CN108168558A - Unmanned aerial vehicle flight path planning algorithm applied to river target search task - Google Patents
Unmanned aerial vehicle flight path planning algorithm applied to river target search task Download PDFInfo
- Publication number
- CN108168558A CN108168558A CN201711403481.0A CN201711403481A CN108168558A CN 108168558 A CN108168558 A CN 108168558A CN 201711403481 A CN201711403481 A CN 201711403481A CN 108168558 A CN108168558 A CN 108168558A
- Authority
- CN
- China
- Prior art keywords
- search
- curved section
- high value
- unmanned plane
- river
- 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
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
- G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
- G01C21/20—Instruments for performing navigational calculations
Landscapes
- Engineering & Computer Science (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Automation & Control Theory (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Control Of Position, Course, Altitude, Or Attitude Of Moving Bodies (AREA)
Abstract
Unmanned aerial vehicle flight path planning algorithm applied to river target search task, target river region abstract modeling is two-dimensional curve by the present invention, target is in each curved section known to existing probability, search high value curved section is extracted using gauss hybrid models, that is high value region of search, and the search order of high value curved section is ranked up, the unmanned plane for performing search mission is distributed to, so that unmanned plane carries out target search by the curved section after sequence.This method can obtain near-optimization result and solving speed is very fast, improve the target search efficiency of unmanned plane.
Description
Technical field
The present invention relates to Navigation of Pilotless Aircraft and control technology field, are related to a kind of unmanned aerial vehicle flight path planning algorithm, specially
A kind of unmanned aerial vehicle flight path planning algorithm applied to river target search task.
Background technology
Unmanned plane (Unmanned Aerial Vehicle, abbreviation UAV) refer to by it is unmanned on power drive, machine, by
Wireless remotecontrol manipulation or a kind of aircraft of self-contained program's control.The production merged as aeronautical technology with information technology height
Object, unmanned plane because its is cost-effective, using flexible, executable high risk task, the advantages such as is not limited by pilot's physiological condition,
It is widely applied military with civil field.Over nearest more than 30 years, countries in the world give more sustained attention and add to unmanned plane field
Big input, unmanned air vehicle technique achieve tremendous development and progress, represent the developing direction of current new and high technology.
One of the typical case of target search as unmanned plane is widely used to accident and meets an urgent need monitoring or rescue etc.
Scene.Since the survival probability for being searched target (such as wander away personnel, shipwreck) will reduce rapidly as time go on,
It is required that unmanned plane finds target in the shortest possible time.The above problem can substantially regard track optimization problem as so that nothing
Maximum search return rate (i.e. accumulated probability) can be obtained during the man-machine flight along optimal trajectory, realizes the efficient investigation to mission area
Covering and target fast search.
Domestic and foreign scholars have carried out numerous studies for target search problem and have proposed serial of methods, mainly including geometry
Method, Monte Carlo analysis, method based on search graph etc..Geometric method is by planning the acquisition track such as parallel lines, spiral shell of specific shape
Spin line etc. realizes traversal or all standing of the unmanned plane to mission area, although such Method And Principle is simple, believes in target priori
Its search efficiency is significant lower in the case of breath is known.Monte Carlo analysis guides unmanned plane random motion in mission area, from
And gradually overlay area and target being searched, the sharpest edges of such method are not need to accurately position to determine with the optimization of complexity
Plan process, but equally not using target prior information and it is not suitable for a wide range of complex region.Method based on search graph is first
A series of grid cells are turned to by mission area is discrete, the target information of each unit storage is then based on, takes suitable optimization
Strategy causes unmanned plane to be moved to most promising direction, although such method can flexibly handle all kinds of complex situations, nobody
Machine may hover for a long time ignores other high value regions in regional area, and search efficiency has to be hoisted.In addition, to solve
Above-mentioned local optimum problem simultaneously simplifies multiple no-manned plane collaboration target search task, and a kind of effective thinking is to decompose mission area
For multiple subregions and each unmanned plane is distributed to, so as to which complicated Collaborative Control problem is converted into multiple simple single unmanned planes
Search problem, main Domain Decomposition Method include barycenter Voronoi Division Samplings, fuzzy C-means clustering, polygon segmentation
Deng.
Existing target search research is applied to two-dimensional level region (such as rectangle regular domain or irregular area) mostly,
It is and less for the target search Study on Problems in river region.As a kind of special mission area, river can regard band one as
Determine the curve of topographic constraints, therefore selected compared to the even full course in more courses of the unmanned plane in ordinary two dimensional horizontal zone,
River region limits the offline mode of unmanned plane, increases the difficulty of target search problem.For such problem, existing solution
Scheme is mostly qualitatively passive tactful, as unmanned plane carries out the search of all standing formula in river flying overhead or directly rushes towards target
Previous location area carry out Greedy search, lack quantitative analysis and heuristic strategies and instruct.
Invention content
It is an object of the invention to combine the natural conditions feature in river, propose a kind of suitable for river target search task
Unmanned aerial vehicle flight path planning algorithm, to realize the fast search of river goal task.
In order to achieve the goal above, the present invention provides following technical solution：
Unmanned aerial vehicle flight path planning algorithm applied to river target search task, which is characterized in that the unmanned plane regards
Field range is more than flow width, and the planning algorithm includes the following steps：
S1：Target river to be searched is abstracted as two dimensional surface curve, with to front distance LsFor independent variable, to Plane Curved
Line does sliding-model control as M=L/LsA discretization unit, wherein L be river total length, LsLength for each discretization unit
Degree；
S2：In any one sampling period, unmanned plane is in two discretization unit smInterior movement；Target to be searched is each
Discretization unit smInterior existing Probability p (sm) it is known that and p (sm) ∈ [0,1], and meet
S3：Arbitrary discretization unit composition search curved section, Sk={ sm,sm+1,...,sn, 1≤m ＜ n≤M are utilized
Gauss hybrid models describe target to be searched existing probabilistic information in each search curved section, extract with Pk,1And Pk,2For area
The high value search curved section S on domain boundaryk；
S4：It is current search curved section boundary point P to define subregion transition timesi,pGo to next search curve segment boundary
Point Pj,qThe most short flight time, defining the transition times is：
T(Pi,p,Pj,q)=λ1·Dubins_cost(Pi,p,Pj,q), wherein i ≠ j and i, j ∈ 1,2 ... K }, and p, q
∈{1,2}；Wherein λ1Represent the proportionality coefficient of subregion transition times；Dubins_cost(Pi,p,Pj,q) represent Dubins curves
Length, p represent i-th search curved section the flag bit that flies out, q represent j-th search curved section inlet point flag bit,
K represents the quantity of high value search curved section；
Time of the subregion cover time for unmanned plane scouting flight in search curved section is defined, is Ck=λ2·Lk,
Middle λ2Represent the proportionality coefficient of subregion cover time, LkRepresent the length of search curved section, Lk=(n-m+1) Ls；
It defines subregion covering and returns the accumulated probability that high value search curved section is completely covered and is obtained by unmanned plane：
Rk；
S5：Evaluation index is used as using the covering return of above-mentioned subregion transition times, subregion cover time and subregion,
Nearest insertion algorithm is taken to scan for the iterative sequencing of sequence to each high value search curved section, sequence is modeled as：
S6：It is sequentially connected the high value search curved section after optimization in order, obtains the final searching route of unmanned plane.
As preferred：The method of extraction high value search curved section includes the following steps：
Using K one-dimensional Gaussian functionsForm gauss hybrid models, wherein k=
1,2 ... K, ukFor the mean value of each Gaussian function, σkStandard deviation for each Gaussian function；
If the weight coefficient shared by each Gaussian function is αk, and meet
Destination probability by curved section to be searched is
Estimate the weight coefficient α of each Gaussian functionk, mean value uk, standard deviation sigmak, estimation is iterated to p (s), until full
The sufficient condition of convergence；
The correspondence section of each 95.4% probability of Gaussian function is got, with Pk,1=uk-2σkAnd Pk,2=uk+2σkFor region
The high value of boundary point searches for curved section, then Lk=4 σk；Rk=0.954 αk。
As preferred：The condition of convergence is | p (s)-p'(s) | ＜ ξ, wherein p (s) and p'(s) be respectively iteration before and after
Destination probability value, ξ=10-5。
As preferred：If multiple UAVs perform search mission, further comprise the steps：To the high price after sequence
Value search curved section distributes to multiple UAVs, carries out region distribution.
As preferred：The method of region distribution, includes the following steps：
High value search curved section set is distributed into every frame unmanned plane, obtaining a collection for every frame unmanned plane is combined into Ai's
Distribute high value search curved section set；
By the method in step S5, to every frame unmanned plane set AiInterior high value is searched for curved section and is calculated into row index,
Obtain Ji；
Region distribution is modeled asWherein, NuRepresent the total of unmanned plane
Number, ρ1Represent the proportionality coefficient of total search income, ρ2The proportionality coefficient of the expression task degree of balance；
Using alternative manner, model solution is distributed to region, determines region allocation strategy.
As preferred：The high value search curved section set that all unmanned planes are distributed meets following constraints：
Beneficial effects of the present invention are：
(1) target river region abstract modeling is two-dimensional curve by the present invention, greatly reduces problem complexity with calculating
Amount；
(2) present invention come approximate description river region feature and extracts high value sub-district using adaptive GMM
Domain can quantify to extract the correspondence section of 95.4% probability of Gaussian function, targetedly high value subregion be scanned for appointing
Business distribution, the precision of quantized result is higher and is conducive to the solution of contingency question；
(3) present invention carries out region ordering using nearest insertion algorithm, and this method can obtain near-optimization result and solve speed
Degree is very fast；
(4) algorithm provided by the invention is suitable for the planning of a frame or multiple UAVs searching route.
Description of the drawings
Fig. 1 is flow chart of the method for the present invention；
Fig. 2 is that river is abstracted as two-dimensional curve schematic diagram；
Fig. 3 is the destination probability figure in river region；
Fig. 4 is approximation of the destination probability based on adaptive GMM；
Fig. 5 is the subregion ordering iteration process based on nearest insertion algorithm；
Fig. 6 be with three frame unmanned planes carry out target search using the method for the present invention optimization come unmanned aerial vehicle flight path.
Specific embodiment
The specific embodiment of the present invention is clearly fully described by below with reference to attached drawing.Obviously, it is embodied
Mode described embodiment is only the part of the embodiment of the present invention, instead of all the embodiments.Based in the present invention
Embodiment, those of ordinary skill in the art's all other embodiments obtained under the premise of creative work is not made,
Belong to protection scope of the present invention.
The present invention provides a kind of algorithms for being planned the search flight path of unmanned plane, specific to work as nothing for one kind
It is man-machine be used for river target search task when unmanned aerial vehicle flight path planning algorithm.Using the planning algorithm, can greatly improve
The search efficiency of unmanned plane.
In order to improve the search efficiency of unmanned plane, the sight of the visual sensor of UAV flight is directed toward underface always,
And its field range is more than the width in river, so, it is ensured that unmanned plane may search for entire during advancing forward
The range of river width.The overall flow figure of method is with reference to figure 1.
The present embodiment provides a kind of planning algorithm for being suitable for a frame unmanned plane and performing the search flight path of search mission first.
Applied to the unmanned aerial vehicle flight path planning algorithm of river target search task, include the following steps：
S1：Target river is abstracted as two dimensional surface curve and carries out sliding-model control.
With reference to figure 2, since the search wide viewing angle of unmanned plane is more than the width in river, the width in river can be ignored
Degree, two dimensional surface curve is abstracted as by target river to be searched.With to front distance LsFor independent variable, plane curve is done discrete
Change processing as M=L/LsA discretization unit, wherein L be target river total length, LsLength for each discretization unit；Into
After row sliding-model control, the discretization unit in M river will be obtained.
S2：The probabilistic information of each discretization unit in structure unit's sampling period.
In the unit sampling period, the search range of unmanned plane is the length of the discretization unit in above-mentioned river.Specifically,
In any one sampling period, unmanned plane is in two discretization unit smInterior movement；Target to be searched is in each discretization unit sm
Interior existing Probability p (sm) it is known that and p (sm) ∈ [0,1], and meetProbability graph is with reference to figure 3.
S3：Extract high value search curved section.
Due to target, existing probability is different in each search curved section, has height to have low, in order to improve search efficiency, needs
The curved section for wanting selection target existing probability high scans for, and therefore, it is necessary to extract high value search curved section.
Arbitrary discretization unit composition search curved section, Sk={ sm,sm+1,...,sn, 1≤m ＜ n≤M utilize Gauss
Mixed model describes target to be searched existing probabilistic information in each search curved section, extracts with Pk,1And Pk,2For regional edge
The high value search curved section S on boundaryk；The quantity of wherein high value search curved section is K, k ∈ { 1,2 ..., K }.
Specific method is：Using K one-dimensional Gaussian functions：
Composition gauss hybrid models, wherein k=1,2 ... K, ukFor the mean value of each Gaussian function, σkFor each Gaussian function
Several standard deviations；
If the weight coefficient shared by each Gaussian function is αk, and meet
Destination probability by curved section to be searched is
Estimate the weight coefficient α of each Gaussian functionk, mean value uk, standard deviation sigmak, define one and be made of D cell position
Training sample, estimation is iterated p (s) using expectation maximization, until meet the condition of convergence；The condition of convergence is p
(s)-p'(s) ＜ ξ, wherein p (s) and p'(s) it is respectively destination probability value before and after iteration, ξ values are needed regarding specific convergence precision
Depending on asking, ξ=10 in the present embodiment-5.Iteration result is with reference to figure 4.
In above-mentioned parameter iterative estimation procedure, the corresponding training individuals quantity accounting of each cell position and destination probability
Unanimously, i.e. Dm=p (sm)D；Elimination, merging and the dynamic mechanisms such as division can be introduced and adaptively adjust model quantity.If certain Gaussian mode
The weight very little of type and there is certain distance with other models, illustrate that the component for unwanted noise component, can be eliminated directly；If two
Gauss model is little apart from close and weight, then it is assumed that they reflect identical feature distribution, over-fitting, therefore it occur
Can merge into a Gaussian component；If the weight of certain Gaussian component, standard deviation are larger, illustrate poor fitting phenomenon occur, need
It is split into two Gaussian components.
According to parameter estimation result, you can quantization extracts the correspondence section of each 95.4% probability of Gaussian function, as height
The river subregion (i.e. high value search curved section) of value, gets with Pk,1=uk-2σkAnd Pk,2=uk+2σkFor regional edge
The length L of the high value curved section of boundary's point, then high value search curved sectionk=4 σk。
S4：Definition quantization evaluation index carries out grade assessment and row using nearest insertion algorithm to high value search curved section
Sequence determines search order of the unmanned plane to each high value search curved section.The following three kinds of evaluation index sub-districts of the present embodiment definition
Domain transition times, subregion cover time and subregion covering return.
It is current search curved section boundary point P to define subregion transition timesi,p(or search initial position Pinitial) go to
Next search curved section boundary point Pj,qThe most short flight time, due to finally obtain multiple high values search curved section be phase
It is discontinuous point to the curved section of separation, between the boundary point of each curved section, and what subregion transition times represented as exists
The time moved between mutual discontinuous curved section.Usual Dubins curves are the nothing between any two points for considering course constraint
Man-machine Minimal Flight Path, defining the transition times is：
T(Pi,p,Pj,q)=λ1·Dubins_cost(Pi,p,Pj,q), wherein i ≠ j and i, j ∈ 1,2 ... K }, and p, q
∈{1,2}；Wherein λ1Represent the proportionality coefficient of subregion transition times；Dubins_cost(Pi,p,Pj,q) represent Dubins curves
Length, p represent i-th search curved section the flag bit that flies out, q represent j-th search curved section inlet point flag bit,
K represents the quantity of high value search curved section；
Time of the subregion cover time for unmanned plane scouting flight in search curved section is defined, is Ck=λ2·Lk,
Middle λ2Represent the proportionality coefficient of subregion cover time, LkIt represents the length of search curved section, is Lk=(n-m+1) Ls；Due to
The length of high value curved section that Gauss model extracts is used as Lk=4 σk, therefore, Ck=λ2·4σk。
λ1And λ2Relationship be inversely proportional with flying speed relationship of the unmanned plane in different phase, for example, unmanned plane is in subregion
Between the flying speed of transition be 2 times of flying speed when it covers subregion (i.e. in subregion overhead), then above-mentioned parameter is answered
Meet
It defines subregion covering and returns the accumulated probability R that high value search curved section is completely covered and is obtained by unmanned planek,
The probability searches for the sum of curved section target existing probability for each high value；And relative to the high value by Gauss model extraction
Curved section, since river subregion is the corresponding interval range of each 95.4% probability of Gaussian function, and weight coefficient is αk, therefore
Subregion covering return is defined as follows：Rk=0.954 αk。
S5：The search order of each high value search curved section is ranked up, to realize optimal efficient search strategy.
Using the covering return of above-mentioned subregion transition times, subregion cover time and subregion as evaluation index, adopt
Nearest insertion algorithm is taken to each high value search curved section LpThe iterative sequencing of sequence is scanned for, sequence is modeled as：
Wherein, PinitialTo search for initial position,First high value curve for starting to search for is represented respectively
Two endpoints of section,Represent that the subregion covering for starting first high value curved section of search is returned,Expression starts to search
The subregion cover time of first high value curved section of rope.Correspondingly,It represents respectively
The corresponding index of kth high value curved section.
Herein it should be noted that two boundary point flag bits of subregion are respectively 1 and 2, theirs and for 3.1 and 2
Do not refer to number, but digit synbol, for example, two boundary points of the 5th sub-regions are P5,1, P5,2.Since unmanned plane is to kth -1
The inlet point flag bit of sub-regions is nk-1, then the flag bit that flies out to the unmanned plane after the sub-district domain search is 3-
nk-1。
Subregion sequencing problem can regard typical traveling salesman problem as, therefore nearest insertion algorithm is taken to obtain near-optimization
Ranking results.In each iterative process, from remaining unsorted subregion setIn arbitrarily choose certain sub-regions
And insert it into the subregion sequence { l that sorted1,...,lIAny position, and calculate new sort area by order models formula
The index value J (I+1) in domain, and then increment Delta J=J (I+1)-J (I) of parameter value, are selected from above-mentioned be possible to situation
Situation when Δ J is maximized is selected, then certain the unsorted subregion selected is inserted into the subregion sequence that sorted.It is above-mentioned
Iterative process repeats K steps, that is, completes the approximate sorting of K sub-regions.
Fig. 5 is given by searching for initial position Pinitial, with respect to three sub-regions (P1,1,P1,2)、(P2,1,P2,2) and (P3,1,
P3,2), using nearest insertion algorithm to the process of 3 sub-regions approximate sortings, wherein solid line represents river subregion, and dotted line represents
All possible transition flight path section, solid line represent determining transition flight path section.As can be seen that from initial position PinitialIt is possible
Transition path includes (Pinitial,P1,1)、(Pinitial,P1,2)、(Pinitial,P2,1)、(Pinitial,P2,2)、(Pinitial,P3,1)、
(Pinitial,P3,2)；After 1 iterative calculation, (P is determinedinitial,P1,1) it is optimal path, and proceed by road next time
Diameter iterates to calculate, and possible transition path is included as shown in figure (c), after iterative calculation, determines (P1,2,P2,1) it is optimal road
Diameter；So repeatedly, the final search path schemed shown in (f).
After sequence, it may be determined that the grade (i.e. the search order of subregion) of each river subregion and each height
The flag bit (entering subregion from which of two endpoints of every sub-regions endpoint) of region inlet point, unmanned plane will
Enter subregion by flag bit, scanned for one by one by the sequence of subregion.
Above searching algorithm is suitable for the acquisition track planning that a frame unmanned plane performs search mission.
And in the prior art, in order to realize efficient search, usual multiple UAVs perform the task of search jointly.Cause
This, the present embodiment is further provided in the case that a kind of multiple UAVs perform search mission jointly, unmanned plane acquisition track rule
The algorithm drawn.
When there are multiple UAVs to perform search mission jointly, need to consider the balance of total search income and task
Degree, each frame unmanned plane is distributed to by subregion.
Specifically, if multiple UAVs perform search mission, further comprise the steps：To the value after sequence
Search curved section distributes to multiple UAVs, carries out region distribution.
The method of region distribution, includes the following steps：
High value search curved section is distributed into every frame unmanned plane, every frame unmanned plane obtains a collection and is combined into AiDistribution it is high
Value search curved section set, curved section set is interior to include a plurality of high value search curved section；
In order to avoid colliding between unmanned plane, it is desirable that a frame unmanned plane is only assigned to per sub-regions, and in order to protect
Demonstrate,prove the efficient utilization of unmanned plane, it is desirable that for every frame unmanned plane at least distribute a sub-regions, therefore all unmanned planes are distributed
High value search curved section set meets following constraints：
Wherein, NuRepresent the quantity of unmanned plane, K represents the quantity of high value search curved section.
To the process that multiple UAVs searching route is allocated with reference to the method in step S4, S5, to every frame unmanned plane
The curved section set A of acquisitioniInterior high value is searched for curved section and is calculated into row index (including subregion transition times, subregion
Cover time and subregion covering return), obtain Ji；
Region distribution is modeled asWherein, NuRepresent the total of unmanned plane
Number, ρ1Represent the proportionality coefficient of total search income, ρ2The proportionality coefficient of the expression task degree of balance；ρ1And ρ2Value need to be according to tool
Body mission requirements carry out assignment, such as when each machine detectivity is close, it is possible to increase ρ2Value；Each machine detectivity difference is larger
When, it is possible to increase ρ1Value.
Using alternative manner, model solution is distributed to region, determines region allocation strategy, as shown in Figure 6.Final optimization
Flight path is represented by：
The planning of unmanned aerial vehicle flight path is carried out, it can be achieved that the effective search of target river section using algorithm of the present invention.
Claims (6)
1. the unmanned aerial vehicle flight path planning algorithm applied to river target search task, which is characterized in that the visual field of the unmanned plane
Range is more than flow width, and the planning algorithm includes the following steps：
S1：Target river to be searched is abstracted as two dimensional surface curve, with to front distance LsFor independent variable, plane curve is done
Sliding-model control is M=L/LsA discretization unit, wherein L are river total length sm, LsLength for each discretization unit；
S2：In any one sampling period, unmanned plane is in two discretization unit smInterior movement；Target to be searched is each discrete
Change unit smInterior existing Probability p (sm) it is known that and p (sm) ∈ [0,1], and meet
S3：Arbitrary discretization unit composition search curved section, Sk={ sm,sm+1,...,sn, 1≤m ＜ n≤M utilize Gauss
Mixed model describes target to be searched existing probabilistic information in each search curved section, extracts with Pk,1And Pk,2For regional edge
The high value search curved section S on boundaryk, the quantity of wherein high value search curved section is K, k ∈ { 1,2 ..., K }；
S4：It is current search curved section boundary point P to define subregion transition timesi,pGo to next search curved section boundary point Pj,q
The most short flight time, defining the transition times is：
T(Pi,p,Pj,q)=λ1·Dubins_cost(Pi,p,Pj,q), wherein i ≠ j and i, j ∈ 1,2 ... K }, and p, q ∈ 1,
2}；Wherein λ1Represent the proportionality coefficient of subregion transition times；Dubins_cost(Pi,p,Pj,q) represent Dubins curves length
Degree, p represent the flag bit that flies out of i-th of search curved section, and q represents the inlet point flag bit of j-th of search curved section, K tables
Show the quantity of high value search curved section；
Time of the subregion cover time for unmanned plane scouting flight in search curved section is defined, is Ck=λ2·Lk, wherein λ2
Represent the proportionality coefficient of subregion cover time, LkRepresent the length of search curved section, Lk=(n-m+1) Ls；
It defines subregion covering and returns the accumulated probability that high value search curved section is completely covered and is obtained by unmanned plane：Rk；
S5：Using the covering return of above-mentioned subregion transition times, subregion cover time and subregion as evaluation index, take
Nearest insertion algorithm scans for each high value search curved section the iterative sequencing of sequence, and sequence is modeled as：
S6：It is sequentially connected the high value search curved section after optimization in order, obtains the final searching route of unmanned plane.
2. it is applied to the unmanned aerial vehicle flight path planning algorithm of river target search task as described in claim 1, which is characterized in that
The method of extraction high value search curved section includes the following steps：
Using K one-dimensional Gaussian functionsComposition gauss hybrid models, wherein k=1,
2 ... K, ukFor the mean value of each Gaussian function, σkStandard deviation for each Gaussian function；
If the weight coefficient shared by each Gaussian function is αk, and meet
Destination probability by curved section to be searched is
Estimate the weight coefficient α of each Gaussian functionk, mean value uk, standard deviation sigmak, estimation is iterated to p (s), is received until meeting
Hold back condition；
The correspondence section of each 95.4% probability of Gaussian function is got, with Pk,1=uk-2σkAnd Pk,2=uk+2σkFor zone boundary point
High value search curved section, then Lk=4 σk；Rk=0.954 αk。
3. it is applied to the unmanned aerial vehicle flight path planning algorithm of river target search task as claimed in claim 2, which is characterized in that
The condition of convergence is | p (s)-p'(s) | ＜ ξ, wherein p (s) and p'(s) it is respectively destination probability value before and after iteration, the ξ
=10-5。
4. it is applied to the unmanned aerial vehicle flight path planning algorithm of river target search task as described in claim 1, which is characterized in that
If multiple UAVs perform search mission, further comprise the steps：Value search curved section after sequence is distributed to
Multiple unmanned planes carry out region distribution.
5. it is applied to the unmanned aerial vehicle flight path planning algorithm of river target search task as claimed in claim 4, which is characterized in that
The method of region distribution, includes the following steps：
High value search curved section is distributed into every frame unmanned plane, every frame unmanned plane obtains a collection and is combined into AiDistribution high value
Search for curved section set；
By the method in step S5, to every frame unmanned plane set AiInterior high value is searched for curved section and is calculated into row index, obtains
Ji；
Region distribution is modeled asWherein, NuRepresent the sum of unmanned plane, ρ1
Represent the proportionality coefficient of total search income, ρ2The proportionality coefficient of the expression task degree of balance；
Using alternative manner, model solution is distributed to region, determines region allocation strategy.
6. it is applied to the unmanned aerial vehicle flight path planning algorithm of river target search task as claimed in claim 5, which is characterized in that
The high value search curved section set that all unmanned planes are distributed meets following constraints：
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711403481.0A CN108168558B (en) | 2017-12-22 | 2017-12-22 | Unmanned aerial vehicle track planning algorithm applied to river target search task |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711403481.0A CN108168558B (en) | 2017-12-22 | 2017-12-22 | Unmanned aerial vehicle track planning algorithm applied to river target search task |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108168558A true CN108168558A (en) | 2018-06-15 |
| CN108168558B CN108168558B (en) | 2020-04-10 |
Family
ID=62523264
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711403481.0A Active CN108168558B (en) | 2017-12-22 | 2017-12-22 | Unmanned aerial vehicle track planning algorithm applied to river target search task |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108168558B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108955645A (en) * | 2018-07-16 | 2018-12-07 | 福州日兆信息科技有限公司 | Three-dimensional modeling method and device applied to communication iron tower intelligent patrol detection |
| CN111487986A (en) * | 2020-05-15 | 2020-08-04 | 中国海洋大学 | Underwater robot cooperative target searching method based on global information transfer mechanism |
| CN111596675A (en) * | 2020-05-15 | 2020-08-28 | 中国海洋大学 | Underwater robot optimization decision-making method facing non-wide area target search task |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2011057323A1 (en) * | 2009-11-10 | 2011-05-19 | Bae Systems Australia Limited | Method and system to aid craft movement prediction |
| CN103472850A (en) * | 2013-09-29 | 2013-12-25 | 合肥工业大学 | Multi-unmanned aerial vehicle collaborative search method based on Gaussian distribution prediction |
| US8825383B1 (en) * | 2010-12-10 | 2014-09-02 | Google Inc. | Extracting patterns from location history |
| CN104536454A (en) * | 2014-12-05 | 2015-04-22 | 中国运载火箭技术研究院 | Space-time synchronization matching method used for double unmanned aerial vehicle cooperation |
| CN104596516A (en) * | 2014-11-24 | 2015-05-06 | 中国海洋大学 | Unmanned aerial vehicle coverage flight path planning based on dynamic newly-added adjacent area |
| CN104867134A (en) * | 2015-05-04 | 2015-08-26 | 国家电网公司 | Identification method for transmission line tower inspection by unmanned aerial vehicle |
| CN106406346A (en) * | 2016-11-01 | 2017-02-15 | 北京理工大学 | Plan method for rapid coverage track search coordinated by multiple UAVs (Unmanned Aerial Vehicles) |
-
2017
- 2017-12-22 CN CN201711403481.0A patent/CN108168558B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2011057323A1 (en) * | 2009-11-10 | 2011-05-19 | Bae Systems Australia Limited | Method and system to aid craft movement prediction |
| US8825383B1 (en) * | 2010-12-10 | 2014-09-02 | Google Inc. | Extracting patterns from location history |
| CN103472850A (en) * | 2013-09-29 | 2013-12-25 | 合肥工业大学 | Multi-unmanned aerial vehicle collaborative search method based on Gaussian distribution prediction |
| CN104596516A (en) * | 2014-11-24 | 2015-05-06 | 中国海洋大学 | Unmanned aerial vehicle coverage flight path planning based on dynamic newly-added adjacent area |
| CN104536454A (en) * | 2014-12-05 | 2015-04-22 | 中国运载火箭技术研究院 | Space-time synchronization matching method used for double unmanned aerial vehicle cooperation |
| CN104867134A (en) * | 2015-05-04 | 2015-08-26 | 国家电网公司 | Identification method for transmission line tower inspection by unmanned aerial vehicle |
| CN106406346A (en) * | 2016-11-01 | 2017-02-15 | 北京理工大学 | Plan method for rapid coverage track search coordinated by multiple UAVs (Unmanned Aerial Vehicles) |
Non-Patent Citations (5)
| Title |
|---|
| SIVAKUMAR RATHINAM等: "Autonomous Searching and Tracking of a River using an UAV", 《2007 AMERICAN CONTROL CONFERENCE》 * |
| YAO PENG等: "UAV feasible path planning based on disturbed fluid and trajectory propagation", 《CHINESE JOURNAL OF AERONAUTICS》 * |
| 丁家如: "多无人机任务分配与路径规划算法研究", 《中国优秀硕士学位论文全文数据库工程科技Ⅱ辑》 * |
| 简康: "无人机航迹规划算法研究", 《中国优秀硕士学位论文全文数据库工程科技Ⅱ辑》 * |
| 霍霄华: "多UCAV动态协同任务规划建模与滚动优化方法研究", 《中国博士学位论文全文数据库信息科技辑》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108955645A (en) * | 2018-07-16 | 2018-12-07 | 福州日兆信息科技有限公司 | Three-dimensional modeling method and device applied to communication iron tower intelligent patrol detection |
| CN111487986A (en) * | 2020-05-15 | 2020-08-04 | 中国海洋大学 | Underwater robot cooperative target searching method based on global information transfer mechanism |
| CN111596675A (en) * | 2020-05-15 | 2020-08-28 | 中国海洋大学 | Underwater robot optimization decision-making method facing non-wide area target search task |
| CN111596675B (en) * | 2020-05-15 | 2021-06-15 | 中国海洋大学 | Underwater robot optimization decision-making method facing non-wide area target search task |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108168558B (en) | 2020-04-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110930770B (en) | Four-dimensional track prediction method based on control intention and airplane performance model | |
| CN111598473B (en) | Multi-platform combined task planning method for complex observation task | |
| CN108229719B (en) | Multi-objective optimization method and device for unmanned aerial vehicle formation task allocation and flight path planning | |
| CN107622327B (en) | Multi-unmanned aerial vehicle flight path planning method based on culture ant colony search mechanism | |
| US8260485B1 (en) | Adaptive multi-vehicle area coverage optimization system and method | |
| CN110428111A (en) | Multi-Tasking method for planning track when UAV/UGV collaboration is long | |
| CN107807665B (en) | Unmanned aerial vehicle formation detection task cooperative allocation method and device | |
| CN105957076A (en) | Clustering based point cloud segmentation method and system | |
| CN108168558A (en) | Unmanned aerial vehicle flight path planning algorithm applied to river target search task | |
| CN105953800B (en) | A kind of unmanned vehicle trajectory planning grid space division methods | |
| CN106778856A (en) | A kind of object identification method and device | |
| CN110426039A (en) | Multi-unmanned aerial vehicle path planning method based on task completion time minimization | |
| CN112116830A (en) | Unmanned aerial vehicle dynamic geo-fence planning method based on airspace meshing | |
| CN110262563A (en) | Multiple no-manned plane collaboratively searching mesh calibration method waterborne | |
| CN102324041B (en) | Pixel classification method, joint body gesture recognition method and mouse instruction generating method | |
| CN107330422A (en) | A kind of method for carrying out mima type microrelief classification to semiarid zone based on high accuracy number elevation model | |
| CN109459929A (en) | The parsing Homotopy Method that martian atmosphere approach section longitudinal direction accessoble region generates | |
| CN112733251B (en) | Collaborative flight path planning method for multiple unmanned aerial vehicles | |
| CN106875755A (en) | A kind of air traffic conflict management method and device based on complexity | |
| CN116257082B (en) | Distributed active cooperative detection method for multiple unmanned aerial vehicles | |
| CN110986957A (en) | Three-dimensional flight path planning method and device for unmanned aerial vehicle | |
| CN104729509B (en) | A kind of path planning method based on non-dominated sorted genetic algorithm II | |
| CN111199243B (en) | Aerial target identification method and system based on improved decision tree | |
| CN115454132A (en) | Unmanned aerial vehicle inspection three-dimensional path planning method and system for mountain photovoltaic power station | |
| CN117556979A (en) | Unmanned plane platform and load integrated design method based on group intelligent search |
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 |