CN114417730B - Aircraft glide section flight range online prediction method based on integrated neural network - Google Patents
Aircraft glide section flight range online prediction method based on integrated neural network Download PDFInfo
- Publication number
- CN114417730B CN114417730B CN202210102721.8A CN202210102721A CN114417730B CN 114417730 B CN114417730 B CN 114417730B CN 202210102721 A CN202210102721 A CN 202210102721A CN 114417730 B CN114417730 B CN 114417730B
- Authority
- CN
- China
- Prior art keywords
- aircraft
- flight
- neural network
- longitude
- latitude
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
Abstract
An integrated neural network-based aircraft glide section flight range online prediction method belongs to the technical field of aircraft flight. The method comprises the following steps: before the aircraft takes off, selecting sample points on a preset standard track and dividing the sample points into a plurality of groups randomly; calculating the longitude and latitude of the trajectory drop point; forming a sample library which inputs the initial value of trajectory calculation and outputs the maximum value and the minimum value of the trajectory drop point longitude and latitude; respectively training each group of sample points to obtain a flight range prediction neural network; inputting the current flight state of the aircraft, adopting a flight range prediction neural network to predict the flight range, performing integrated neural network prediction on the prediction result, and then obtaining the flight prediction result. The invention improves the accuracy of prediction, considers the strong nonlinear dynamics process of the glide section of the aircraft, can obtain the maximum flight range according to the current aircraft state, adopts the mode of offline training and online use of a neural network, calculates the flight range in real time, and has good engineering practicability.
Description
Technical Field
The invention relates to an integrated neural network-based online predication method for the flight range of a glide section of an aircraft, and belongs to the technical field of aircraft flight.
Background
The dynamics process of the gliding section of the aircraft has strong nonlinear characteristics. The traditional flight range online predication method based on the approximate linearization model has low calculation precision, and the flight range predication method based on the nonlinear model has long calculation time and cannot meet the requirement of online calculation. Therefore, it is urgently needed to develop a new online prediction method of the flight range.
Disclosure of Invention
In order to solve the problems in the background technology, the invention provides an integrated neural network-based online prediction method for the flight range of the glide section of an aircraft.
The invention adopts the following technical scheme: an integrated neural network-based online predication method for flight range of a glide flight of an aircraft comprises the following steps:
s1: before the aircraft takes off, selecting a plurality of sample points of flight state quantity on a preset standard track, and dividing the plurality of sample points into a plurality of groups at random;
s2: calculating the longitude and latitude of the trajectory drop point;
s3: for each group of sample points, respectively calculating the corresponding maximum value of the trajectory drop point longitude, the minimum value of the trajectory drop point longitude, the maximum value of the trajectory drop point latitude and the minimum value of the trajectory drop point latitude, and taking the two maximum values and the corresponding minimum value as a range to form a sample library which is input as an initial trajectory calculation value and output as the maximum value and the minimum value of the trajectory drop point longitude and latitude;
s4: respectively training each group of sample points to obtain a flight range prediction neural network;
s5: when the aircraft flies online, inputting the current flying state of the aircraft, adopting each training flying range prediction neural network obtained in S4 to predict the flying range, performing integrated neural network prediction on the prediction result, and further obtaining the flying prediction result.
Compared with the prior art, the invention has the beneficial effects that:
according to the invention, the predicted flight range results of a plurality of neural networks are integrated, so that the accuracy of prediction is improved; the low consuming time, the aircraft of high accuracy section of flying flight based on integrated neural network presupposes on line, has considered the strong nonlinear dynamics process of aircraft section of flying, can obtain the biggest flight scope according to current aircraft state, adopts the mode that neural network line down trains and uses on line, calculates the flight scope in real time, has good engineering practicality.
Drawings
FIG. 1 is a schematic flow diagram of the present invention;
FIG. 2 is a flow chart of integrated neural network prediction.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments, and based on the embodiments of the present invention, all other embodiments obtained by a person skilled in the art without making creative efforts belong to the protection scope of the present invention.
An integrated neural network-based online predication method for flight range of a glide flight of an aircraft comprises the following steps:
s1: before the aircraft takes off, selecting a plurality of flight state quantity sample points on a preset standard track, wherein the flight state quantity sample points comprise: ground center distance r before flight of aircraft 0 Longitude theta 0 Latitude phi 0 Velocity V 0 Track angle gamma 0 And heading angle psi 0 Dividing the plurality of sample points into a plurality of groups at random;
s2: taking each group of sample points as initial ballistic calculation values, randomly obtaining and determining a global attack angle alpha and a roll angle sigma of the aircraft according to the given section forms of the attack angle and the roll angle of the aircraft, setting the global attack angle alpha and the roll angle sigma of the aircraft to different values respectively, and calculating the longitude theta of a ballistic drop point by adopting a nonlinear dynamics model f And ballistic drop point latitude phi f ；
The ballistic drop longitude θ f And ballistic drop point latitude phi f The calculation process of (2) is as follows:
s201: setting the aircraft mass m and the drag coefficient C of the aircraft D Coefficient of lift C L Parameters of reference area S and atmospheric density rho, and initializing parameters of a whole-course attack angle alpha and a roll angle sigma of the aircraft;
s202: setting an initial point of trajectory calculation:
namely: initializing a state quantity of an aircraft, comprising: distance r between the centers of the earth 0 And the longitude theta is theta 0 Latitude phi ═ phi 0 Velocity V ═ V 0 The track angle gamma is gamma 0 And heading angle psi 0 ；
S203: and (3) simulating the next state by adopting a nonlinear dynamics model:
in formula (1):
r is the ground center distance of the aircraft in flight;
v is the flight speed of the aircraft;
gamma is the flight path angle of the aircraft while flying;
θ is the longitude of the aircraft when in flight;
psi is the heading angle of the aircraft while flying;
phi is the latitude of the aircraft in flight;
d is the drag of the aircraft in flight;
m is the mass of the aircraft;
g 0 is the earth's surface gravitational acceleration;
r e is the radius of the earth;
ω e is the angular velocity of the earth's rotation;
l is the lift of the aircraft in flight;
σ is the roll angle of the aircraft in flight;
s204: judging whether the following convergence conditions are met:
namely: judging whether the earth center distance is smaller than the earth radius;
if the geocentric distance is smaller than the radius of the earth, the ballistic trajectory calculation is finished, and the longitude and the latitude of the output calculation result are the longitude of the ballistic trajectory drop point and the latitude of the ballistic trajectory drop point, namely theta f ＝θ,φ f ＝φ；
And if the geocentric distance is not smaller than the radius of the earth, returning to S203, and performing a new round of iterative solution.
S3: for each group of sample points, respectively calculating the maximum value theta of the longitude of the corresponding trajectory drop point fmax (iii), minimum value of trajectory drop point longitude θ fmin Maximum value phi of trajectory drop point latitude fmax And the minimum value of the trajectory drop point latitude, phi fmin And the two maximum values and the corresponding minimum values are combinedValues as ranges, forming inputs as initial values r for ballistic calculations 0 ,θ 0 ,φ 0 ,V 0 ,γ 0 ,ψ 0 Outputting a sample library which is the maximum value and the minimum value of the trajectory drop point longitude and latitude;
s4: for each group of sample points, training by adopting a back propagation method to obtain a flight range prediction neural network;
the training method for predicting the back propagation of the neural network in the flight range comprises the following steps:
s401: initializing a network parameter w 0 ；
S402: calculating a network prediction error Loss:
namely:
in formula (2):
θ fmax the maximum value of the trajectory drop point longitude;
θ fmin is the minimum value of the trajectory drop point longitude;
φ fmax is the maximum value of the trajectory drop point latitude;
,φ fmin is the minimum value of the trajectory drop point latitude;
if Loss is less than threshold L max If so, ending the training, otherwise, turning to S403;
s403: adjusting a network parameter w by adopting a gradient descent method:
in formula (3):
eta is the set learning rate;
s402 is repeated after completion.
S5: when the aircraft flies online, inputting the current flying state of the aircraft, adopting each training flying range prediction neural network obtained in S4 to predict the flying range, performing integrated neural network prediction on the prediction result, and further obtaining the flying prediction result.
The integrated neural network presupposes the following steps:
s501: setting n net A neural network, and respectively inputting the current flight state r of the aircraft into each neural network s ,θ s ,φ s ,V s ,γ s ,ψ s And indicating the latitude and longitude range theta of the landing point fmax i ,φ fmax i ,θ fmin i ,φ fmin i ,i＝1,2…n net ；
S502: and geometrically averaging the calculation results of each neural network obtained in the step S501 to obtain a final prediction range result:
it will be evident to those skilled in the art that the invention is not limited to the details of the foregoing illustrative embodiments, and that the present invention may be embodied in other specific forms without departing from the spirit or essential attributes thereof. The present embodiments are therefore to be considered in all respects as illustrative and not restrictive, the scope of the invention being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein. Any reference sign in a claim should not be construed as limiting the claim concerned.
Furthermore, it should be understood that although the present description refers to embodiments, not every embodiment may contain only a single embodiment, and such description is for clarity only, and those skilled in the art should integrate the description, and the embodiments may be combined as appropriate to form other embodiments understood by those skilled in the art.
Claims (4)
1. An on-line prediction method for flight range of a glide section of an aircraft based on an integrated neural network is characterized in that: the method comprises the following steps:
s1: before the aircraft takes off, selecting a plurality of flight state quantity sample points on a preset standard track, wherein the flight state quantity sample points comprise: ground center distance r before flight of aircraft 0 Longitude θ, longitude 0 Latitude phi 0 Velocity V 0 Track angle gamma 0 And heading angle psi 0 Dividing the plurality of sample points into a plurality of groups at random;
s2: respectively taking each group of sample points as initial ballistic calculation values, and calculating the longitude and latitude of a ballistic falling point;
s3: for each group of sample points, respectively calculating the maximum value of the trajectory drop point longitude, the minimum value of the trajectory drop point longitude, the maximum value of the trajectory drop point latitude and the minimum value of the trajectory drop point latitude, and taking the two maximum values and the corresponding minimum value as ranges to form a sample library which is input as an initial trajectory calculation value and output as the maximum value and the minimum value of the trajectory drop point longitude and latitude;
s4: respectively training each group of sample points to obtain a flight range prediction neural network;
s5: when the aircraft flies online, inputting the current flying state of the aircraft, adopting each training flying range prediction neural network obtained in S4 to predict the flying range, performing integrated neural network prediction on the prediction result, and further obtaining the flying prediction result.
2. The integrated neural network-based aircraft glide slope flight range online prediction method of claim 1, wherein: s2, the calculation process of the trajectory drop point longitude and trajectory drop point latitude is as follows:
s201: setting the aircraft mass m and the drag coefficient C of the aircraft D Coefficient of lift C L Reference area S and atmospheric density rho, and attack the whole course of the aircraftInitializing parameters of an angle alpha and a roll angle sigma;
s202: setting an initial point of trajectory calculation:
namely: initializing a state quantity of an aircraft, comprising: distance r between the centers of the earth 0 And the longitude theta is theta 0 Latitude phi ═ phi 0 Velocity V ═ V 0 The track angle gamma is gamma 0 And heading angle psi ═ psi 0 ；
S203: and (3) simulating the next state by adopting a nonlinear dynamics model:
in formula (1):
r is the ground center distance of the aircraft in flight;
v is the flight speed of the aircraft;
gamma is the flight path angle of the aircraft while flying;
θ is the longitude of the aircraft when in flight;
ψ is the heading angle of the aircraft when in flight;
phi is the latitude of the aircraft in flight;
d is the drag of the aircraft in flight;
m is the mass of the aircraft;
g 0 is the earth's surface gravitational acceleration;
r e is the radius of the earth;
ω e is the angular velocity of the earth's rotation;
l is the lift of the aircraft in flight;
σ is the roll angle of the aircraft in flight;
s204: judging whether the following convergence conditions are met:
namely: judging whether the earth center distance is smaller than the earth radius;
if the geocentric distance is smaller than the radius of the earth, the ballistic trajectory calculation is finished, and the longitude theta and the latitude phi of the output calculation result are the longitude theta of the ballistic trajectory drop point f And trajectory drop point latitude phi f Namely: theta f ＝θ,φ f ＝φ；
And if the geocentric distance is not smaller than the radius of the earth, returning to S203, and performing a new round of iterative solution.
3. The integrated neural network-based aircraft glide slope flight range online prediction method of claim 2, wherein: s4 the training method for predicting neural network back propagation in the flight range is as follows:
s401: initializing a network parameter w 0 ；
S402: calculating a network prediction error Loss:
namely:
in formula (2):
θ f max is the maximum value of the trajectory drop point longitude;
θ f min is the minimum value of the trajectory drop longitude;
φ f max is the maximum value of the trajectory drop point latitude;
φ f min is the minimum value of the trajectory drop point latitude;
if Loss is less than threshold L max If so, ending the training, otherwise, turning to S403;
s403: adjusting a network parameter w by adopting a gradient descent method:
in formula (3):
eta is the set learning rate;
s402 is repeated after completion.
4. The integrated neural network-based aircraft glide slope flight range online prediction method of claim 3, wherein: s5 the integrated neural network presupposes the following steps:
s501: setting n net A neural network, and respectively inputting the current flight state r of the aircraft into each neural network s ,θ s ,φ s ,V s ,γ s ,ψ s And indicating the latitude and longitude range theta of the landing point f max i ,φ f max i ,θ f min i ,φ f min i ,i＝1,2…n net ；
S502: and geometrically averaging the calculation results of each neural network obtained in the step S501 to obtain a final prediction range result:
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210102721.8A CN114417730B (en) | 2022-01-27 | 2022-01-27 | Aircraft glide section flight range online prediction method based on integrated neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210102721.8A CN114417730B (en) | 2022-01-27 | 2022-01-27 | Aircraft glide section flight range online prediction method based on integrated neural network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114417730A CN114417730A (en) | 2022-04-29 |
| CN114417730B true CN114417730B (en) | 2022-09-16 |
Family
ID=81278964
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210102721.8A Active CN114417730B (en) | 2022-01-27 | 2022-01-27 | Aircraft glide section flight range online prediction method based on integrated neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114417730B (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5774831A (en) * | 1996-12-06 | 1998-06-30 | Gupta; Surender Kumar | System for improving average accuracy of signals from global positioning system by using a neural network to obtain signal correction values |
| AU2014201918A1 (en) * | 2012-03-12 | 2014-04-24 | The Boeing Company | Method and apparatus for identifying structural deformation |
| CN105278545A (en) * | 2015-11-04 | 2016-01-27 | 北京航空航天大学 | Active-disturbance-rejection trajectory linearization control method suitable for hypersonic velocity maneuvering flight |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5631830A (en) * | 1995-02-03 | 1997-05-20 | Loral Vought Systems Corporation | Dual-control scheme for improved missle maneuverability |
| CN104049640B (en) * | 2014-06-27 | 2016-06-15 | 金陵科技学院 | Unmanned vehicle attitude robust fault tolerant control method based on Neural Network Observer |
| US9589210B1 (en) * | 2015-08-26 | 2017-03-07 | Digitalglobe, Inc. | Broad area geospatial object detection using autogenerated deep learning models |
| US9911339B2 (en) * | 2015-11-05 | 2018-03-06 | Ge Aviation Systems Llc | Experimental real-time performance enhancement for aircraft |
| CN106168807B (en) * | 2016-09-09 | 2018-01-09 | 腾讯科技（深圳）有限公司 | The flight control method and flight control assemblies of a kind of aircraft |
-
2022
- 2022-01-27 CN CN202210102721.8A patent/CN114417730B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5774831A (en) * | 1996-12-06 | 1998-06-30 | Gupta; Surender Kumar | System for improving average accuracy of signals from global positioning system by using a neural network to obtain signal correction values |
| AU2014201918A1 (en) * | 2012-03-12 | 2014-04-24 | The Boeing Company | Method and apparatus for identifying structural deformation |
| CN105278545A (en) * | 2015-11-04 | 2016-01-27 | 北京航空航天大学 | Active-disturbance-rejection trajectory linearization control method suitable for hypersonic velocity maneuvering flight |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114417730A (en) | 2022-04-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2021036778A1 (en) | Method for directly planning reentry trajectory in altitude-velocity profile | |
| US10520389B2 (en) | Aerodynamic modeling using flight data | |
| CN111399541B (en) | Unmanned aerial vehicle whole-region reconnaissance path planning method of unsupervised learning type neural network | |
| Paul et al. | Wind-aware trajectory planning for fixed-wing aircraft in loss of thrust emergencies | |
| Baklacioglu et al. | Aero-propulsive modelling for climb and descent trajectory prediction of transport aircraft using genetic algorithms | |
| CN114065398B (en) | Flight performance calculation method for high-aspect-ratio flexible aircraft | |
| CN109858137B (en) | Complex maneuvering aircraft track estimation method based on learnable extended Kalman filtering | |
| CN113325735B (en) | Particle swarm algorithm-based rapid water-feeding fire-extinguishing strategy planning method | |
| CN112506218B (en) | Reentry vehicle any no-fly zone around-flying method based on intelligent trajectory prediction | |
| CN113126643A (en) | Intelligent robust reentry guidance method and system for hypersonic aircraft | |
| CN113325706B (en) | Hypersonic aircraft reentry trajectory optimization method based on improved control parameterization | |
| CN114417730B (en) | Aircraft glide section flight range online prediction method based on integrated neural network | |
| CN113761662B (en) | Generation method of trajectory prediction pipeline of gliding target | |
| CN114442673B (en) | Intelligent aircraft trajectory planning method based on chaotic model and DNN | |
| CN108398883B (en) | RLV approach landing trajectory rapid deduction and determination method | |
| WO2019186151A1 (en) | Methods and apparatus for simulating liquid collection on aerodynamic components | |
| CN113110558B (en) | Hybrid propulsion unmanned aerial vehicle demand power prediction method | |
| CN115963855A (en) | Unpowered reentry aircraft landing area prediction method based on deep learning | |
| CN106483967B (en) | A kind of dirigible pitch angle antihunt means based on angular velocity information measurement and sliding formwork | |
| Paul | Emergency trajectory generation for fixed-wing aircraft | |
| CN114578846A (en) | AGIMM tracking method based on maneuver detection sorting | |
| McKenna et al. | Online Parameter Estimation Within Trajectory Optimization for Dynamic Soaring | |
| He et al. | Improved Pigeon-Inspired Optimization Algorithm and Its Application to Minimum-Fuel Ascent Trajectory Optimization | |
| CN114115314B (en) | Sweep angle determining method and system for variant aircraft | |
| EP4328594A1 (en) | Air data estimation in aircraft |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |