is denoted as b i,j , the discretization segmentation criterion at the starting point of the i-th segment is expressed as follows:

[0199] Point b on the optimal trajectory i,j-1 Distance b i,j Flight distance of a point The calculation method is as follows:

[0200]

[0201] Calculate b in sequence i,j+1 Distance b i,j Flight distance b i,j+2 Distance b i,j+1 Flight distance Set a value s representing the distance between the current point and the starting point of the i-th segment p , if the starting point of the i-th segment is b i,j , then for the first point after the starting point, Then judge by the following formula:

[0202]

[0203] That is, when s p When ≤τ1, continue to take the next point as the current point until s p >τ1, where

[0204] a i+1 is the starting point of the next segment, τ1 is a constant, and the above criteria are repeatedly used to divide a trajectory into multiple segments.

[0205] Step 3.2: Use the proportional guidance law to fit each segment. The fitting formula is as follows:

[0206]

[0207] Where,

[0208] The symbol subscript i represents the i-th segment, the symbol subscript j represents the j-th point, Δt i,j is the time interval between adjacent points in the fitted segment, t i,j is the moment of the jth trajectory point in the i-th segment, t i,j-1 is the moment of the j-1th trajectory point in the i-th segment, K yi,j ,K zi,j are the longitudinal proportional guidance coefficient and lateral proportional guidance coefficient to be optimized, are the estimated longitudinal overload and lateral overload, q αi,j ,q βi,j are the longitudinal target sight angular rate and the lateral target sight angular rate, V i,j is the flight speed, g0 is the acceleration due to gravity, are the estimated track inclination and track deviation, are the estimated positions of the aircraft in the launch coordinate system;

[0209] Step 3.3: Use the genetic algorithm to optimize the proportional guidance coefficient on each segment. The optimization process is as follows:

[0210] 1. Determine the parameter value range and chromosome code length, chromosome code, establish the population size, and determine the selection, crossover, and mutation probabilities;

[0211] The fitness function is established as follows:

[0212]

[0213] Where,

[0214] m=(x,y,θ) is the state point on the optimal trajectory, The longitudinal proportional guidance coefficient is solved by using a discrete calculation method to obtain the state point estimate on the optimal trajectory;

[0215] 2. Initialize the population and randomly generate 200 individuals. Each individual carries chromosome information as the proportional guidance coefficient to obtain the first generation population.

[0216] 3. Use the fitness function to evaluate each individual in the population and obtain the fitness value of each individual;

[0217] 4. Perform selection, crossover, and mutation operations on the population;

[0218] 5. Perform 200 generations of iteration until the predetermined population iteration number is reached, stop the simulation and output the results.

[0219] Step 3.4: Repeat steps 3.1, 3.2, and 3.3 for each optimal trajectory in the optimal trajectory data set obtained in step 2, and convert the optimal trajectory state x, y, z, V, θ, ψ v And the optimized proportional guidance coefficient K yi,j ,K zi,j Merge and construct the proportional guidance coefficient dataset.

[0220] Step 4: Perform neural network learning and fitting on the proportional guidance coefficient dataset obtained in step 3, and design a multi-constraint guidance method based on neural network;

[0221] Step 4.1: Preprocess the proportional guidance coefficient data set and map each data into the range [-1, 1]. The specific mapping method is as follows:

[0222]

[0223] Where,

[0224] n=[x r ,y r ,z r ,V,ψ v ,θ] is the data vector input to the neural network, n min 、n max are the minimum and maximum vectors of the corresponding states in the proportional guidance coefficient data set; y is the output normalized data vector, y min =[-1,-1,-1,-1], y max =[1,1,1,1];

[0225] Step 4.2: Divide the entire data set into training and test sets in a ratio of 9:1, and perform neural network training and fitting to obtain the proportional guidance coefficient network;

[0226] Step 4.3: Design a neural network-based multi-constraint guidance method:

[0227] The parameter time-varying proportional guidance law output by the neural network in the initial and intermediate guidance stages is:

[0228]

[0229] Where,

[0230] N y (t),N z (t) is the time-varying proportional guidance coefficient obtained by network training, n y1 ,n z1 Overload instructions for the initial and intermediate guidance segments;

[0231] At 1000m from the target, switch to the final guidance segment and use the proportional guidance law in the final guidance segment:

[0232]

[0233] Where,

[0234] N=3 is the proportional guidance constant, n y2 ,n z2 Overload instruction for the last leading segment.

[0235] Step 5: Based on the neural network-based multi-constraint guidance method obtained in step 4, further design a three-dimensional collaborative guidance law for the aircraft that meets multiple constraints and multiple flight modes to achieve three-dimensional collaborative control of the aircraft;

[0236] Step 5.1: Define the remaining flight time t of the aircraft go The time t when the aircraft arrives at the target point end Subtract the flight time t of the aircraft and based on the optimal trajectory data set obtained in step 2, the trajectory state x=[x,y,z,V,θ,ψ v ] and the corresponding remaining flight time t go Merge and establish a vehicle state-remaining time dataset. The obtained vehicle state-remaining time dataset is divided into a training set and a test set. After data normalization, a neural network is used to fit the relationship between trajectory state and flight time to obtain a remaining time estimation network.

[0237] Step 5.2: Design a cooperative guidance law based on accurate estimation of the remaining time;

[0238] The design idea is to put the remaining time error as feedback into the lateral guidance instruction, and its basic form is as follows:

[0239] a M =a1+a2 (21)

[0240] Where,

[0241] a1 is the basic guidance law for the aircraft's flight target obtained in step 4, and is in the following form:

[0242]

[0243] a2 is the time constraint bias term containing the residual time error, which is designed as follows:

[0244]

[0245] Where,

[0246] Ω(t) is a time-varying parameter, and its expression is:

[0247] Ω(t)=k1rε(t) (24)

[0248] Where,

[0249] r is the remaining distance, k1>0 is a parameter, and ε(t) is the relative remaining time error;

[0250] Step 5.3: Design relative remaining time error instructions for different flight modes. Typical flight modes are as follows:

[0251] The remaining time error command of the aircraft in the fixed flight time arrival mode is:

[0252] ε(t)=t d -tt go (25)

[0253] Where,

[0254] t d is the set flight time, t is the aircraft flight time;

[0255] The remaining time error instruction ε of the i-th aircraft in the multi-aircraft sequential flight mode i (t) is:

[0256] ε i (t) = min(t go,1 ,t go,2 ,…,t go,n )+Δt-t go,i (26)

[0257] Where,

[0258] Δt is the time interval between the i-th aircraft and the first aircraft to reach the target;

[0259] The remaining time error instruction ε of the i-th aircraft in the multi-aircraft simultaneous flight mode i (t) is:

[0260]

[0261] Where,

[0262] is the set of aircraft that can communicate with aircraft i, and the number of aircraft in this set is m.

[0263] Example 1 and Example 2 respectively conduct simulation verification on the multi-aircraft sequential flight mode and the multi-aircraft simultaneous flight mode:

[0264] Example 1:

[0265] The implementation of the method is illustrated by taking three aircraft in a multi-aircraft sequential flight mode to coordinately reach a fixed ground target. The initial position of the given target is (0,0,0), and the initial conditions of the aircraft are shown in Table 1:

[0266] Table 1 Simulation scenario of Example 1

[0267]

[0268] Figure 2 The flight trajectory of three aircraft flying cooperatively to a stationary target is given. It can be seen that the trajectory of the entire flight process is relatively smooth and can reach the target point;

[0269] Figure 3 The track inclination variation curves of the three aircraft during flight are given. At the time of impact, the track inclinations of the three aircraft were -76.59°, -77.42°, and -79.47°, respectively, which met the flight angle constraints.

[0270] Figure 4 The remaining flight time variation curves of the three aircraft during flight are given. The arrival time intervals of the three aircraft are 0.809s and 0.778s, and the sequential flight mode with an arrival interval of 0.8s is successfully achieved.

[0271] Figure 5 The changing curves of the relative distances between the three aircraft and the target are given. It can be seen that the relative distances between the three aircraft and the target all converge to zero. The final relative distances are 0.04m, 0.17m, and 0.12m respectively, and the guidance accuracy is high.

[0272] Example 2:

[0273] The implementation of the method is illustrated by taking three aircraft in multi-aircraft simultaneous flight mode to coordinately reach a fixed ground target. The initial position of the given target is (0,0,0), and the initial conditions of the aircraft are shown in Table 2:

[0274] Table 2 Simulation scenarios of Example 2

[0275]

[0276] Figure 6 The cooperative communication topology structure in the scenario where three aircraft fly simultaneously towards a target is given;

[0277] Figure 7 The flight trajectory of three aircraft flying cooperatively to a stationary target is given. It can be seen that the trajectory of the entire flight process is relatively smooth and can reach the target point;

[0278] Figure 8The track inclination variation curves of the three aircraft during flight are given. At the time of impact, the track inclinations of the three aircraft were -76.62°, -77.67°, and -79.96°, respectively, which met the flight angle constraints.

[0279] Figure 9 The remaining flight time variation curves of the three aircraft during flight are given. The arrival times of the three aircraft are 55.265s, 55.287s, and 55.272s respectively, with the maximum arrival time being 0.022s, successfully achieving the simultaneous flight mode.

[0280] Figure 10 The changing curves of the relative distances between the three aircraft and the target are given. It can be seen that the relative distances between the three aircraft and the target all converge to zero. The final relative distances are 0.04m, 0.21m, and 0.14m respectively, and the guidance accuracy is high.

[0281] In summary, this method has a relatively smooth trajectory for the coordinated arrival process of multiple flight modes such as sequential flight mode and simultaneous flight mode, can achieve accurate arrival at the target, and can ensure that the remaining flight time of multiple aircraft meets the requirements of the corresponding flight mode.

[0282] The above description is only a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with this technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed in the present invention, and these modifications or replacements should all be included in the scope of protection of the present invention.

Claims

1. A method for intelligent trajectory planning and coordinated control of an aircraft, characterized in that: The specific steps include: Step 1: Establish a six-degree-of-freedom motion model of the aircraft in three-dimensional space and a relative motion model between the aircraft and the target; Step 2: Establish an optimal trajectory dataset for the entire airspace and all feature points of the aircraft in three-dimensional space; Step 3: Parameterize the optimal trajectory of the aircraft obtained in step 2 to obtain a proportional guidance coefficient dataset; Step 4: Perform neural network learning and fitting on the proportional guidance coefficient dataset obtained in step 3, and design a multi-constraint guidance method based on neural network; Step 5: Based on the neural network-based multi-constraint guidance method obtained in step 4, further design a three-dimensional collaborative guidance law for the aircraft that meets multiple constraints and multiple flight modes to achieve three-dimensional collaborative control of the aircraft.

2. The method for intelligent trajectory planning and coordinated control of an aircraft according to claim 1, wherein: In step 1, the six-degree-of-freedom motion model of the aircraft includes: The dynamic equation of the center of mass motion of the aircraft is: Where, m is the mass of the aircraft, g is the acceleration of gravity, P is the engine thrust, V is the speed of the aircraft, X, Y, Z are the drag, lift and lateral force respectively, α, β are the angle of attack and sideslip angle respectively, θ, ψ v are the track inclination and track deviation, γ v is the velocity tilt angle; Linearized model of aerodynamic forces and aerodynamic moments acting on the aircraft: Where, C x 、C y 、C z are the drag coefficient, lift coefficient and lateral force coefficient respectively, M x 、M y 、M z are rolling moment, yaw moment and pitching moment, m x 、m y 、m z are the rolling moment coefficient, yaw moment coefficient and pitching moment coefficient respectively, q is the dynamic pressure, S is the characteristic area, and L is the characteristic length; Kinematic equations for the motion of the center of mass of an aircraft: Where, x, y, and z are the coordinates of the aircraft's position in the launch coordinate system; The dynamic equation of the aircraft rotating around the center of mass: Where, J x 、J y 、J z are the moment of inertia of the aircraft on the three axes of the body coordinate system, ω x 、ω y 、ω z are the components of the angular velocity ω of the aircraft relative to the ground coordinate system on the three axes of the body coordinate system, M x 、M y 、M z are the components of the total external moment acting on the aircraft in the three axes of the body coordinate system, namely, the rolling moment, the yaw moment and the pitching moment; The kinematic equations of the aircraft rotating around its center of mass are: Where, ψ and γ are the pitch, yaw, and roll angles of the aircraft, respectively; Angle geometric relationship equation: The relative motion model between the aircraft and the target includes: The relative distance r between the aircraft and the target and its derivative Sight height angle q α , sight azimuth q β and line of sight height and angular velocity Line of sight azimuth Where, is the component of the target position in the launch system, is the component of the target velocity in the launch frame.

3. The method for intelligent trajectory planning and coordinated control of an aircraft according to claim 1, wherein: The step 2 specifically includes: Step 2.1: Perform discretization modeling of the entire airspace of the flight mission to obtain the feature point set of the entire airspace of the aircraft mission; First, the launch altitude is discretized to obtain a launch altitude sequence. Then, the range of flight distance and lateral intercept at a typical launch altitude is determined and discretized to obtain a flight distance sequence and a lateral intercept sequence. Finally, the launch altitude sequence, flight distance sequence, and lateral intercept sequence are sequentially taken to form feature points and stored to obtain a feature point set for the entire airspace of the aircraft mission. Step 2.2: Establish a mathematical description of the trajectory planning problem from the aircraft to the ground target point: The aircraft state equation is the mathematical model of the aircraft. The aircraft's flight process satisfies the initial conditions, process constraints, and terminal constraints. Ultimately, the aircraft can hit the target. The trajectory planning ultimately obtains a set of control instructions that maximizes the aircraft's speed when it reaches the target point. Step 2.3: Based on a feature point in the full airspace feature point set of the aircraft mission, the aircraft trajectory planning problem statement established in step 2.2 is solved using the HP adaptive pseudo-spectral method and the sequential quadratic programming method; Step 2.4: Repeat step 2.3 for all feature points in the feature point set obtained in step 2.1 to obtain the optimal trajectory data set for the entire airspace and all feature points of the aircraft, which constitutes the optimal trajectory of the aircraft.

4. The method for intelligent trajectory planning and coordinated control of an aircraft according to claim 3, wherein: The aircraft state equation is: Where, x=[x,y,z,V,θ,ψ v ], where x, y, and z are the coordinates of the aircraft position in the launch coordinate system, V is the velocity of the aircraft, θ, ψ v are the aircraft’s track inclination and track deviation respectively; u=[α,β], where α and β are the aircraft’s angle of attack and sideslip angle respectively; t0 is the initial time, t f End time The initial conditions are: Where, x0,y0,z0,V0,θ0,ψ v0 They are the three components of the aircraft's initial position, initial velocity, initial track inclination, and initial track deviation at the initial moment; The process constraints are: The terminal constraints are: Where, x f ,y f ,z f ,θ f They are the three components of the target point position and the terminal track inclination; The trajectory planning is to maximize the terminal velocity of the optimization objective function, that is, the optimization objective is to make J = -V(t f ) takes the minimum value.

5. The method for intelligent trajectory planning and coordinated control of an aircraft according to claim 1, wherein: The step 3 specifically includes: Step 3.1: Discretize and segment the optimal trajectory according to the discretization segmentation criterion; Step 3.2: Fit the proportional guidance law on each segment; Step 3.3: Use genetic algorithm to optimize the proportional guidance coefficient on each segment: Step 3.4: Repeat steps 3.1, 3.2, and 3.3 for each optimal trajectory in the optimal trajectory data set obtained in step 2, and convert the optimal trajectory state x, y, z, V, θ, ψ v And the optimized proportional guidance coefficient K yi,j ,K zi,j Merge and construct the proportional guidance coefficient dataset.

6. The method for intelligent trajectory planning and coordinated control of an aircraft according to claim 5, wherein: In step 3.1, the discretization segmentation criterion is expressed as: Where, s p is the distance between the current point and the starting point of the i-th segment; a i+1 is the starting point of the next segment; τ1 is a constant; b is the optimal trajectory i,j+1 Distance b i,j Flight distance, b is the optimal trajectory i,j+2 Distance b i,j+1 The flight distance is calculated by referring to the optimal trajectory b i,j-1 Distance b i,j Flight distance of the point In step 3.2, the fitting formula is as follows: Where, The symbol subscript i represents the i-th segment, the symbol subscript j represents the j-th point, Δt i,j is the time interval between adjacent points in the fitted segment, t i,j is the moment of the jth trajectory point in the i-th segment, t i,j-1 is the moment of the j-1th trajectory point in the i-th segment, K yi,j ,K zi,j are the longitudinal proportional guidance coefficient and lateral proportional guidance coefficient to be optimized, are the estimated longitudinal overload and lateral overload, q αi,j ,q βi,j are the longitudinal target sight angular rate and the lateral target sight angular rate, V i,j is the flight speed, g0 is the acceleration due to gravity, are the estimated track inclination and track deviation, are the estimated positions of the aircraft in the launch coordinate system; In step 3.3, the optimization process is as follows:

1. Determine the parameter range and chromosome code length, chromosome code, establish the population size, determine the selection, crossover, and mutation probabilities, and establish the fitness function: Where, m=(x,y,θ) is the state point on the optimal trajectory, The longitudinal proportional guidance coefficient is solved by using a discrete calculation method to obtain the state point estimate on the optimal trajectory; 2. Initialize the population and randomly generate 200 individuals. Each individual carries chromosome information as the proportional guidance coefficient to obtain the first generation population.

3. Use the fitness function to evaluate each individual in the population and obtain the fitness value of each individual; 4. Perform selection, crossover, and mutation operations on the population; 5. Perform 200 generations of iteration until the predetermined population iteration number is reached, stop the simulation and output the results.

7. The method for intelligent trajectory planning and coordinated control of an aircraft according to claim 1, wherein: The step 4 specifically includes: Step 4.1: Preprocess the proportional guidance coefficient data set and map each data into the interval [-1, 1]. Step 4.2: Divide the entire data set into training and test sets in a ratio of 9:1, and perform neural network training and fitting to obtain the proportional guidance coefficient network; Step 4.3: Design a neural network-based multi-constraint guidance method.

8. The method for intelligent trajectory planning and coordinated control of an aircraft according to claim 7, wherein: The mapping method in step 4.1 is: Where, n=[x r ,y r ,z r ,V,ψ v ,θ] is the data vector input to the neural network, n min 、n max are the minimum and maximum vectors of the corresponding states in the proportional guidance coefficient data set; y is the output normalized data vector, y min =[-1,-1,-1,-1], y max =[1,1,1,1]; The multi-constraint guidance method in step 4.3 is: The parameter time-varying proportional guidance law output by the neural network in the initial and intermediate guidance stages is: Where, N y (t),N z (t) is the time-varying proportional guidance coefficient obtained by network training, n y1 ,n z1 Overload instructions for the initial and intermediate guidance segments; At 1000m from the target, switch to the final guidance segment and use the proportional guidance law in the final guidance segment: Where, N=3 is the proportional guidance constant, n y2 ,n z2 Overload instruction for the last leading segment.

9. The method for intelligent trajectory planning and coordinated control of an aircraft according to claim 1, wherein: The step 5 specifically includes: Step 5.1: Define the remaining flight time t of the aircraft go The time t when the aircraft arrives at the target point end Subtract the flight time t of the aircraft and based on the optimal trajectory data set obtained in step 2, the trajectory state x=[x,y,z,V,θ,ψ v ] and the corresponding remaining flight time t go Merge and establish a vehicle state-remaining time dataset. The obtained vehicle state-remaining time dataset is divided into a training set and a test set. After data normalization, a neural network is used to fit the relationship between trajectory state and flight time to obtain a remaining time estimation network. Step 5.2: Design a cooperative guidance law based on accurate estimation of the remaining time; Step 5.3: Design relative remaining time error instructions for different flight modes.

10. The method for intelligent trajectory planning and coordinated control of an aircraft according to claim 9, wherein: The cooperative guidance law in step 5.2 is to put the remaining time error as feedback into the lateral guidance instruction, and its basic form is: <h2 style=";text-align:left;direction:ltr">a<h2 style=";text-align:left;direction:ltr"> M <h2 style=";text-align:left;direction:ltr"> =a1+a2 Where, Obtain the basic guidance law for the aircraft's flight target for step 4; is the time-constrained bias term containing the residual time error, where Ω(t) = k1rε(t) is a time-varying parameter, where r is the residual distance, k1>0 is a parameter, and ε(t) is the relative residual time error; The relative remaining time error instructions designed in step 5.3 include: The remaining time error command of the aircraft in fixed flight time arrival mode: ε(t)=t d -t-t go Where, t d is the set flight time, t is the aircraft flight time; The remaining time error instruction of the i-th aircraft in the multi-aircraft sequential flight mode: ε i (t)=min(t go,1 ,t go,2 ,…,t go,n )+Δt-t go,i Where, Δt is the time interval between the i-th aircraft and the first aircraft to reach the target; The remaining time error command of the i-th aircraft in the multi-aircraft simultaneous flight mode: Where, is the set of aircraft that can communicate with aircraft i, and the number of aircraft in this set is m.

Images