CN108168558B - Unmanned aerial vehicle track planning algorithm applied to river target search task - Google Patents

Unmanned aerial vehicle track planning algorithm applied to river target search task Download PDF

Info

Publication number
CN108168558B
CN108168558B CN201711403481.0A CN201711403481A CN108168558B CN 108168558 B CN108168558 B CN 108168558B CN 201711403481 A CN201711403481 A CN 201711403481A CN 108168558 B CN108168558 B CN 108168558B
Authority
CN
China
Prior art keywords
search
unmanned aerial
aerial vehicle
region
curve segment
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN201711403481.0A
Other languages
Chinese (zh)
Other versions
CN108168558A (en
Inventor
姚鹏
王琨
解则晓
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Ocean University of China
Original Assignee
Ocean University of China
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Ocean University of China filed Critical Ocean University of China
Priority to CN201711403481.0A priority Critical patent/CN108168558B/en
Publication of CN108168558A publication Critical patent/CN108168558A/en
Application granted granted Critical
Publication of CN108168558B publication Critical patent/CN108168558B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

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

The invention relates to an unmanned aerial vehicle track planning algorithm applied to a river target search task, which is characterized in that a target river region is abstractly modeled into a two-dimensional curve, under the condition that the probability of a target in each curve segment is known, a Gaussian mixture model is adopted to extract and search a high-value curve segment, namely a high-value search region, the search sequence of the high-value curve segment is sequenced and distributed to an unmanned aerial vehicle executing the search task, so that the unmanned aerial vehicle searches the target according to the sequenced curve segments. The method can obtain an approximate optimal result, has high solving speed and improves the target searching efficiency of the unmanned aerial vehicle.

Description

Unmanned aerial vehicle track planning algorithm applied to river target search task
Technical Field
The invention relates to the technical field of unmanned aerial vehicle navigation and control, in particular to an unmanned aerial vehicle track planning algorithm applied to a river target search task.
Background
Unmanned Aerial Vehicle (UAV) refers to a type of aircraft that is powered, Unmanned on board, remotely controlled by radio, or self-contained. As a product of high integration of aviation technology and information technology, the unmanned aerial vehicle has been widely applied to military and civil fields due to the advantages of high cost performance, flexible use, capability of executing high-risk tasks, no limitation of pilot physiological conditions and the like. In more than 30 years recently, all countries in the world pay continuous attention to the field of unmanned aerial vehicles and increase investment, and the unmanned aerial vehicle technology has been developed and advanced enough to represent the development direction of the current high and new technology.
As one of typical applications of an unmanned aerial vehicle, target search has been widely applied to emergency monitoring or rescue and other scenes. Since the probability of survival of the searched target (e.g. lost person, sunken ship, etc.) will decrease rapidly as time passes, it is required that the drone find the target in as short a time as possible. The problem can be essentially regarded as a flight path optimization problem, so that the maximum search return rate (namely, cumulative probability) can be obtained when the unmanned aerial vehicle flies along the optimal flight path, and efficient investigation coverage and target quick search of a task area are realized.
Scholars at home and abroad carry out a great deal of research on the target search problem and propose a series of methods, mainly comprising a geometric method, a random search method, a method based on a search map and the like. The geometric method realizes the traversal or full coverage of the unmanned aerial vehicle on the task area by planning the search tracks with specific shapes such as parallel lines, spiral lines and the like, and although the principle of the method is simple, the search efficiency is obviously lower under the condition that the target prior information is known. The random search method guides the unmanned aerial vehicle to move randomly in the task area, so that the area is covered gradually and the target is searched. The method based on the search graph firstly discretizes a task area into a series of grid units, and then adopts a proper optimization strategy to enable the unmanned aerial vehicle to move towards the most promising direction based on target information stored in each unit. In addition, in order to solve the above-mentioned local optimum problem and simplify the multi-drone cooperative target search task, an effective idea is to decompose the task area into a plurality of sub-areas and allocate each sub-area to each drone, thereby converting the complex cooperative control problem into a plurality of simple single drone search problems, and the main area decomposition method includes centroid Voronoi segmentation sampling, fuzzy C-means clustering, polygon segmentation, and the like.
Most of the existing target search researches are applied to a two-dimensional horizontal area (such as a regular area or an irregular area like a rectangle), and the research on the target search problem in the river area is less. As a special task area, the river can be regarded as a curve with certain terrain constraint, so that compared with multi-course or even full-course selection of the unmanned aerial vehicle in a common two-dimensional horizontal area, the river area limits the flight mode of the unmanned aerial vehicle, and the difficulty of target search is increased. For the problems, most of the existing solutions are qualitative passive strategies, such as full-coverage search for unmanned aerial vehicles flying over rivers or greedy search for previous position areas directly reaching targets, and quantitative analysis and heuristic strategy guidance are lacked.
Disclosure of Invention
The invention aims to provide an unmanned aerial vehicle track planning algorithm suitable for a river target search task by combining with the natural condition characteristics of rivers so as to realize the rapid search of the river target task.
In order to achieve the above purpose, the present invention provides the following technical solutions:
an unmanned aerial vehicle track planning algorithm applied to a river target search task is characterized in that the field of view range of the unmanned aerial vehicle is larger than the width of a river, and the planning algorithm comprises the following steps:
s1: abstracting a target river to be searched into a two-dimensional plane curve by a forward distance LsDiscretizing the plane curve into M-L/L as independent variablesA discretization unit, wherein L is the total river length, LsFor each discretized cell length;
s2: in any sampling period, the unmanned aerial vehicle is in two discretization units smMoving inwards; the target to be searched is arranged in each discretization unit smProbability of intrinsic presence p(s)m) Is known, and p(s)m)∈[0,1]And satisfy
Figure RE-BDA0001519860940000031
S3: any number of discretization units form a search curve segment, Sk＝{sm,sm+1,...,snM is more than or equal to 1 and n is more than or equal to M, probability information of the target to be searched in each search curve section is described by a Gaussian mixture model, and P is extractedk,1And Pk,2High-value search curve segment S for region boundaryk；
S4: defining the transition time of the sub-region as the boundary point P of the current search curve segmenti,pGo to the next search curve segment boundary point Pj,qDefining said transition time as:
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 λ1A scale factor representing a transition time of the sub-region; dubins _ cost (P)i,p,Pj,q) Representing the length of a Dubins curve, p representing the departure point flag of the ith search curve segment, q representing the entry point flag of the jth search curve segment, and K representing the number of high-value search curve segments;
defining the sub-area coverage time as the time of searching flight of the unmanned aerial vehicle in the searching curve segment, and Ck＝λ2·LkWherein λ is2Scale factor, L, representing the time covered by a sub-regionkIndicating the length of the search curve segment, Lk＝(n-m+1)·Ls；
Defining the sub-area coverage return as the cumulative probability obtained by the unmanned plane completely covering the high-value search curve segment: rk；
S5: the sub-region transition time, the sub-region coverage time and the sub-region coverage return are used as evaluation indexes, a nearest insertion method is adopted to carry out iterative sequencing of the search sequence on each high-value search curve segment, and the sequencing modeling is as follows:
Figure RE-BDA0001519860940000032
s6: and sequentially connecting the optimized high-value search curve segments in sequence to obtain the final search path of the unmanned aerial vehicle.
Preferably, the method comprises the following steps: the method for extracting the high-value search curve segment comprises the following steps:
using K one-dimensional Gaussian functions
Figure RE-BDA0001519860940000041
Forming a gaussian mixture model, wherein K is 1,2kFor the mean, σ, of each Gaussian functionkThe standard deviation for each gaussian function;
let the weight coefficient occupied by each Gaussian function be αkAnd satisfy
Figure RE-BDA0001519860940000042
The target probability of the curve segment to be searched is
Figure RE-BDA0001519860940000043
Estimating the weighting coefficients α for each Gaussian functionkMean value ukStandard deviation σkIteratively estimating p(s) until a convergence condition is met;
obtaining the corresponding interval of 95.4% probability of each Gaussian function by Pk,1＝uk-2σkAnd Pk,2＝uk+2σkFor high value search curve segments of region boundary points, then Lk＝4σk；Rk＝0.954αk。
Preferably, the convergence condition is | p(s) -p '(s) | < ξ, where p(s) and p'(s) are target probability values before and after iteration, respectively, and ξ ═ 10-5。
Preferably, the method comprises the following steps: if the plurality of unmanned aerial vehicles execute the search task, the method further comprises the following steps: and distributing the sorted high-value search curve segments to a plurality of unmanned aerial vehicles for region distribution.
Preferably, the method comprises the following steps: a method of region allocation comprising the steps of:
assigning a set of high value search curve segments to each drone, obtaining a set A for each droneiAssigning a set of high value search curve segments;
according to the method in the step S5, a set is formed for each unmanned aerial vehicleiThe index calculation is carried out on the inner high-value search curve segment to obtain Ji；
Modeling region assignmentsIs composed of
Figure RE-BDA0001519860940000044
Wherein N isuRepresenting the total number of drones, p1Proportionality coefficient, p, representing total search yield2A scaling factor representing a degree of task balance;
and (4) adopting an iteration method to model and solve the region allocation and determining a region allocation strategy.
Preferably, the method comprises the following steps: the high-value search curve segment set distributed by all unmanned planes meets the following constraint conditions:
Figure RE-BDA0001519860940000051
Figure RE-BDA0001519860940000052
Figure RE-BDA0001519860940000053
the invention has the beneficial effects that:
(1) according to the method, the target river region is abstractly modeled into a two-dimensional curve, so that the problem complexity and the calculated amount are greatly reduced;
(2) according to the method, the characteristics of the river region are approximately described by using the self-adaptive Gaussian mixture model, the high-value sub-region is extracted, the corresponding region with 95.4% probability of Gaussian functions can be extracted in a quantized mode, the search task allocation is performed on the high-value sub-region in a targeted mode, the accuracy of the quantized result is high, and the subsequent problems can be solved easily;
(3) according to the method, the nearest interpolation method is used for carrying out region sequencing, the method can obtain an approximately optimal result, and the solving speed is high;
(4) the algorithm provided by the invention is suitable for planning the search path of one or more unmanned aerial vehicles.
Drawings
FIG. 1 is a flow chart of a method of the present invention;
FIG. 2 is a schematic diagram of a river abstracted into two-dimensional curves;
FIG. 3 is a graph of target probabilities within a river region;
FIG. 4 is an approximation of a target probability based on an adaptive Gaussian mixture model;
FIG. 5 is an iterative process of sub-region ordering based on the nearest insertion method;
fig. 6 shows the flight path of an unmanned aerial vehicle optimized by the method of the present invention for target search with three unmanned aerial vehicles.
Detailed Description
The embodiments of the present invention will be described more fully hereinafter with reference to the accompanying drawings. It should be apparent that the embodiments described in the detailed description are only a part of the embodiments of the present invention, and not all embodiments. All other embodiments, which can be obtained by a person skilled in the art without any inventive step based on the embodiments of the present invention, belong to the scope of the present invention.
The invention provides an algorithm for planning a search track of an unmanned aerial vehicle, in particular to an unmanned aerial vehicle track planning algorithm when the unmanned aerial vehicle is used for a river target search task. By adopting the planning algorithm, the search efficiency of the unmanned aerial vehicle can be greatly improved.
In order to improve unmanned aerial vehicle's search efficiency, the sight of the visual sensor that unmanned aerial vehicle carried on points to under all the time, and its field of view scope is greater than the width of river, so, can guarantee that unmanned aerial vehicle can search for the scope of whole river course width at the in-process of marcing forward. The overall flow chart of the method refers to fig. 1.
The embodiment firstly provides a planning algorithm suitable for a search track of an unmanned aerial vehicle executing a search task.
The unmanned aerial vehicle track planning algorithm applied to the river target search task comprises the following steps:
s1: abstracting the target river into a two-dimensional plane curve and carrying out discretization processing.
Referring to fig. 2, since the search view angle width of the unmanned aerial vehicle is greater than the width of the river, the width of the river can be ignored, and the target to be searched forThe river is abstracted into two-dimensional plane curves. By a forward distance LsDiscretizing the plane curve into M-L/L as independent variablesA discretization unit, wherein L is the total length of the target river, LsFor each discretized cell length; after discretization, discretization units of M rivers are obtained.
S2: and constructing probability information of each discretization unit in a unit sampling period.
In a unit sampling period, the search range of the unmanned aerial vehicle is the length of the discretization unit of the river. Specifically, in any sampling period, the unmanned aerial vehicle is in two discretization units smMoving inwards; the target to be searched is arranged in each discretization unit smProbability of intrinsic presence p(s)m) Is known, and p(s)m)∈[0,1]And satisfy
Figure RE-BDA0001519860940000061
The probability map refers to fig. 3.
S3: and extracting a high-value search curve segment.
Since the probability of the target existing in each search curve segment is different, and the target exists in different levels, in order to improve the search efficiency, the curve segment with high probability of the target existing needs to be selected for searching, and therefore, the high-value search curve segment needs to be extracted.
Any number of discretization units form a search curve segment, Sk＝{sm,sm+1,...,snM is more than or equal to 1 and n is more than or equal to M, probability information of the target to be searched in each search curve section is described by a Gaussian mixture model, and P is extractedk,1And Pk,2High-value search curve segment S for region boundaryk(ii) a Wherein the number of high-value search curve segments is K, and K belongs to {1, 2.
The specific method comprises the following steps: k one-dimensional gaussian functions are used:
Figure RE-BDA0001519860940000071
forming a gaussian mixture model, wherein K is 1,2kFor each heightMean of the gaussian function, σkThe standard deviation for each gaussian function;
let the weight coefficient occupied by each Gaussian function be αkAnd satisfy
Figure RE-BDA0001519860940000072
The target probability of the curve segment to be searched is
Figure RE-BDA0001519860940000073
Estimating the weighting coefficients α for each Gaussian functionkMean value ukStandard deviation σkDefining a training sample consisting of D unit positions, and iteratively estimating p(s) by adopting a maximum expectation method until a convergence condition is met, wherein the convergence condition is p(s) -p '(s) < ξ, p(s) and p'(s) are target probability values before and after iteration respectively, the value of ξ is determined according to the specific convergence precision requirement, and ξ is 10 in the embodiment-5. The iteration result refers to fig. 4.
In the above iterative parameter estimation process, the ratio of the number of training individuals corresponding to each unit position is consistent with the target probability, i.e. Dm＝p(sm) D; dynamic mechanisms such as elimination, combination and splitting can be introduced to adjust the number of the models in a self-adaptive manner. If the weight of a certain Gaussian model is very small and has a certain distance with other models, the component is an unnecessary noise component and can be directly eliminated; if the two Gaussian models are close in distance and not heavy, the two Gaussian models are considered to reflect the same characteristic distribution and have an overfitting phenomenon, so that the two Gaussian models can be combined into a Gaussian component; if the weight and standard deviation of a certain Gaussian component are large, the phenomenon of under-fitting is shown, and the certain Gaussian component needs to be split into two Gaussian components.
According to the parameter estimation result, corresponding intervals of 95.4% probability of each Gaussian function can be quantitatively extracted to serve as river sub-areas with high value (namely high-value search curve segments), and P is obtainedk,1＝uk-2σkAnd Pk,2＝uk+2σkA high-value search curve is formed for the high-value curve segment of the region boundary pointLength L of segmentk＝4σk。
S4: and defining quantitative evaluation indexes, performing grade evaluation and sequencing on the high-value search curve segments by adopting a nearest insertion method, and determining the search sequence of the unmanned aerial vehicle on each high-value search curve segment. The present embodiment defines the following three evaluation indicators, namely, sub-area transition time, sub-area coverage time, and sub-area coverage return.
Defining the transition time of the sub-region as the boundary point P of the current search curve segmenti,p(or search for the starting position Pinitial) Go to the next search curve segment boundary point Pj,qThe finally obtained high-value search curve segments are relatively separated curve segments, the boundary points of each curve segment are discontinuous points, and the transition time of the sub-region represents the time of movement between the mutually discontinuous curve segments. Generally, a Dubins curve is the shortest route of the unmanned aerial vehicle between any two points considering the heading constraint, and the transition time is defined as:
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 λ1A scale factor representing a transition time of the sub-region; dubins _ cost (P)i,p,Pj,q) Representing the length of a Dubins curve, p representing the departure point flag of the ith search curve segment, q representing the entry point flag of the jth search curve segment, and K representing the number of high-value search curve segments;
defining the sub-area coverage time as the time of searching flight of the unmanned aerial vehicle in the searching curve segment, and Ck＝λ2·LkWherein λ is2Scale factor, L, representing the time covered by a sub-regionkLength of the search curve segment is represented as Lk＝(n-m+1)·Ls(ii) a The length of the high-value curve segment extracted by adopting the Gaussian model is Lk＝4σkThus, Ck＝λ2·4σk。
λ1And λ2Is inversely proportional to the relationship of the flight speed of the drone at different phases, e.g. droneThe flight speed of the transition between sub-areas is 2 times the flight speed of the transition covering the sub-area (i.e. over the sub-area), then the above parameters should be satisfied
Figure RE-BDA0001519860940000081
Defining the sub-area coverage return as the cumulative probability R obtained by completely covering the high-value search curve segment by the unmanned aerial vehiclekThe probability is the sum of the target existence probabilities of all the high-value search curve segments, and compared with the high-value curve segments extracted by the Gaussian model, the river subarea is the interval range corresponding to 95.4 percent of the probability of each Gaussian function, and the weight coefficient is αkTherefore, the sub-region coverage return is defined as follows: rk＝0.954αk。
S5: and sequencing the search sequence of each high-value search curve segment to realize an optimal high-efficiency search strategy.
Adopting the sub-region transition time, the sub-region coverage time and the sub-region coverage return as evaluation indexes, and adopting a nearest insertion method to search each high-value search curve segment LpAnd carrying out iterative sequencing of the search sequence, wherein the sequencing is modeled as:
Figure RE-BDA0001519860940000091
wherein, PinitialIn order to search for the start position,
Figure RE-BDA0001519860940000092
respectively representing the two end points of the first high value curve segment from which the search is initiated,
Figure RE-BDA0001519860940000093
the sub-region representing the first high value curve segment from which the search is initiated covers the reward,
Figure RE-BDA0001519860940000094
the sub-region coverage time representing the first high value curve segment from which the search is initiated. Accordingly, the method can be used for solving the problems that,
Figure RE-BDA0001519860940000095
respectively representing the corresponding indexes of the k-th high-value curve segment.
It should be noted here that the two boundary point flags of the sub-region are 1 and 2, respectively, and their sum is 3. 1 and 2 do not refer to a number but to a mark, e.g. the two boundary points of the 5 th sub-area are P5,1，P5,2. The entry point flag bit of the unmanned aerial vehicle to the k-1 sub-area is nk-1And the flag bit of the flying-out point of the unmanned aerial vehicle after the sub-area is searched is 3-nk-1。
The sub-region sorting problem can be regarded as a typical traveler problem, and therefore a nearest insertion method is adopted to obtain a near-optimal sorting result. From the remaining set of unordered subregions in each iteration
Figure RE-BDA0001519860940000096
In the sequence of ordered sub-regions, { l }, by randomly selecting a sub-region and inserting it into the sequence of ordered sub-regions1,...,lIAnd calculating an index value J (I +1) of the new sorting region according to a sorting model formula, further calculating an increment delta J of the index value, wherein the increment delta J is J (I +1) -J (I), selecting a case when the delta J takes the maximum value from all possible cases, and then inserting a selected certain unsorted subregion into the sorted subregion sequence. The iterative process repeats K steps, and the approximate ordering of the K sub-regions is completed.
FIG. 5 shows the starting position P from the searchinitialRelative to three sub-regions (P)1,1,P1,2)、(P2,1,P2,2) And (P)3,1,P3,2) And (3) adopting a nearest interpolation method to approximately sequence the 3 sub-regions, wherein a solid line represents a river sub-region, a dotted line represents all possible transition track segments, and a solid line represents a determined transition track segment. It can be seen that from the starting position PinitialPossible transition paths include (P)initial,P1,1)、(Pinitial,P1,2)、(Pinitial,P2,1)、 (Pinitial,P2,2)、(Pinitial,P3,1)、(Pinitial,P3,2) (ii) a After 1 iteration of the calculation, determine (P)initial,P1,1) For the optimal path and starting the next iterative calculation of the path, the possible transition paths include as shown in fig. 5(c), after the iterative calculation, the determination (P) is carried out1,2,P2,1) Is an optimal path; this is repeated to finally obtain the search path shown in fig. 5 (f).
After the sorting is finished, the grade of each river sub-region (namely the searching sequence of the sub-regions) and the zone bit of each sub-region entry point (namely the sub-region is entered from one of the two end points of each sub-region) can be determined, and the unmanned aerial vehicle enters the sub-regions according to the zone bit and searches one by one according to the sequence of the sub-regions.
The search algorithm is suitable for planning the search track of the unmanned aerial vehicle for executing the search task.
In the prior art, in order to realize efficient searching, a plurality of unmanned aerial vehicles commonly perform a searching task. Therefore, the present embodiment further provides an algorithm for planning a search trajectory of a plurality of drones when the drones jointly execute a search task.
When a plurality of unmanned aerial vehicles jointly execute the search task, the total search yield and the balance degree of the task need to be comprehensively considered, and the sub-areas are distributed to the unmanned aerial vehicles.
Specifically, if a plurality of drones execute a search task, the method further includes the following steps: and distributing the sorted value search curve segments to a plurality of unmanned aerial vehicles for region distribution.
A method of region allocation comprising the steps of:
assigning a high value search curve segment to each drone, each drone obtaining a set of AiThe high-value search curve segment set is distributed, and the curve segment set comprises a plurality of high-value search curve segments;
in order to avoid collision among the drones, each sub-area is required to be allocated to only one drone, and in order to ensure efficient utilization of the drones, at least one sub-area is required to be allocated to each drone, so that the high-value search curve segment set allocated to all the drones meets the following constraint conditions:
Figure RE-BDA0001519860940000111
Figure RE-BDA0001519860940000112
Figure RE-BDA0001519860940000113
wherein N isuRepresenting the number of drones and K representing the number of high value search curve segments.
Process for distributing search routes for multiple drones referring to the method in steps S4, S5, the curve segment set A obtained for each droneiThe high-value search curve segment in the graph carries out index calculation (including sub-region transition time, sub-region coverage time and sub-region coverage return), and J is obtainedi；
Modeling region assignments as
Figure RE-BDA0001519860940000114
Wherein N isuRepresenting the total number of drones, p1Proportionality coefficient, p, representing total search yield2A scaling factor representing a degree of task balance; rho1And ρ2The values need to be assigned according to specific task requirements, for example, when the detection capabilities of all machines are similar, rho can be increased2A value of (d); when the difference of the detection capability of each machine is large, rho can be increased1The value of (c).
And (3) adopting an iterative method to model and solve the region allocation and determining a region allocation strategy, as shown in FIG. 6. The final optimized trajectory can be expressed as:
Figure RE-BDA0001519860940000115
the algorithm of the invention is adopted to plan the unmanned aerial vehicle track, and the high-efficiency search of the target river section can be realized.

Claims (6)

1. The unmanned aerial vehicle track planning method applied to the river target search task is characterized in that the field of view range of the unmanned aerial vehicle is larger than the width of a river, and the planning algorithm comprises the following steps:
s1: abstracting a target river to be searched into a two-dimensional plane curve by a forward distance LsDiscretizing the plane curve into M-L/L as independent variablesA discretization unit, wherein L is the total river length, LsFor each discretized cell length;
s2: in any sampling period, the unmanned aerial vehicle is in two discretization units smMoving inwards; the target to be searched is arranged in each discretization unit smProbability of intrinsic presence p(s)m) Is known, and p(s)m)∈[0,1]And satisfy
Figure FDA0002373981880000011
S3: any number of discretization units form a search curve segment, Sk＝{sm,sm+1,…,sn}，1≤m<n is less than or equal to M, probability information of the target to be searched in each search curve segment is described by a Gaussian mixture model, and P is extractedk,1And Pk,2High-value search curve segment S for region boundarykWherein the number of the high-value search curve segments is K, and K belongs to {1, 2.., K };
s4: defining the transition time of the sub-region as the boundary point P of the current search curve segmenti,pGo to the next search curve segment boundary point Pj,qDefining said transition time as:
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 λ1A scale factor representing a transition time of the sub-region; dubins _ cost (P)i,p,Pj,q) Representing the length of the Dubins curve, p represents the departure point flag for the ith search curve segment,q represents the entry point flag of the jth search curve segment, and K represents the number of high-value search curve segments;
defining the sub-area coverage time as the time of searching flight of the unmanned aerial vehicle in the searching curve segment, and Ck＝λ2·LkWherein λ is2Scale factor, L, representing the time covered by a sub-regionkIndicating the length of the search curve segment, Lk＝(n-m+1)·Ls；
Defining the sub-area coverage return as the cumulative probability obtained by the unmanned plane completely covering the high-value search curve segment: rk；
S5: the sub-region transition time, the sub-region coverage time and the sub-region coverage return are used as evaluation indexes, a nearest insertion method is adopted to carry out iterative sequencing of the search sequence on each high-value search curve segment, and the sequencing modeling is as follows:
Figure FDA0002373981880000021
Pinitialis a search starting position;
Figure FDA0002373981880000022
respectively representing two end points of a first high-value curve segment for starting searching;
Figure FDA0002373981880000023
representing the two end points of the kth high-value curve segment;
Figure FDA0002373981880000026
a sub-region coverage reward representing a kth high-value curve segment;
Figure FDA0002373981880000027
sub-region coverage times representing the kth high-value curve segment;
s6: and sequentially connecting the optimized high-value search curve segments in sequence to obtain the final search path of the unmanned aerial vehicle.
2. The unmanned aerial vehicle track planning method applied to river target search mission of claim 1, wherein the method of extracting high-value search curve segment comprises the following steps:
using K one-dimensional Gaussian functions
Figure FDA0002373981880000024
Forming a gaussian mixture model, wherein K is 1,2kFor the mean, σ, of each Gaussian functionkThe standard deviation for each gaussian function;
let the weight coefficient occupied by each Gaussian function be αkAnd satisfy
Figure FDA0002373981880000025
The target probability of the curve segment to be searched is
Figure FDA0002373981880000031
Estimating the weighting coefficients α for each Gaussian functionkMean value ukStandard deviation σkIteratively estimating p(s) until a convergence condition is met;
obtaining the corresponding interval of 95.4% probability of each Gaussian function by Pk,1＝uk-2σkAnd Pk,2＝uk+2σkFor high value search curve segments of region boundary points, then Lk＝4σk；Rk＝0.954αk。
3. The unmanned aerial vehicle track planning method applied to river target search task of claim 2, wherein the convergence condition is | p(s) -p'(s) & ltY & gt<ξ, wherein p(s) and p'(s) are target probability values before and after iteration, respectively, and ξ is 10-5。
4. The method for planning the flight path of the unmanned aerial vehicle applied to the river target search task according to claim 1, wherein if a plurality of unmanned aerial vehicles perform the search task, the method further comprises the following steps: and distributing the sorted value search curve segments to a plurality of unmanned planes for region distribution.
5. The unmanned aerial vehicle track planning method applied to river target search task of claim 4, wherein the region allocation method comprises the following steps:
assigning a high value search curve segment to each drone, each drone obtaining a set of AiAssigning a set of high value search curve segments;
according to the method in the step S5, a set is formed for each unmanned aerial vehicleiThe index calculation is carried out on the inner high-value search curve segment to obtain Ji；
Modeling region assignments as
Figure FDA0002373981880000032
Wherein N isuRepresenting the total number of drones, p1Proportionality coefficient, p, representing total search yield2A scaling factor representing a degree of task balance;
and (4) adopting an iteration method to model and solve the region allocation and determining a region allocation strategy.
6. The unmanned aerial vehicle track planning method applied to river target search mission of claim 5, wherein the set of high value search curve segments allocated to all unmanned aerial vehicles satisfies the following constraint conditions:
Figure FDA0002373981880000041
Figure FDA0002373981880000042
Figure FDA0002373981880000043
CN201711403481.0A 2017-12-22 2017-12-22 Unmanned aerial vehicle track planning algorithm applied to river target search task Active CN108168558B (en)

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 CN108168558A (en) 2018-06-15
CN108168558B true 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)

Families Citing this family (3)

* Cited by examiner, † Cited by third party
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
CN111487986B (en) * 2020-05-15 2021-03-16 中国海洋大学 Underwater robot cooperative target searching method based on global information transfer mechanism
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

Citations (7)

* Cited by examiner, † Cited by third party
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)

Patent Citations (7)

* Cited by examiner, † Cited by third party
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)

* Cited by examiner, † Cited by third party
Title
Autonomous Searching and Tracking of a River using an UAV;Sivakumar Rathinam等;《2007 American Control Conference》;20070730;359-364 *
UAV feasible path planning based on disturbed fluid and trajectory propagation;Yao Peng等;《Chinese Journal of Aeronautics》;20150622;第28卷(第4期);1163-1177 *
多UCAV动态协同任务规划建模与滚动优化方法研究;霍霄华;《中国博士学位论文全文数据库信息科技辑》;20090715(第7期);138-130 *
多无人机任务分配与路径规划算法研究;丁家如;《中国优秀硕士学位论文全文数据库工程科技Ⅱ辑》;20160715(第7期);031-121 *
无人机航迹规划算法研究;简康;《中国优秀硕士学位论文全文数据库工程科技Ⅱ辑》;20160315(第3期);031-430 *

Also Published As

Publication number Publication date
CN108168558A (en) 2018-06-15

Similar Documents

Publication Publication Date Title
Wu et al. Cooperative path planning for heterogeneous unmanned vehicles in a search-and-track mission aiming at an underwater target
CN106441308B (en) A kind of Path Planning for UAV based on adaptive weighting dove group&#39;s algorithm
CN109254588B (en) Unmanned aerial vehicle cluster cooperative reconnaissance method based on cross variation pigeon swarm optimization
US8260485B1 (en) Adaptive multi-vehicle area coverage optimization system and method
CN102915652B (en) aircraft traffic separation system and method
CN108168558B (en) Unmanned aerial vehicle track planning algorithm applied to river target search task
CN111598473B (en) Multi-platform combined task planning method for complex observation task
CN110109477A (en) Unmanned plane cluster multi objective control optimization method based on dove colony intelligence backward learning
CN110262563A (en) Multiple no-manned plane collaboratively searching mesh calibration method waterborne
CN103557867A (en) Three-dimensional multi-UAV coordinated path planning method based on sparse A-star search (SAS)
CN110288001A (en) Target identification method based on the training study of target data feature
CN105654500A (en) Unmanned aerial vehicle target detection method for optimizing visual attention mechanism based on bionic pigeons
CN104317305B (en) A kind of front flight path of penetrating threatened towards complicated battlefield determines method
CN112432649A (en) Heuristic unmanned aerial vehicle swarm flight path planning method introducing threat factors
CN113238232B (en) Target searching method of autonomous underwater vehicle system for ocean static target
CN116257082B (en) Distributed active cooperative detection method for multiple unmanned aerial vehicles
CN112733251A (en) Multi-unmanned aerial vehicle collaborative track planning method
Yan et al. A safe landing site selection method of UAVs based on LiDAR point clouds
CN116954233A (en) Automatic matching method for inspection task and route
CN114167898A (en) Global path planning method and system for data collection of unmanned aerial vehicle
CN115060263A (en) Flight path planning method considering low-altitude wind and energy consumption of unmanned aerial vehicle
CN112987765B (en) Precise autonomous take-off and landing method of unmanned aerial vehicle/boat simulating attention distribution of prey birds
CN113805609A (en) Unmanned aerial vehicle group target searching method based on chaos lost pigeon group optimization mechanism
CN105760813A (en) Unmanned aerial vehicle target detection method based on plant branch and root evolution behaviors
CN111506078B (en) Robot navigation method and system

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