CN109634309A - A kind of aircraft automatic obstacle avoiding system, method and aircraft - Google Patents
A kind of aircraft automatic obstacle avoiding system, method and aircraft Download PDFInfo
- Publication number
- CN109634309A CN109634309A CN201910132801.6A CN201910132801A CN109634309A CN 109634309 A CN109634309 A CN 109634309A CN 201910132801 A CN201910132801 A CN 201910132801A CN 109634309 A CN109634309 A CN 109634309A
- Authority
- CN
- China
- Prior art keywords
- aircraft
- flight
- module
- image
- automatic obstacle
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 42
- 238000004458 analytical method Methods 0.000 claims abstract description 25
- 238000013461 design Methods 0.000 claims abstract description 25
- 230000004888 barrier function Effects 0.000 claims description 31
- 238000001514 detection method Methods 0.000 claims description 23
- 238000012545 processing Methods 0.000 claims description 16
- 238000001914 filtration Methods 0.000 claims description 15
- 238000012549 training Methods 0.000 claims description 10
- 230000010006 flight Effects 0.000 claims description 8
- 238000003062 neural network model Methods 0.000 claims description 8
- 210000002569 neuron Anatomy 0.000 claims description 8
- 230000003044 adaptive effect Effects 0.000 claims description 7
- 230000008859 change Effects 0.000 claims description 7
- 230000033001 locomotion Effects 0.000 claims description 7
- 230000003287 optical effect Effects 0.000 claims description 7
- 230000004913 activation Effects 0.000 claims description 6
- 230000004075 alteration Effects 0.000 claims description 6
- 238000004364 calculation method Methods 0.000 claims description 6
- 230000005284 excitation Effects 0.000 claims description 6
- 238000013528 artificial neural network Methods 0.000 claims description 5
- 238000005070 sampling Methods 0.000 claims description 5
- 230000011664 signaling Effects 0.000 claims description 5
- 108010076504 Protein Sorting Signals Proteins 0.000 claims description 3
- 230000008878 coupling Effects 0.000 claims description 3
- 238000010168 coupling process Methods 0.000 claims description 3
- 238000005859 coupling reaction Methods 0.000 claims description 3
- 238000013016 damping Methods 0.000 claims description 3
- 238000009826 distribution Methods 0.000 claims description 3
- 238000012360 testing method Methods 0.000 claims description 3
- 210000005036 nerve Anatomy 0.000 claims 1
- 238000005516 engineering process Methods 0.000 abstract description 4
- 230000003247 decreasing effect Effects 0.000 abstract description 2
- 239000000523 sample Substances 0.000 description 28
- 239000003795 chemical substances by application Substances 0.000 description 14
- 239000002245 particle Substances 0.000 description 7
- 239000013598 vector Substances 0.000 description 7
- 238000003066 decision tree Methods 0.000 description 6
- 239000004615 ingredient Substances 0.000 description 6
- 230000000694 effects Effects 0.000 description 5
- 230000008569 process Effects 0.000 description 3
- 230000008901 benefit Effects 0.000 description 2
- 238000005260 corrosion Methods 0.000 description 2
- 230000007797 corrosion Effects 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000000877 morphologic effect Effects 0.000 description 2
- 230000011218 segmentation Effects 0.000 description 2
- 238000013112 stability test Methods 0.000 description 2
- 238000012512 characterization method Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 230000010339 dilation Effects 0.000 description 1
- 238000006073 displacement reaction Methods 0.000 description 1
- 230000008030 elimination Effects 0.000 description 1
- 238000003379 elimination reaction Methods 0.000 description 1
- 230000003628 erosive effect Effects 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000013138 pruning Methods 0.000 description 1
- 239000013589 supplement Substances 0.000 description 1
- 239000011800 void material Substances 0.000 description 1
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course or altitude of land, water, air, or space vehicles, e.g. automatic pilot
- G05D1/10—Simultaneous control of position or course in three dimensions
- G05D1/101—Simultaneous control of position or course in three dimensions specially adapted for aircraft
Abstract
The invention belongs to vehicle technology fields, a kind of aircraft automatic obstacle avoiding system, method and aircraft are disclosed, automatic obstacle avoiding system includes: solar powered module, image capture module, complaint message detecting module, central control module, analyzes and determines module, directive generation module, path planning module, stability analysis module, display module.Passage path planning module of the present invention adjusts the flight program track when judging that other described aircraft and the aircraft can bump against, therefore can be ensured to be the flight path of aircraft planning safety, greatly improves flight safety;Meanwhile calculating cost can be greatly decreased while guaranteeing precision of prediction by stability analysis module, to be easy to carry out the qualitative and quantitative study of the full flight domain dynamic stability characteristic of aircraft, obtain the design direction for having directive significance to Flight Vehicle Design.
Description
Technical field
The invention belongs to vehicle technology field more particularly to a kind of aircraft automatic obstacle avoiding systems, method and aircraft.
Background technique
Aircraft (flight vehicle) is the instrument to fly in endoatmosphere or exoatmosphere space (space).Flight
Device is divided into 3 classes: aircraft, spacecraft, rocket and guided missile.In endoatmosphere, the referred to as aircraft of flight, such as balloon, dirigible, flies
Machine etc..They go up to the air by the air force that the quiet buoyancy of air or air relative motion generate and fly.In being known as space flight
Spacecraft, such as artificial earth satellite, manned spaceship, space probe, space shuttle.They are obtained under the promotion of carrier rocket
It obtains necessary speed and enters space, the track movement similar with celestial body is then done by inertia.However, existing aircraft is independently kept away
It is low to hinder security of system;Meanwhile it is complicated, cumbersome to the stability test of aircraft flight, analytical error is big.
In conclusion problem of the existing technology is:
(1) existing aircraft automatic obstacle avoiding security of system is low;Meanwhile complicated to the stability test of aircraft flight,
It is cumbersome, analytical error is big.
(2) during acquiring the ambient image in front of aircraft flight in the prior art, weight is blocked in multiple mobile object
In the case where folded, using conventional target track algorithm, motion target tracking success rate cannot be effectively improved.
(3) millimetre-wave radar carries out multi-target detection to the barrier in the detectable distance in front in the prior art, obtains
During the position of barrier, speed and azimuth information, in work in complicated detection background, it is difficult in echo-signal
It avoids adulterating many clutter ingredients, using traditional algorithm, clutter ingredient cannot be effectively eliminated in the form of false target
It carries over.
(4) analyze and determine according to the image of acquisition, complaint message the process of the risk of barrier in the prior art
In, the decision tree of the degree of danger of barrier is established using traditional algorithm according to the image of acquisition, obstruction data information, it cannot
Quickly determine barrier risk.
Summary of the invention
In view of the problems of the existing technology, the present invention provides a kind of aircraft automatic obstacle avoiding system, method and flights
Device.
The invention is realized in this way a kind of aircraft automatic obstacle avoiding method, the aircraft automatic obstacle avoiding method include
Following steps:
Step 1 is converted solar energy into electrical energy using solar panel and is powered for aircraft;It is adopted by image
Collect module using the ambient image in front of optical camera device acquisition aircraft flight;
Step 2 is detected mould by complaint message and is carried out using millimetre-wave radar to the barrier in the detectable distance in front
Multi-target detection, position, speed and the azimuth information of acquired disturbance object;
Step 3 analyze and determine according to image, the complaint message of acquisition the danger of barrier using data analysis program
It is dangerous;
Step 4 generates change of voyage according to judging result using instruction generator and instructs;Passage path planning module benefit
The flight path of aircraft is planned according to alteration command with navigation system;
Step 5 analyzes the stability of aircraft using analysis program;
Step 6 shows acquisition ambient image, complaint message, flight path, stability data using display.
Further, the path planning module planing method includes:
(1) change of voyage instruction is received, generates flight program track for the aircraft, the flight program track includes
Flight program route and flight program height layer;
(2) when having other aircraft flights except the aircraft in the flight program height layer, according to default
Rule judges whether other described aircraft can bump against with the aircraft;
(3) when judging that other described aircraft and the aircraft can bump against, the flight program track is adjusted, until
Judge that other described aircraft will not bump against with the aircraft according to the preset rules.
Further, described to judge other described aircraft include: with whether the aircraft bump against according to preset rules
Obtain the flight path and flying speed of other aircraft;
When the flight path of the flight program track and other aircraft has overlapping region, according to it is described other
The flight path and flying speed of aircraft and the flight program track of the aircraft and flying speed, calculate it is described other
Aircraft and the aircraft reach the time point interval of the overlapping region;
When the time point interval is less than or equal to prefixed time interval, other described aircraft and the aircraft are judged
It can bump against.
Further, the flight program track of the adjustment aircraft includes:
By adjusting flight program height layer of the aircraft at at least described overlapping region and/or flight program road
Line adjusts the flight program track;The flight program height layer adjusted of the aircraft is located under the aircraft
It limits between flying height and upper limit flying height.
Further, method for analyzing stability includes:
1) aerocraft flying parameter is set;
2) Aerodynamic Characteristics are directed to, determine aimed at precision requirement and the calculation amount upper limit, and in full flight domain range
Choose the initial sample point of predetermined quantity;
3) at initial sample point x (i), one group of given training signal sequence u (x (i), k) is calculated using CFD
Obtain aircraft forced movement unsteady aerodynamic force and torque coefficient when ordinal series y (x (i), k), then with u (x (i), k)
It is sample data with y (x (i), k), constructs Kriging agent model as initial target proxy model;
4) building refers to agent model, assesses initial sample point, to determine present sample precision;
5) new candidate sampled point is generated using test design method;And by target proxy model and refer to agent model
All candidate sampled points are respectively calculated and are assessed, target proxy model and reference is added to obtain each candidate sampled point
The precision changing value generated after agent model;
6) the candidate sampled point for meeting present sample required precision in precision changing value is chosen by adaptively sampled criterion
It is trained as next addition target proxy model and with reference to the sample point of agent model;
7) step 4) is repeated to 5), then judges whether current goal agent model precision is up to standard and counts in each repeat
Whether calculation amount reaches the upper limit, and previous cycle is exited when meeting one of condition, completes adopting for current goal agent model
Sample simultaneously determines full flight domain unsteady aerodynamic force agency-reduced-order model training sample with this;
8) a large amount of random distributions, the design point full of full flight domain are generated in full flight domain with arbitrary sampling method;It is right
Each design point is obtained the input/output signal of each design point using Kriging interpolation in training sample, recycles each input
The discrete space that output signal is formed determines unsteady aerodynamic force reduced-order model;
9) unsteady aerodynamic force reduced-order model is converted to the state space equation of continuous space;Rigid body dynamic equation is turned
Turn to the state equation under continuous space;By the state space equation of unsteady aerodynamic force reduced-order model and rigid body dynamic equation
State equation carries out feedback link to get the coupled dynamic stability analysis equation of current flight device is arrived;Solve coupled dynamic stability
The eigenmatrix characteristic root of analysis equation, characteristic root real part characterize system damping, and imaginary part characterizes system frequency;Wherein, when all
When characteristic root real part is all negative, the aircraft dynamic stability of the design point;When there is the characteristic root of positive real part, the design point
Aircraft moves unstable;
10) after the dynamic stability feature for obtaining the aircraft of each design point through the above steps, aircraft is obtained whole
Dynamic stability feature in a flight domain.
Further, step 3 carries out analytical judgment obstacle according to image, the complaint message of acquisition using data analysis program
In the risk of object,
Obstacle micro-image is detected using the Pulse-coupled Neural Network Model of suitable processing obstacle image information;
Obstacle micro-image is handled by the lesser impulsive noise pollution of density by adaptive weighted filter;Obstacle micro-image by
The biggish impulsive noise pollution of density is secondary using keeping the introducing binode constitutive element mathematical morphology of edge detail information to carry out
Barrier hazard information is obtained after filtering.
Further, it is suitble to the Pulse-coupled Neural Network Model of processing obstacle image information:
Fij[n]=Sij;
Uij[n]=Fij[n](1+βij[n]Lij[n]);
θij[n]=θ0e-αθ(n-1);
Wherein, βij[n] is adaptive link strength factor;
Sij、Fij[n]、Lij[n]、Uij[n]、θij[n] is respectively received image signal, feed back input, link input, inside
Active entry and dynamic threshold, NwFor the sum of all pixels in selected window W to be processed, Δ is adjustment factor, chooses 1~3.
Further, when Pulse-coupled Neural Network Model detects obstacle micro-image, make ash using network characteristic
Degree is SijmaxPixel light a fire activation, then second of Pulse Coupled Neural Network iterative processing is carried out, between [Sijmax/1+β
ijLij,Sijmax] between pixel capture activation, make the corresponding Y of the pixel activated twiceijOutput is 1;Then dirty to former noise
Dye image highlights processing, then to treated image SijIt is iterated processing by aforementioned, and makes corresponding output Yij=1, it utilizes
Picture noise pixel and surrounding pixel correlation are small, the big characteristic of gray scale difference, where the excitation of a neuron does not cause
When the excitation of the most of neurons of areas adjacent, just illustrate that the neuron respective pixel may be noise spot;
Tentatively screen out Yij=0 corresponding pixel is the signaling point of obstacle micro-image, is protected;To YijOutput
It counts for 1 pixel within the scope of 3*3 template B to export Yij=1 is the number N that center neighborhood element value is 1YDifferentiation is returned
Class: 1≤NY≤ 8, it is noise spot, works as NY=9, it is determined as image slices vegetarian refreshments.
Another object of the present invention is to provide a kind of flights of aircraft automatic obstacle avoiding method described in implementation claim
Device automatic obstacle avoiding system, the aircraft automatic obstacle avoiding system include:
Solar powered module is connect with central control module, for being converted solar energy by solar panel
Electric energy is powered for aircraft;
Image capture module is connect with central control module, for being acquired in front of aircraft flight by optical camera device
Ambient image;
Complaint message detecting module, connect with central control module, for by millimetre-wave radar it is detectable to front away from
Multi-target detection, position, speed and the azimuth information of acquired disturbance object are carried out from interior barrier;
Central control module, with solar powered module, image capture module, complaint message detecting module, analytical judgment
Module, directive generation module, path planning module, stability analysis module, display module connection, for passing through central controller
Modules are controlled to work normally;
It analyzes and determines module, is connect with central control module, for image, the barrier by data analysis program according to acquisition
Information is hindered analyze and determine the risk of barrier;
Directive generation module is connect with central control module, is navigated for being generated by instruction generator according to judging result
Line alteration command;
Path planning module is connect with central control module, is flown for being planned by navigation system according to alteration command
The flight path of device;
Stability analysis module, connect with central control module, for by analysis program to the stability of aircraft into
Row analysis;
Display module is connect with central control module, for by display display acquire ambient image, complaint message,
Flight path, stability data.
Another object of the present invention is to provide a kind of aircraft.
Advantages of the present invention and good effect are as follows: passage path planning module of the present invention in flight program height layer
When stating other aircraft flights except aircraft, judge whether are other described aircraft and the aircraft according to preset rules
It can bump against, and when judging that other described aircraft and the aircraft can bump against, adjust the flight program track, therefore energy
It is enough ensured to be the flight path of aircraft planning safety, greatly improves flight safety;Meanwhile passing through stability analysis module pair
The full flight domain multi-channel coupling dynamic stability problem analysis of aircraft, propose it is a kind of will be based on the non-of the efficient self-adapted method of sampling
Unsteady Flow reduced-order model and Dynamical Equations of Rigid Body are carried out in the method that State-Space Coupling solves eigenmatrix characteristic root
The full flight domain dynamic stability feature prediction of aircraft, can be greatly decreased calculating cost while guaranteeing precision of prediction, thus
It is easy to carry out the qualitative and quantitative study of the full flight domain dynamic stability characteristic of aircraft, acquisition has directive significance to Flight Vehicle Design
Design direction.
Image capture module acquires the process of the ambient image in front of aircraft flight by optical camera device in the present invention
In, in the case where multiple mobile object blocks overlapping, using the multiple target tracking algorithm based on particle filter, improve moving target
Success rate is tracked, target following success rate is made to increase 39.5 percentage points, algorithm is averaged time-consuming 0.78s.
Complaint message detecting module carries out the barrier in the detectable distance in front by millimetre-wave radar in the present invention
Multi-target detection, during the position of acquired disturbance object, speed and azimuth information, in work in complicated detection background
In, it is difficult to avoid that in echo-signal and adulterates many clutter ingredients, using LVQ clustering algorithm, clutter ingredient can be eliminated with void
The form of decoy carries over.
Analyze and determine that module carries out analysis according to image, the complaint message of acquisition by data analysis program and sentences in the present invention
During the risk of disconnected barrier, algorithm is analyzed and determined based on improved decision tree data using one kind, it can be according to adopting
The image of collection, obstruction data information quickly establish the decision tree of the degree of danger of barrier, quickly to determine barrier risk
Foundation is provided.
The present invention is by improvement Pulse Coupled Neural Network without detecting obstacle automatically in the case where setting detection threshold value
Noise in micro-image completes the removal of noise using multistage combination filter, while effectively filtering out noise jamming very
The information such as image edge detailss are protected well.
The present invention has the effect that
The present invention provides characteristic using the lock-out pulse of Pulse Coupled Neural Network and distinguishes position pulse noise spot and signal
Pixel position, it is relatively traditional that higher noise detection property is had based on value detection method in intermediate value detection or related improvement
Can, relative to other threshold value noise detection methods;For the present invention without setting detection threshold value, noise fallout ratio and omission factor are low, make an uproar
Sound detection precision is higher;Meanwhile relative to other noise iteration detection methods;The method of the present invention detection time is short, and automaticity is strong;
There is presently no any impulse noise correction methods to apply in the detection of obstacle micro-image impulsive noise;Hindering
Micro-image impulsive noise is hindered to filter out the stage, the present invention is first according to the above-mentioned noise detected and signaling point, to image pixel
Carry out classification processing;Only the noise spot of detection is filtered when using first order adaptive weighted filter, relative to
The methods of other median filterings, Wiener filtering protect signaling point information while effectively filtering out noise;In second level mathematics
It is to carry out supplement auxiliary to the related noise missed in prime filtering to filter out when morphologic filtering, can not only has while denoising
Effect filters out noise jamming, and can protect the information such as image edge detailss well;
With stronger subjective vision effect and index is objectively evaluated, noise removal capability is strong, signal-to-noise ratio is high and adaptability is good, special
It is not to the obstacle micro-image by serious noise pollution, it is shown that bigger filtering superiority.
Detailed description of the invention
Fig. 1 is aircraft automatic obstacle avoiding method flow chart provided in an embodiment of the present invention.
Fig. 2 is aircraft automatic obstacle avoiding system construction drawing provided in an embodiment of the present invention.
In Fig. 2: 1, solar powered module;2, image capture module;3, complaint message detecting module;4, center control mould
Block;5, module is analyzed and determined;6, directive generation module;7, path planning module;8, stability analysis module;9, display module.
Specific embodiment
In order to further understand the content, features and effects of the present invention, the following examples are hereby given, and cooperate attached drawing
Detailed description includes.
Structure of the invention is explained in detail with reference to the accompanying drawing.
As shown in Figure 1, aircraft automatic obstacle avoiding method provided by the invention the following steps are included:
S101: it is powered firstly, solar panel converts solar energy into electrical energy for aircraft.
S102: the ambient image in front of acquisition aircraft flight, by radar to the barrier in the detectable distance in front
Carry out multi-target detection, position, speed and the azimuth information of acquired disturbance object.
S103: according to the data information of above-mentioned acquisition, carrying out the risk for analyzing and determining barrier, raw according to judging result
It is instructed at change of voyage, plans the flight path of aircraft.
S104: the state of flight information of aircraft is obtained, the stability of aircraft is analyzed.
S105: acquisition ambient image, complaint message, flight path, stability data are shown by display.
Step S103 analyze and determine according to image, the complaint message of acquisition the danger of barrier using data analysis program
In dangerous,
Obstacle micro-image is detected using the Pulse-coupled Neural Network Model of suitable processing obstacle image information;
Obstacle micro-image is handled by the lesser impulsive noise pollution of density by adaptive weighted filter;Obstacle micro-image by
The biggish impulsive noise pollution of density is secondary using keeping the introducing binode constitutive element mathematical morphology of edge detail information to carry out
Barrier hazard information is obtained after filtering.
It is suitble to the Pulse-coupled Neural Network Model of processing obstacle image information:
Fij[n]=Sij;
Uij[n]=Fij[n](1+βij[n]Lij[n]);
θij[n]=θ0e-αθ(n-1);
Wherein, βij[n] is adaptive link strength factor;
Sij、Fij[n]、Lij[n]、Uij[n]、θij[n] is respectively received image signal, feed back input, link input, inside
Active entry and dynamic threshold, NwFor the sum of all pixels in selected window W to be processed, Δ is adjustment factor, chooses 1~3.
When Pulse-coupled Neural Network Model detects obstacle micro-image, make gray scale S using network characteristicijmax
Pixel light a fire activation, then second of Pulse Coupled Neural Network iterative processing is carried out, between [Sijmax/1+βijLij,Sijmax]
Between pixel capture activation, make the corresponding Y of the pixel activated twiceijOutput is 1;Then place is highlighted to former image polluted by noise
Reason, then to treated image SijIt is iterated processing by aforementioned, and makes corresponding output Yij=1, utilize picture noise pixel
It is small with surrounding pixel correlation, the big characteristic of gray scale difference, when the excitation of a neuron do not cause it is most near region
When the excitation of number neuron, just illustrate that the neuron respective pixel may be noise spot;
Tentatively screen out Yij=0 corresponding pixel is the signaling point of obstacle micro-image, is protected;To YijOutput
It counts for 1 pixel within the scope of 3*3 template B to export Yij=1 is the number N that center neighborhood element value is 1YDifferentiation is returned
Class: 1≤NY≤ 8, it is noise spot, works as NY=9, it is determined as image slices vegetarian refreshments;
The implementation method of obstacle micro-image adaptive weighting filter noise filtering;
When pulse exports Yij=1 and NY=1~8, NYIt is to work as in 3*3 template B for 1 number, chooses filter window M, it is right
Image polluted by noise fijAdaptive-filtering, filtering equations are as follows:
In formula, xrsIt is the coefficient of respective pixel in filter window, SrsFor the gray value of respective pixel in filter window, fij
To correspond to the output valve of window center position after filtering:
D in formulaijFor pixel grey scale intermediate value in box filter window M, ΩijEach pixel of filter window and center gray scale difference are exhausted
To mean value, max is maximizing symbol;
It chooses filter window M and chooses the filter window M that size is m*m, the selection principle of window size is:
The specific method of binode constitutive element mathematical morphology second level filtering:
The obstacle micro-image of residual impulse noise is f, and E is structural element SE, then expansion has following relational expression:
In formulaFor dilation operation symbol, F and G are the domain of f and E respectively, and x-z is displacement parameter;
Above formula extension relationship is all to be merged into all background dots contacted with object in object, expands boundary to outside
Process, fill up the hole in object;
Above formula Θ is erosion operation symbol, and corrosion is to eliminate boundary point, and boundary is internally shunk, while in the base of corrosion expansion
On plinth, in conjunction with morphologic opening and closing operation:
As shown in Fig. 2, aircraft automatic obstacle avoiding system provided in an embodiment of the present invention includes: solar powered module 1, figure
As acquisition module 2, complaint message detecting module 3, central control module 4, analyze and determine module 5, directive generation module 6, path
Planning module 7, stability analysis module 8, display module 9.
Solar powered module 1 is connect with central control module 4, for being converted solar energy by solar panel
It is that aircraft is powered for electric energy;
Image capture module 2 is connect with central control module 4, before acquiring aircraft flight by optical camera device
The ambient image of side;
Complaint message detecting module 3 is connect with central control module 4, for detectable to front by millimetre-wave radar
Barrier in distance carries out multi-target detection, position, speed and the azimuth information of acquired disturbance object;
Central control module 4, with solar powered module 1, image capture module 2, complaint message detecting module 3, analysis
Judgment module 5, directive generation module 6, path planning module 7, stability analysis module 8, display module 9 connect, for passing through
Central controller controls modules work normally;
Analyze and determine module 5, connect with central control module 4, for pass through data analysis program according to the image of acquisition,
Complaint message analyze and determine the risk of barrier;
Directive generation module 6 is connect with central control module 4, for being generated by instruction generator according to judging result
Change of voyage instruction;
Path planning module 7 is connect with central control module 4, is flown for being planned by navigation system according to alteration command
The flight path of row device;
Stability analysis module 8 is connect with central control module 4, for the stability by analysis program to aircraft
It is analyzed;
Display module 9 is connect with central control module 4, for showing that acquisition ambient image, obstacle are believed by display
Breath, flight path, stability data.
During described image acquisition module 2 acquires the ambient image in front of aircraft flight by optical camera device,
In order to motion target tracking success rate be improved, using based on the more of particle filter in the case where multiple mobile object blocks overlapping
Target tracking algorism, specifically includes the following steps:
Step 1 opens M thread to M target, and Parallel Tracking improves speed;
Step 2 determines the range of moving target in t=0 according to corresponding budget law for each tracking target,
According to the algorithm of particle filter within the scope of this, particle sample set is initializedBy the weight w of t=0 moment all particles '0It is set as 1/N, i=1,
2 ..., N;Assuming that target exists in the moment position t=0
Step 3 enables t=t+1, according to formula
It calculates
Wherein g (x)=- k ' (x);
According to formula
WhereinWithFor candidate target;
Pasteur's coefficient ρ is calculated, which indicates the similarity for the position and target that tracking obtains, take G1=0.85, G2=
0.35;
As ρ >=G1When, it indicates that tracking is normal, executes step 6;Work as G2ρ < G1When, expression is blocked, but part is hidden
Gear executes step 4;As ρ≤G2When, expression is blocked, and is largely blocked, and step 5 is executed;
Step 4 is tracked according to segmented image module, is retrievedReturn step three recalculates Pasteur system
Number;
Step 6, withNew position, redistributes particle position, and update obtains new particle assemblyNew weightCalculating formula is
λ is controling parameter in formula, while according to the far and near with Pasteur's distance setting weight coefficient, Pasteur's distance of tracking target
Nearlyr weight is higher, and the remoter weight of distance is lower, and weight is normalized
Step 7 judges whether the whole frame for having handled video, if not, carrying out the processing of next frame image, executes step
Rapid three;If so, executing step 8;Algorithm terminates, release processing thread.
The complaint message detecting module 3 carries out more mesh to the barrier in the detectable distance in front by millimetre-wave radar
Mark detection during the position of acquired disturbance object, speed and azimuth information, in work in complicated detection background, is returned
It is difficult to avoid that in wave signal and adulterates many clutter ingredients, in order to which the clutter ingredient of elimination is left down in the form of false target
Come, using LVQ clustering algorithm, specifically includes the following steps:
Step 1, the sample data acquired under the conditions of true echo for specific objective (aircraft, ship), distance,
The information such as orientation, the elevation angle are of substantially known;True, the false label of this part known target in sample is set as very, the mark of other targets
Label are all preset as vacation;
Step 2 is respectively to randomly select one to initialize prototype vector in true/false sample set from initial labelsWithAnd it willWithCluster label be respectively set as it is true/false;
Step 3 randomly selects one group of sample vector from all samplesCalculate itself and two prototype vectorsWithDistance, prototype vector lesser for distance valueAccording to sample drawnCategory label yjWithLabel
It is whether consistent, following formula is used respectively
P '=pi·+α·(xj-pj·);
P '=pi·-α·(xj-pi·);
To prototype vectorIt is updated, wherein α ∈ (0,1) is learning rate;
Step 4, when meeting maximum number of iterations or prototype vector updates very little or the condition not updated, output characterization
Very, the prototype vector of decoy point mark characteristicWith
The analytical judgment module 5 is analyzed and determined by data analysis program according to image, the complaint message of acquisition
During the risk of barrier, in order to quickly establish the dangerous journey of barrier according to the image of acquisition, obstruction data information
The decision tree of degree, determines that barrier risk provides foundation to be quick, is sentenced using one kind based on the analysis of improved decision tree data
Disconnected algorithm, is included the steps that in detail below:
Step 1, establishes training set sample T, and T meets the condition for stopping extension, then returns;
Step 2 updates the best division subset that count matrix determines attribute for Category Attributes scan attribute list;
Attribute is divided into q section using wide histogram method for connection attribute, establishes partitioned histogram by step 3
List;
Step 4, to the Gini value of each pure interval computation boundary;
Step 5 carries out beta pruning to section using TESTCASE algorithm;
Step 6 judges whether the tuple of candidate section inside division point segmentation is identical, deletes the identical time of segmentation tuple
Reconnaissance;
Step 7 finds out the best splitting point of entire attribute;The best division of more each attribute, selection one are optimal
Split point divides T for T1And T2;
Step 8, recursively to T1And T2Generate decision tree.
7 planing method of path planning module provided by the invention includes:
(1) change of voyage instruction is received, generates flight program track for the aircraft, the flight program track includes
Flight program route and flight program height layer;
(2) when having other aircraft flights except the aircraft in the flight program height layer, according to default
Rule judges whether other described aircraft can bump against with the aircraft;
(3) when judging that other described aircraft and the aircraft can bump against, the flight program track is adjusted, until
Judge that other described aircraft will not bump against with the aircraft according to the preset rules.
It is provided by the invention to judge whether other described aircraft bump against with the aircraft according to preset rules, it wraps
It includes:
Obtain the flight path and flying speed of other aircraft;
When the flight path of the flight program track and other aircraft has overlapping region, according to it is described other
The flight path and flying speed of aircraft and the flight program track of the aircraft and flying speed, calculate it is described other
Aircraft and the aircraft reach the time point interval of the overlapping region;
When the time point interval is less than or equal to prefixed time interval, other described aircraft and the aircraft are judged
It can bump against.
The flight program track of the adjustment aircraft provided by the invention, comprising:
By adjusting flight program height layer of the aircraft at at least described overlapping region and/or flight program road
Line adjusts the flight program track;The flight program height layer adjusted of the aircraft is located under the aircraft
It limits between flying height and upper limit flying height.
8 analysis method of stability analysis module provided by the invention includes:
1) aerocraft flying parameter is set;
2) Aerodynamic Characteristics are directed to, determine aimed at precision requirement and the calculation amount upper limit, and in full flight domain range
Choose the initial sample point of predetermined quantity;
3) at initial sample point x (i), one group of given training signal sequence u (x (i), k) is calculated using CFD
Obtain aircraft forced movement unsteady aerodynamic force and torque coefficient when ordinal series y (x (i), k), then with u (x (i), k)
It is sample data with y (x (i), k), constructs Kriging agent model as initial target proxy model;
4) building refers to agent model, assesses initial sample point, to determine present sample precision;
5) new candidate sampled point is generated using test design method;And by target proxy model and refer to agent model
All candidate sampled points are respectively calculated and are assessed, target proxy model and reference is added to obtain each candidate sampled point
The precision changing value generated after agent model;
6) the candidate sampled point for meeting present sample required precision in precision changing value is chosen by adaptively sampled criterion
It is trained as next addition target proxy model and with reference to the sample point of agent model;
7) step 4) is repeated to 5), then judges whether current goal agent model precision is up to standard and counts in each repeat
Whether calculation amount reaches the upper limit, and previous cycle is exited when meeting one of condition, completes adopting for current goal agent model
Sample simultaneously determines full flight domain unsteady aerodynamic force agency-reduced-order model training sample with this;
8) a large amount of random distributions, the design point full of full flight domain are generated in full flight domain with arbitrary sampling method;It is right
Each design point is obtained the input/output signal of each design point using Kriging interpolation in training sample, recycles each input
The discrete space that output signal is formed determines unsteady aerodynamic force reduced-order model;
9) unsteady aerodynamic force reduced-order model is converted to the state space equation of continuous space;Rigid body dynamic equation is turned
Turn to the state equation under continuous space;By the state space equation of unsteady aerodynamic force reduced-order model and rigid body dynamic equation
State equation carries out feedback link to get the coupled dynamic stability analysis equation of current flight device is arrived;Solve coupled dynamic stability
The eigenmatrix characteristic root of analysis equation, characteristic root real part characterize system damping, and imaginary part characterizes system frequency;Wherein, when all
When characteristic root real part is all negative, the aircraft dynamic stability of the design point;When there is the characteristic root of positive real part, the design point
Aircraft moves unstable;
10) after the dynamic stability feature for obtaining the aircraft of each design point through the above steps, it can be obtained aircraft
Dynamic stability feature in entire flight domain.
The above is only the preferred embodiments of the present invention, and is not intended to limit the present invention in any form,
Any simple modification made to the above embodiment according to the technical essence of the invention, equivalent variations and modification, belong to
In the range of technical solution of the present invention.
Claims (10)
1. a kind of aircraft automatic obstacle avoiding method, which is characterized in that the aircraft automatic obstacle avoiding method the following steps are included:
Step 1 is converted solar energy into electrical energy using solar panel and is powered for aircraft;Pass through Image Acquisition mould
Block utilizes the ambient image in front of optical camera device acquisition aircraft flight;
Step 2 detects mould by complaint message and carries out more mesh to the barrier in the detectable distance in front using millimetre-wave radar
Mark detection, position, speed and the azimuth information of acquired disturbance object;
Step 3 analyze and determine according to image, the complaint message of acquisition the risk of barrier using data analysis program;
Step 4 generates change of voyage according to judging result using instruction generator and instructs;Passage path planning module is utilized and is led
Boat system plans the flight path of aircraft according to alteration command;
Step 5 analyzes the stability of aircraft using analysis program;
Step 6 shows acquisition ambient image, complaint message, flight path, stability data using display.
2. aircraft automatic obstacle avoiding method as described in claim 1, which is characterized in that the path planning module planing method
Include:
(1) change of voyage instruction is received, generates flight program track for the aircraft, the flight program track includes flight
Programme path and flight program height layer;
(2) when having other aircraft flights except the aircraft in the flight program height layer, according to preset rules
Judge whether other described aircraft can bump against with the aircraft;
(3) when judging that other described aircraft and the aircraft can bump against, the flight program track is adjusted, until according to
The preset rules judge that other described aircraft will not bump against with the aircraft.
3. the high hair of aircraft automatic obstacle avoiding as claimed in claim 2, which is characterized in that it is described judge according to preset rules described in
Other aircraft include: with whether the aircraft can bump against
Obtain the flight path and flying speed of other aircraft;
When flight path in the flight program track and other aircraft has overlapping region, according to other described flights
The flight path and flying speed of device and the flight program track of the aircraft and flying speed calculate other described flights
Device and the aircraft reach the time point interval of the overlapping region;
When the time point interval is less than or equal to prefixed time interval, judge that other described aircraft can phase with the aircraft
It hits.
4. aircraft automatic obstacle avoiding method as claimed in claim 3, which is characterized in that the flight rule of the adjustment aircraft
Drawing track includes:
By adjusting flight program height layer and/or flight program route of the aircraft at at least described overlapping region
Adjust the flight program track;The lower limit that the flight program height layer adjusted of the aircraft is located at the aircraft flies
Between row height and upper limit flying height.
5. aircraft automatic obstacle avoiding method as described in claim 1, which is characterized in that method for analyzing stability includes:
1) aerocraft flying parameter is set;
2) Aerodynamic Characteristics are directed to, determine aimed at precision requirement and the calculation amount upper limit, and choose in full flight domain range
The initial sample point of predetermined quantity;
3) at initial sample point x (i), one group of given training signal sequence u (x (i), k) is calculated using CFD and is obtained
The unsteady aerodynamic force of aircraft forced movement and torque coefficient when ordinal series y (x (i), k), then with u (x (i), k) and y
(x (i), k) is sample data, constructs Kriging agent model as initial target proxy model;
4) building refers to agent model, assesses initial sample point, to determine present sample precision;
5) new candidate sampled point is generated using test design method;And by target proxy model and with reference to agent model to institute
The candidate sampled point having is respectively calculated and assesses, and target proxy model is added and with reference to agency to obtain each candidate sampled point
The precision changing value generated after model;
6) the candidate sampled point conduct for meeting present sample required precision in precision changing value is chosen by adaptively sampled criterion
It is next that target proxy model is added and is trained with reference to the sample point of agent model;
7) step 4) is repeated to 5), then judges whether current goal agent model precision up to standard and calculation amount in each repeat
Whether reach the upper limit, and exit previous cycle when meeting one of condition, completes the sampling of current goal agent model simultaneously
Full flight domain unsteady aerodynamic force agency-reduced-order model training sample is determined with this;
8) a large amount of random distributions, the design point full of full flight domain are generated in full flight domain with arbitrary sampling method;To each
Design point is obtained the input/output signal of each design point using Kriging interpolation in training sample, recycles each input and output
The discrete space that signal is formed determines unsteady aerodynamic force reduced-order model;
9) unsteady aerodynamic force reduced-order model is converted to the state space equation of continuous space;It is by rigid body dynamic is equations turned
State equation under continuous space;By the state of the state space equation of unsteady aerodynamic force reduced-order model and rigid body dynamic equation
Equation carries out feedback link to get the coupled dynamic stability analysis equation of current flight device is arrived;Solve coupled dynamic stability analysis
The eigenmatrix characteristic root of equation, characteristic root real part characterize system damping, and imaginary part characterizes system frequency;Wherein, when all features
When root real part is all negative, the aircraft dynamic stability of the design point;When there is the characteristic root of positive real part, the flight of the design point
Device moves unstable;
10) it after the dynamic stability feature for obtaining the aircraft of each design point through the above steps, obtains aircraft and is entirely flying
Dynamic stability feature in row domain.
6. aircraft automatic obstacle avoiding method as described in claim 1, which is characterized in that step 3 using data analysis program according to
Image, the complaint message of acquisition carry out in the risk of analytical judgment barrier,
Obstacle micro-image is detected using the Pulse-coupled Neural Network Model of suitable processing obstacle image information;Obstacle
Micro-image is handled by the lesser impulsive noise pollution of density by adaptive weighted filter;Obstacle micro-image is by density
Biggish impulsive noise pollution is using the introducing binode constitutive element mathematical morphology progress secondary filtering for keeping edge detail information
After obtain barrier hazard information.
7. aircraft automatic obstacle avoiding method as claimed in claim 6, which is characterized in that Pulse-coupled Neural Network Model is to obstacle
When micro-image is detected, make gray scale S using network characteristicijmaxPixel light a fire activation, then carry out the second subpulse coupling
Neural network iterative processing is closed, between [Sijmax/1+βijLij,Sijmax] between pixel capture activation, make the picture activated twice
The corresponding Y of vegetarian refreshmentsijOutput is 1;Then processing highlighted to former image polluted by noise, then to treated image SijBy it is aforementioned into
Row iteration processing, and make corresponding output Yij=1, small using picture noise pixel and surrounding pixel correlation, gray scale difference is big
Characteristic, as soon as illustrate the nerve when the excitation of neuron does not cause the excitation of most of neurons near region
First respective pixel may be noise spot.
8. aircraft automatic obstacle avoiding method as claimed in claim 7, which is characterized in that tentatively screen out Yij=0 corresponding pixel
Point is the signaling point of obstacle micro-image, is protected;To YijOutput counts for 1 pixel within the scope of 3*3 template B with defeated
Y outij=1 is the number N that center neighborhood element value is 1YDifferentiate and sort out: 1≤NY≤ 8, it is noise spot, works as NY=9, it is determined as figure
As pixel.
9. a kind of aircraft automatic obstacle avoiding system for implementing aircraft automatic obstacle avoiding method described in claim 1, which is characterized in that
The aircraft automatic obstacle avoiding system includes:
Solar powered module is connect with central control module, for being converted solar energy into electrical energy by solar panel
It is powered for aircraft;
Image capture module is connect with central control module, for acquiring the ring in front of aircraft flight by optical camera device
Border image;
Complaint message detecting module, connect with central control module, for passing through millimetre-wave radar in the detectable distance in front
Barrier carry out multi-target detection, position, speed and the azimuth information of acquired disturbance object;
Central control module, with solar powered module, image capture module, complaint message detecting module, analyze and determine module,
Directive generation module, path planning module, stability analysis module, display module connection, for passing through central controller controls
Modules work normally;
It analyzes and determines module, is connect with central control module, for being believed by data analysis program according to image, the obstacle of acquisition
Breath analyze and determine the risk of barrier;
Directive generation module is connect with central control module, is become for generating course line according to judging result by instruction generator
More instruct;
Path planning module is connect with central control module, for planning aircraft according to alteration command by navigation system
Flight path;
Stability analysis module, connect with central control module, for being divided by analyzing program the stability of aircraft
Analysis;
Display module is connect with central control module, for showing acquisition ambient image, complaint message, flight by display
Path, stability data.
10. a kind of aircraft for carrying aircraft automatic obstacle avoiding system described in claim 9.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910132801.6A CN109634309B (en) | 2019-02-21 | 2019-02-21 | Autonomous obstacle avoidance system and method for aircraft and aircraft |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910132801.6A CN109634309B (en) | 2019-02-21 | 2019-02-21 | Autonomous obstacle avoidance system and method for aircraft and aircraft |
Publications (2)
Publication Number | Publication Date |
---|---|
CN109634309A true CN109634309A (en) | 2019-04-16 |
CN109634309B CN109634309B (en) | 2024-03-26 |
Family
ID=66065782
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910132801.6A Active CN109634309B (en) | 2019-02-21 | 2019-02-21 | Autonomous obstacle avoidance system and method for aircraft and aircraft |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109634309B (en) |
Cited By (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110299030A (en) * | 2019-06-28 | 2019-10-01 | 汉王科技股份有限公司 | Handheld terminal, aircraft and its airspace measurement method, control method |
CN111175730A (en) * | 2020-01-08 | 2020-05-19 | 中国石油大学(华东) | Millimeter wave radar target trace condensing method for unmanned ship |
CN112069889A (en) * | 2020-07-31 | 2020-12-11 | 北京信息科技大学 | Civil aircraft trajectory prediction method, electronic device and storage medium |
CN112306092A (en) * | 2020-10-29 | 2021-02-02 | 吉林化工学院 | Unmanned aerial vehicle system of patrolling and examining |
CN112356761A (en) * | 2020-09-22 | 2021-02-12 | 成都飞机工业(集团)有限责任公司 | Aircraft transportation method based on whole machine erection |
CN113311857A (en) * | 2021-04-29 | 2021-08-27 | 重庆交通大学 | Environment sensing and obstacle avoidance system and method based on unmanned aerial vehicle |
CN113467504A (en) * | 2021-07-26 | 2021-10-01 | 广东电网有限责任公司 | Aircraft flight stability control method, system, equipment and storage medium |
CN113467495A (en) * | 2021-07-07 | 2021-10-01 | 中国人民解放军火箭军工程大学 | Aircraft rapid trajectory optimization method meeting strict time and position constraints |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104732500A (en) * | 2015-04-10 | 2015-06-24 | 天水师范学院 | Traditional Chinese medicinal material microscopic image noise filtering system and method adopting pulse coupling neural network |
CN107272731A (en) * | 2017-06-05 | 2017-10-20 | 陈金良 | The automatic anti-collision system of unmanned plane |
CN107831777A (en) * | 2017-09-26 | 2018-03-23 | 中国科学院长春光学精密机械与物理研究所 | A kind of aircraft automatic obstacle avoiding system, method and aircraft |
CN108121856A (en) * | 2017-12-06 | 2018-06-05 | 中国科学院力学研究所 | A kind of full flight domain aerocraft dynamic stability analysis method |
WO2019015158A1 (en) * | 2017-07-21 | 2019-01-24 | 歌尔科技有限公司 | Obstacle avoidance method for unmanned aerial vehicle, and unmanned aerial vehicle |
-
2019
- 2019-02-21 CN CN201910132801.6A patent/CN109634309B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104732500A (en) * | 2015-04-10 | 2015-06-24 | 天水师范学院 | Traditional Chinese medicinal material microscopic image noise filtering system and method adopting pulse coupling neural network |
CN107272731A (en) * | 2017-06-05 | 2017-10-20 | 陈金良 | The automatic anti-collision system of unmanned plane |
WO2019015158A1 (en) * | 2017-07-21 | 2019-01-24 | 歌尔科技有限公司 | Obstacle avoidance method for unmanned aerial vehicle, and unmanned aerial vehicle |
CN107831777A (en) * | 2017-09-26 | 2018-03-23 | 中国科学院长春光学精密机械与物理研究所 | A kind of aircraft automatic obstacle avoiding system, method and aircraft |
CN108121856A (en) * | 2017-12-06 | 2018-06-05 | 中国科学院力学研究所 | A kind of full flight domain aerocraft dynamic stability analysis method |
Cited By (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110299030A (en) * | 2019-06-28 | 2019-10-01 | 汉王科技股份有限公司 | Handheld terminal, aircraft and its airspace measurement method, control method |
CN111175730A (en) * | 2020-01-08 | 2020-05-19 | 中国石油大学(华东) | Millimeter wave radar target trace condensing method for unmanned ship |
CN112069889A (en) * | 2020-07-31 | 2020-12-11 | 北京信息科技大学 | Civil aircraft trajectory prediction method, electronic device and storage medium |
CN112356761A (en) * | 2020-09-22 | 2021-02-12 | 成都飞机工业(集团)有限责任公司 | Aircraft transportation method based on whole machine erection |
CN112356761B (en) * | 2020-09-22 | 2022-06-14 | 成都飞机工业(集团)有限责任公司 | Aircraft transportation method based on whole machine erection |
CN112306092A (en) * | 2020-10-29 | 2021-02-02 | 吉林化工学院 | Unmanned aerial vehicle system of patrolling and examining |
CN113311857A (en) * | 2021-04-29 | 2021-08-27 | 重庆交通大学 | Environment sensing and obstacle avoidance system and method based on unmanned aerial vehicle |
CN113467495A (en) * | 2021-07-07 | 2021-10-01 | 中国人民解放军火箭军工程大学 | Aircraft rapid trajectory optimization method meeting strict time and position constraints |
CN113467495B (en) * | 2021-07-07 | 2023-08-11 | 中国人民解放军火箭军工程大学 | Aircraft rapid trajectory optimization method meeting strict time and position constraints |
CN113467504A (en) * | 2021-07-26 | 2021-10-01 | 广东电网有限责任公司 | Aircraft flight stability control method, system, equipment and storage medium |
CN113467504B (en) * | 2021-07-26 | 2023-06-02 | 广东电网有限责任公司 | Method, system, equipment and storage medium for controlling flight stability of aircraft |
Also Published As
Publication number | Publication date |
---|---|
CN109634309B (en) | 2024-03-26 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109634309A (en) | A kind of aircraft automatic obstacle avoiding system, method and aircraft | |
US20220197281A1 (en) | Intelligent decision-making method and system for unmanned surface vehicle | |
CN107909082B (en) | Sonar image target identification method based on deep learning technology | |
CN113567984B (en) | Method and system for detecting artificial small target in SAR image | |
CN111461023B (en) | Method for quadruped robot to automatically follow pilot based on three-dimensional laser radar | |
CN110889324A (en) | Thermal infrared image target identification method based on YOLO V3 terminal-oriented guidance | |
CN110634333B (en) | Landing on emergency or unprepared landing strips in low visibility conditions | |
Xu et al. | A cascade adaboost and CNN algorithm for drogue detection in UAV autonomous aerial refueling | |
CN108256436A (en) | A kind of radar HRRP target identification methods based on joint classification | |
CN109633631A (en) | A kind of multi-functional reconnaissance radar combat duty automatic planning | |
Shi et al. | Objects detection of UAV for anti-UAV based on YOLOv4 | |
CN108647644A (en) | Coal mine based on GMM characterizations blows out unsafe act identification and determination method | |
CN114820702A (en) | Jepsort unmanned aerial vehicle view angle pedestrian multi-target tracking method based on yolov5 | |
CN109241902A (en) | A kind of landslide detection method based on multi-scale feature fusion | |
CN111831010A (en) | Unmanned aerial vehicle obstacle avoidance flight method based on digital space slice | |
CN102298728B (en) | Method for evaluating target threat degree | |
CN115903880A (en) | Unmanned aerial vehicle autonomous image navigation and obstacle avoidance method based on improved reinforcement learning | |
CN113625569A (en) | Small unmanned aerial vehicle prevention and control hybrid decision method and system based on deep reinforcement learning and rule driving | |
CN116907282B (en) | Unmanned target aircraft ultra-low altitude flight control method based on artificial intelligence algorithm | |
CN112556682B (en) | Automatic detection algorithm for underwater composite sensor target | |
Cobano et al. | Thermal detection and generation of collision-free trajectories for cooperative soaring UAVs | |
CN110033490A (en) | A kind of low slow Small object preventing control method in airport based on photoelectric image automatic identification | |
CN113589272A (en) | Automatic generation method for target tracking equipment on-duty log | |
CN112987765A (en) | Precise autonomous take-off and landing method of unmanned aerial vehicle/boat simulating attention distribution of prey birds | |
Liu et al. | Object detection algorithm based on lightweight YOLOv4 for UAV |
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 |