CN-121977568-A - PINN-based reentry glide terminal speed near-optimal track planning method and system

CN121977568ACN 121977568 ACN121977568 ACN 121977568ACN-121977568-A

Abstract

The invention discloses a PINN-based reentry and glide terminal speed near-optimal trajectory planning method and system, and aims to solve the problems that the reentry and glide trajectory planning of a defending aircraft is subjected to complex constraint (process, control, space-time full state, terminal speed near-optimal) and low solving efficiency under the strong opposition of the remote distance of a nearby space. The method comprises the following steps of final speed near-optimal longitudinal planning (parameterized terminal speed, multi-parameter resistance acceleration profile, voyage and time analysis prediction, SQP optimization), PINN-based double-stage self-adaptive lateral planning (course adjustment/maintenance, PINN rapid terminal state prediction), three-degree-of-freedom track correction and generation (lateral large maneuvering range and time correction and maneuvering coefficient self-adaptive determination based on Gaussian process model), and corresponding modules of task scene construction, problem modeling and the like of the system. The invention satisfies the near-optimal terminal speed and space-time full-state constraint, has high terminal precision and less calculation time consumption, and is suitable for remote forward strong countermeasure tasks.

Inventors

Yan Xunliang
XIA WENJIE
YUE XINYANG
LI JING
LIU JIAQI
GONG MAOHUA
ZHANG HAOLONG
LIU RUIFENG

Assignees

西北工业大学
中国航天系统科学与工程研究院

Dates

Publication Date: 20260505
Application Date: 20260203

Claims (10)

1. A PINN-based reentry glide terminal speed near-optimal track planning method is characterized by comprising the following steps: Step 1, constructing a reentry gliding flight task scene facing the remote forward-going countermeasure requirement, wherein a gliding type defending aircraft in the task scene adopts a high-throwing-reentry-gliding type trajectory scheme, realizes remote delivery and approaching through a gliding carrying remote delivery and releasing load approaching mode, and needs to reach a shift position according to preset flight time and meet the terminal time, height, speed, track angle and course angle constraint; Step 2, performing problem description, including establishing a glide type defending aircraft dimensionless three-degree-of-freedom mass center motion equation which ignores earth rotation and takes dimensionless energy as independent variable, constructing a constraint model containing process constraint, control constraint and middle-to-last shift constraint, and setting performance indexes of maximizing terminal speed; Step 3, designing a glide track planning algorithm, which comprises the following steps: 3.1, carrying out parameterization characterization on the terminal speed, designing a multi-parameter resistance acceleration profile based on resistance acceleration reentry corridor boundary interpolation, and optimizing profile parameters through a sequential quadratic programming algorithm by combining a course and time analysis prediction method to realize longitudinal track planning meeting constraints; Dividing the lateral gauge into a course adjustment section and a course maintenance section, rapidly predicting the state of the terminal by utilizing PINN network, solving a tilting inversion point by Newton iteration method, designing tilting inversion logic to determine a tilting angle symbol, and realizing lateral planning meeting the constraint of the position and the course angle of the terminal; 3.3 three-degree-of-freedom glide track correction and rapid generation, namely designing a course and time correction strategy suitable for lateral large maneuver conditions, combining a maneuver coefficient self-adaptive determination method based on a Gaussian process model, correcting the track and rapidly generating the three-degree-of-freedom glide track meeting all constraints.
2. The method for planning a near-optimal trajectory at a reentry and glide terminal speed based on PINN as claimed in claim 1, wherein the glide type defending aircraft in step 1 has a large lift-drag ratio aerodynamic profile, and can fly in a near space in a long distance in a glide mode, and among terminal constraints to be met by the shift position, the terminal position constraint is converted into a terminal to-be-flown constraint so as to be convenient for adjusting a resistance acceleration profile.
3. The method for near-optimal trajectory planning at the end of reentry and glide based on PINN of claim 1, wherein in step 2: the process constraint comprises heat flow, dynamic pressure, overload and balanced gliding conditions, and is converted into a functional relation of resistance acceleration and dimensionless energy to obtain D-E reentry corridor boundaries, and then converted into resistance acceleration corridor constraint Wherein In order to normalize the energy of the light, 、 The lower and upper boundaries of the corridor are respectively; In the control constraint, the attack angle profile is preset as a three-section function, and only the roll angle amplitude is constrained , 、 The lower limit and the upper limit of the tilting angle amplitude are respectively set; The mid-to-last shift constraint comprises a terminal position constraint, a terminal speed direction angle constraint and a flight time constraint, namely , 、、、 Respectively presetting terminal time, altitude, longitude and latitude, 、 Respectively the terminal speed direction angle constraint, 、 The error margin of the course angle and the course angle respectively.
4. The method for planning a near-optimal trajectory at a reentry and glide end speed based on PINN as claimed in claim 1, wherein the terminal speed parameterization in step 3.1 is specifically: wherein To be optimized for parameters and , In order to have a dimensionless terminal speed, 、 The terminal speed is the minimum value and the maximum value of the dimensionless terminal speed respectively; According to the dynamic pressure constraint analysis determination of the upper boundary of the height-speed corridor, the expression is , Is used for the purpose of the maximum dynamic pressure, As the average radius of the earth, Is sea level atmospheric density; And (3) carrying out numerical iteration solution on the nonlinear relation between the height and the speed of the terminal established by the quasi-equilibrium gliding condition or estimating by using a resistance acceleration expression.
5. The method for planning a near-optimal trajectory at a reentry and glide end speed based on PINN as claimed in claim 1, wherein the multi-parameter drag acceleration profile in step 3.1 is a 5-segment polynomial, and the specific expression is: , Wherein, the In order to normalize the energy of the light, 、 To fit a boundary for the reentry corridor, In order to refer to the resistive acceleration profile, 、 To be optimized for the resistive acceleration profile, furthermore, to ensure that the designed reference profile is always inside the corridor to avoid violations of process constraints, 、 The conditions should be satisfied: 。
6. The method for near optimal trajectory planning at the end of reentry and glide based on PINN as claimed in claim 1, wherein the voyage and time resolution prediction in step 3.1 is specifically to ignore earth rotation and assume that 、 Course of voyage Time of The derivative of normalized energy is , Order the Integrating the derivative to obtain the course and time corresponding to the course resistance acceleration profile, which can be expressed as , In the formula, The integral process obtains an analytic prediction formula by deducing a primitive function, and the prediction error is corrected and compensated by the subsequent degree of freedom track.
7. The method for near-optimal trajectory planning at the end of reentry and glide based on PINN of claim 1, wherein in step 3.2: The course adjustment section adjusts the relative position of the aircraft and the target point by adjusting the corresponding normalized energy of the single tilting reverse point to enable the line of sight angle to be close to the expected value of the course angle of the terminal; The PINN network comprises an input layer, a hidden layer and an output layer, wherein the input is dimensionless energy Current state quantity Roll angle of control amount Output is the state quantity of the next moment ; The PINN network has a loss function of , As the weight coefficient of the light-emitting diode, The term loss is fitted to the data, Is a physical guide loss, and can be expressed as , Wherein, the Is PINN network in State quantity predicted value at the time; Is that For a pair of Is a derivative of (2); And (3) with Respectively is The true value of the state quantity and its derivative, To train the number of sample points.
8. The method for near-optimal trajectory planning at the end of reentry and glide based on PINN of claim 1, wherein the roll-over inversion logic in step 3.2 is specifically: Setting the heading angle deviation of the terminal , Normalizing energy for a tilting inversion point by Newton's iteration Solving for The iteration cut-off condition is , Is a convergence error limit; the energy of the tilting inversion point is required to be satisfied , wherein, The maximum value of the tilting inversion energy can be determined according to the lateral movement capability before and after inversion, and can be taken as ; When (when) When the gliding defending aircraft is in the course adjustment section, the roll angle sign of the gliding defending aircraft can be determined by the following formula , In the formula, To reenter the initial roll angle symbol; When (when) When the glide type defending aircraft enters the course maintaining section, the roll angle sign is determined by the course angle error corridor, and the corresponding roll inversion logic is that , In the formula, A roll angle symbol for the previous time; The corridor boundary can be designed as a piecewise linear function of speed, and furthermore, the terminal course angle precision and the end section width of the course angle error corridor are considered Closely related, in order to ensure that the heading angle precision of the terminal strictly meets the requirement, the requirements are met Strictly less than 。
9. The method for near-optimal trajectory planning at the end of reentry and glide based on PINN of claim 1, wherein in step 3.3: the course correction strategy is based on a spherical triangle formula, and the reentry point and the target azimuth angle are set as The arc length of the large arc track is Predicting the track terminal and the great arc length of the reentrant point to be The distance between the terminal and the target is By means of The projected arc length is calculated and, , Azimuth angles of the track terminal and the reentry point are obtained; symbol and terminal point of (c) and target course angle Related, in particular, can represent , Maneuver coefficient self-adaptive determination method based on Gaussian process GP model , For the amount of change in the initial and final heading angle, In order for the time of flight to be a function of, A nonlinear mapping relation represented by a GP model; By means of Correcting target course To reduce the number of track outer loop iterations.
10. A PINN-based reentry glide terminal near-optimal trajectory planning system, comprising: the task scene construction module is used for constructing the reentry gliding flight task scene in the step 1 of the claim 1; The problem modeling module is used for executing the centroid motion equation establishment, the constraint model establishment and the performance index setting described in the step 2; the longitudinal planning module is used for executing the final-speed near-optimal longitudinal planning in the step 3.1; The lateral planning module is used for executing PINN-based double-stage self-adaptive lateral planning in the step 3.2; The track correction module is used for executing the course correction described in the step 3.3 and adaptively determining the maneuvering coefficient based on the GP; The track generation module is used for integrating the longitudinal track, the lateral track and the correction result and rapidly generating a three-degree-of-freedom reentry gliding end speed near-optimal track meeting all constraints.

Description

PINN-based reentry glide terminal speed near-optimal track planning method and system Technical Field The invention relates to the technical field of aerospace, in particular to a method and a system for planning a near-optimal track of a reentry gliding end speed based on PINN. Background The near space high-speed target aircraft has the characteristics of high flying speed, long voyage, strong maneuver burst prevention capability and the like, and the rapid development of the near space high-speed target aircraft brings great threat to the aerospace safety of defenses. However, the conventional countermeasure trajectory scheme for the incoming target mostly adopts a ramp-up type arc trajectory, and the scheme has the defects of limited range slant distance, relatively low speed, weak maneuverability and the like, so that the defending aircraft is difficult to occupy space-time advantages in the countermeasure process, and therefore, development of a novel trajectory pattern and trajectory planning technology for actively and remotely defending the approaching space high-speed reentry target is needed. The high-throwing-reentry-glide trajectory scheme can remarkably improve the effective range and the maneuverability of the defensive aircraft, so that the defensive aircraft can be suitable for remote forward-forward strong countermeasure tasks of reentry of glide targets in the near space. In this type of ballistic scheme, the flying time of the defending aircraft in the reentry glide phase can generally reach more than 65%, so that the overall flying performance of the aircraft is significantly affected. However, the high-precision remote forward strong countermeasure task in the near space makes the trajectories of the glide segments of the defending aircraft face severe terminal time and state constraints, the particularity of the countermeasure task also requires the terminal speed to be as high as possible, and the characteristics result in the problem of planning the glide trajectories of the defending aircraft being evolved into the problems of fast time-varying, strong coupling and nonlinear planning with complex severe constraints and specific optimization indexes, and the feasibility domain is smaller and the solving difficulty is larger. In order to solve the problems, a glide track planning method research taking the optimal conditions of terminal time, state constraint and final speed into consideration needs to be carried out. At present, the conventional rapid reentry glide track planning method based on standard track tracking, prediction correction thought or balance glide hypothesis does not consider complex constraints such as time and terminal angle, and the conventional track planning method meeting the time or angle constraints does not consider terminal speed optimality in most cases, but the track optimization numerical value solving method capable of realizing the terminal speed optimization is generally complex and takes long time, and even cannot be effectively solved, so that the method is difficult to be applied to reentry glide terminal speed near-optimal track planning with complex and multiple constraints. Disclosure of Invention Aiming at the problems, the invention provides a PINN-based reentry gliding near-optimal trajectory planning method and system, which meet near-optimal and multi-constraint requirements of the terminal, improve the terminal precision and the calculation efficiency, reduce the iteration time consumption and are suitable for remote forward strong countermeasure tasks. In order to solve the problems, the invention adopts the following technical scheme: a PINN-based reentry glide terminal speed near-optimal trajectory planning method comprises the following steps: Step 1, constructing a reentry gliding flight task scene facing the remote forward-going countermeasure requirement, wherein a gliding type defending aircraft in the task scene adopts a high-throwing-reentry-gliding type trajectory scheme, realizes remote delivery and approaching through a gliding carrying remote delivery and releasing load approaching mode, and needs to reach a shift position according to preset flight time and meet the terminal time, height, speed, track angle and course angle constraint; Step 2, performing problem description, including establishing a glide type defending aircraft dimensionless three-degree-of-freedom mass center motion equation which ignores earth rotation and takes dimensionless energy as independent variable, constructing a constraint model containing process constraint, control constraint and middle-to-last shift constraint, and setting performance indexes of maximizing terminal speed; Step 3, designing a glide track planning algorithm, which comprises the following steps: 3.1, carrying out parameterization characterization on the terminal speed, designing a multi-parameter resistance acceleration profile based on resistance acceleration reentry corrido