CN109582027A - A kind of USV cluster collision-avoidance planning method based on Modified particle swarm optimization algorithm - Google Patents
A kind of USV cluster collision-avoidance planning method based on Modified particle swarm optimization algorithm Download PDFInfo
- Publication number
- CN109582027A CN109582027A CN201910032577.3A CN201910032577A CN109582027A CN 109582027 A CN109582027 A CN 109582027A CN 201910032577 A CN201910032577 A CN 201910032577A CN 109582027 A CN109582027 A CN 109582027A
- Authority
- CN
- China
- Prior art keywords
- usv
- optimization algorithm
- particle swarm
- swarm optimization
- coordinate system
- 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
- 239000002245 particle Substances 0.000 title claims abstract description 78
- 238000004422 calculation algorithm Methods 0.000 title claims abstract description 59
- 238000005457 optimization Methods 0.000 title claims abstract description 41
- 238000000034 method Methods 0.000 title claims abstract description 24
- 238000005096 rolling process Methods 0.000 claims abstract description 6
- 230000004888 barrier function Effects 0.000 claims description 25
- 230000006870 function Effects 0.000 claims description 20
- 230000002265 prevention Effects 0.000 claims description 12
- 230000008447 perception Effects 0.000 claims description 8
- 238000013461 design Methods 0.000 claims description 6
- 230000007613 environmental effect Effects 0.000 claims description 5
- 230000008569 process Effects 0.000 claims description 4
- 230000033228 biological regulation Effects 0.000 claims description 3
- 230000005484 gravity Effects 0.000 claims description 3
- 230000006872 improvement Effects 0.000 claims description 3
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 abstract description 11
- 230000000007 visual effect Effects 0.000 description 9
- 238000004458 analytical method Methods 0.000 description 4
- 230000015572 biosynthetic process Effects 0.000 description 4
- 238000003786 synthesis reaction Methods 0.000 description 4
- 230000001133 acceleration Effects 0.000 description 3
- 238000011161 development Methods 0.000 description 3
- 238000002474 experimental method Methods 0.000 description 3
- 230000008901 benefit Effects 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000007796 conventional method Methods 0.000 description 2
- 230000007423 decrease Effects 0.000 description 2
- 230000002068 genetic effect Effects 0.000 description 2
- 238000011835 investigation Methods 0.000 description 2
- 241000208340 Araliaceae Species 0.000 description 1
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 1
- 235000003140 Panax quinquefolius Nutrition 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 235000008434 ginseng Nutrition 0.000 description 1
- 230000001788 irregular Effects 0.000 description 1
- 238000012804 iterative process Methods 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 238000004904 shortening Methods 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 238000001356 surgical procedure Methods 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, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
- G05D1/02—Control of position or course in two dimensions
- G05D1/0206—Control of position or course in two dimensions specially adapted to water vehicles
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Automation & Control Theory (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Aviation & Aerospace Engineering (AREA)
- Control Of Position, Course, Altitude, Or Attitude Of Moving Bodies (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention belongs to unmanned water surface ship avoidance fields, and in particular to a kind of USV cluster collision-avoidance planning method based on Modified particle swarm optimization algorithm.The invention mainly includes steps: the comprehensive ken model of USV is established according to the parameter characteristic of pathfinder and photoelectric sensor；Construct coordinate system；Constructing environment model；Rolling optimal strategy and Modified particle swarm optimization algorithm designed for USV cluster collision-avoidance planning；The information input of information and target point that comprehensive sensor is detected obtains the navigation of USV subsequent time and the adjustment instruction of the speed of a ship or plane into Modified particle swarm optimization algorithm.The present invention not only overcomes standard particle colony optimization algorithm and is easy to the shortcomings that falling into local optimum, and combine USV current context information, improve the real-time of USV collision-avoidance planning, the corner optimization that joined USV in fitness function, also improves the flatness in path while obtaining optimal path.
Description
Technical field
The invention belongs to unmanned water surface ship avoidance fields, and in particular to a kind of USV based on Modified particle swarm optimization algorithm
Cluster collision-avoidance planning method.
Background technique
In recent years, more and more countries focused on the development of sea powers, and unmanned water surface ship is small in size with its, flexibility
The high and strong advantage of fight capability becomes the hot spot of research.Unmanned water surface ship not only can be with autonomous navigation and contexture by self, also
The tasks such as target following and bay patrol can be completed, so unmanned water surface ship is all sent out no matter in military field or civil field
Wave increasingly important role.However, single USV seems when facing unknown working environment and diversified task mission
It is powerless to be efficiently completed task, and joined together the group system constituted, robustness and mobility by more USV
By force, operating efficiency is high and job area is also relatively wider, these advantages of USV group system receive the favor of many countries,
In following a period of time, the intelligence of unmanned water surface ship, architecture, standardization by be various countries' development emphasis, unmanned water surface
Ship cluster will also play the part of more importantly role in the synthesis cooperation with other weapons, and unmanned water surface ship cluster system
Revolutionary impact will also be generated to the whole looks in the following marine battlefield by uniting.Marine environment is complicated and changeable, either island,
Reef or some other dynamic objects, which all may must navigate by water unmanned boat cluster, constitutes certain threat, and in cluster
Also certain safe distance must be kept between each USV, this requires unmanned water surface ship clusters during completing each task
Must have certain autonomous collision prevention ability, the ability of the autonomous collision prevention of unmanned water surface ship cluster is to ensure that it is efficiently and safely complete
At one of the key technology of every mission task.
The conventional methods such as Artificial Potential Field Method, Visual Graph method, rolling window method and genetic algorithm, ant group algorithm, population
The intelligent algorithms such as algorithm, neural network algorithm are currently used collision-avoidance planning algorithms.Most of collision-avoidance planning algorithm all cannot
One is cooked up without collision prevention optimal path, conventional method such as Artificial Potential Field Method for USV in a short time, although having simple former
Reason, lesser calculation amount, very fast arithmetic speed, but be easy to can not find globally optimal solution, or even shake during collision prevention
Phenomenon, although swarm intelligence algorithm such as genetic algorithm can be to avoid Local Extremum be fallen into, principle is complicated, computationally intensive, and
Planning time is longer, therefore, for USV design it is a kind of simple, quickly, the collision-avoidance planning algorithm of real-time online be vital.
The autonomous collision prevention of USV cluster may be implemented using the method combined based on Rolling optimal strategy and particle swarm optimization algorithm, it should
Algorithm carries out operation using the local environmental information that sensor real-time perception obtains, and the use of particle swarm optimization algorithm is quickly collection
Each USV in group cooks up one without collision prevention optimal path in the comprehensive ken.
Summary of the invention
The USV cluster collision-avoidance planning method based on Modified particle swarm optimization algorithm that the purpose of the present invention is to provide a kind of.
A kind of USV cluster collision-avoidance planning method based on Modified particle swarm optimization algorithm, mainly comprises the steps that
Step 1: the comprehensive ken model of USV is established according to the parameter characteristic of pathfinder and photoelectric sensor；
Step 2: building coordinate system；
Step 3: constructing environment model；
Step 4: Rolling optimal strategy and Modified particle swarm optimization algorithm designed for USV cluster collision-avoidance planning；
Step 5: the information input of information and target point that comprehensive sensor is detected to Modified particle swarm optimization algorithm
In, obtain the navigation of USV subsequent time and the adjustment instruction of the speed of a ship or plane.
Three coordinate systems: global context coordinate system X are established in step 2GOGYG, local coordinate system X based on USVUOUYUWith
Local coordinate system X based on the comprehensive perception information of sensorRORYR；
Global coordinate system XGOGYGUsing east northeast coordinate system, the lower left corner is coordinate origin, and X-axis is directed toward direct north, and Y-axis refers to
To due east direction；
Local coordinate system is divided into hull coordinate system and sensor coordinate system, and the origin of hull coordinate system is located at the center of gravity of USV
Place, pathfinder and photoelectric sensor be mounted at a of hull center, and X-axis and Y-axis are parallel to the X of USV local coordinate system
Axis and Y-axis.
Constructing environment model in step 3 specifically includes the following steps:
(3.1) data acquire: the sensing range of pathfinder and photoelectric sensor is 360 °, it is assumed that every in 1 ° of angle
Sensor obtains and can only obtain a range information, and angle is changed to 360 ° from 0 ° counterclockwise, i.e., sensor is often swept
It retouches and once just obtains 361 groups of perception data points, these data points include two parameters, and first parameter is between point and sensor
Angle, the distance of second parameter point to sensor；
(3.2) perception data point is extracted as line segment feature: firstly, being divided into point set by way of region division
Then data point set in each region is carried out linear fit again, and then is extracted as line segment feature by multiple regions；
(3.3) extruding, extruding mode master only static-obstacle thing extruding: are carried out to its part to the extruding of static-obstacle thing
It is divided into following three kinds of situations: the extruding of the extruding of single line section, the extruding of adjacent segments and curve；
(3.4) trajectory predictions of dynamic barrier: being puffed to a circle for dynamic barrier first, and then basis is returned certainly
Model prediction is returned to go out position and direction of the dynamic barrier under global coordinate system.
Step 4 specifically includes: USV during the motion, every time to integrate portion, the specific item punctuate regional planning agency path in the ken,
Then using the window in each comprehensive ken of Modified particle swarm optimization algorithm optimization, the improvement of particle swarm optimization algorithm includes used
The designs of property weight and Studying factors, the selection of fitness function.
USV during the motion, every time to integrate portion, the specific item punctuate regional planning agency path in the ken, will be held in motion process
Row following steps:
(4.1) USV obtains the environmental information in the comprehensive ken by self-contained sensor；
(4.2) one is cooked up using particle swarm algorithm from starting point to specific item punctuate to pass through without the optimal road of collision prevention；
(4.3) USV will move to specific item punctuate from starting point along the path cooked up；
(4.4) starting point, return step 1, until USV drives to global terminal are designated as with this specific item.
It is taken based on the value of the inertia weight ω of the inertia adjustable strategies Curve guide impeller of tangent function, formula are as follows:
In formula, α=π/4, m are controlling elements, and k is current iteration number, kmaxFor maximum number of iterations；
It is taken based on Serial regulation stragetic innovation design learning factor c1And c2Value, formula are as follows:
K is current iteration number, k in formulamaxFor maximum number of iterations, c1s=2.5, c2s=1.0, c1e=1.0, c2e=
2.5；
The selection of fitness function:
The starting point of known USV is (x0,y0), target point is (xn,yn), subpath points are (xi,yi), wherein i=0,1,
2 ... n-1, n, then fitness function is defined as:
The fitness function defined according to each corner of USV are as follows:
In conjunction with above-mentioned formula, USV fitness function total during collision-avoidance planning is obtained:
F=λ1f1+λ2f2。
Particle swarm optimization algorithm specifically includes the following steps:
Step 1: the parameter in initialization particle swarm algorithm: ωs、ωe、c1s、c1e、c2s、c2e, population quantity PNUM, most
Big the number of iterations kmax, enable k=0；
Step 2: the history optimal value of particle individual is set as by initialization particle populations, the i.e. speed of particle and position
Pbest, group's optimal value are set as gbest；
Step 3: in the evolution in the present age, the fitness function value f of each particle is calculated；
Step 4: the pbest and gbest of more new particle；
Step 5: the speed of more new particle and position；
Step 6: it if meeting termination condition, exports gbest and terminates, otherwise turn Step 5.
Particle is updated to following formula:
Wherein, c1、c2For Studying factors, c under normal conditions1+c2≤ 4, r1、r2For the random number between [0,1], ω is used
Property weight, pidFor individual optimal value, gidFor global optimum, subscript " i ", " d " respectively indicate i-th of particle and i-th of particle
Dimension.
The beneficial effects of the present invention are:
Concentric circles is generated in the synthesis ken that the present invention passes through each USV in the cluster to model, and utilizes pathfinder
The comprehensive environmental information measured in real time with photoelectric sensor particle swarm optimization algorithm after improvement combines to be come in a rolling manner
Plan in the comprehensive ken of each USV without collision prevention path, not only overcome standard particle colony optimization algorithm and be easy to fall into part most
Excellent disadvantage, and USV current context information is combined, the real-time of USV collision-avoidance planning is improved, is added in fitness function
The corner optimization for having entered USV, also improves the flatness in path while obtaining optimal path.
Detailed description of the invention
Fig. 1 is synthesis ken simulation drawing of the invention；
Fig. 2 is coordinate schematic diagram of the invention；
Fig. 3 (a) is the extruding figure of single line section；Fig. 3 (b) is the extruding figure of adjacent segments；
Fig. 4 (a) is the extruding figure of convex curve；Fig. 4 (b) is the extruding figure of concave curve；
Fig. 5 (a) is the visual tangent line figure after barrier extruding；Fig. 5 (b) is that barrier is swollen after USV travels a distance
Visual tangent line figure after change；
Fig. 6 is algorithm flow chart of the invention；
Fig. 7 is simulated effect figure of the invention.
Specific embodiment
The present invention is described in more detail with reference to the accompanying drawing.
The comprehensive ken model of USV is established according to the parameter characteristic of pathfinder and photoelectric sensor；
The parameter characteristic of pathfinder and photoelectric sensor is as shown in table 1:
1 USV sensors performance parameter table of table
Construct global coordinate system and local coordinate system
As shown in Fig. 2, the local coordinate system X of USVUOUYUOrigin be located at the center of gravity of USV, with the advance side of current USV
To for X-axis positive direction, the direction perpendicular to X-axis and direction USV starboard side is Y-axis positive direction.Pathfinder and photoelectric sensor
It is mounted at a of hull center, X-axis and Y-axis are parallel to the X-axis and Y-axis of USV local coordinate system, QiFor sensor perception
The obstacle arrived, it is d at a distance from sensor, and angle θ, USV are (x in the coordinate of global context coordinate systemu,yu), αuIt is
The bow of USV to global context reference axis XGThe angle of (north orientation).
By the coordinate (x of sensing datai,yi) it is converted into (the x of coordinate in global coordinate systemg,yg) formula are as follows:
The extruding of barrier
1. the extruding of single line section.Assume that two endpoints of certain line segment are respectively a as shown in Fig. 3 (a)1、a2, USV and barrier
Safe navigation distance be ds, then by a1Extend ds length to a11, a2Extend ds length to a22, cross a11It is a11a22Vertical line,
Two o'clock b is taken on vertical line1And b2, b1And b2With a11Distance be ds, similarly, cross a22It is a11a22Vertical line, b3With b4It is vertical line
On distance a22It is the two o'clock of ds, connects b1b2b4b3It is the polygon after the line segment extruding.
2. the extruding of adjacent segments.As shown in Fig. 3 (b), for the broken line of adjacent segments composition, firstly, according to single line section
Processing method every line segment is puffed to a quadrangle, then, merge adjacent quadrangle, two quadrangles intersected
Part, then by their intersection point e4As polygon vertex new after merging, discard portion others intersection point and original part
Vertex, if the vertex of a polygon appears in another polygon, this vertex just will be given up, finally, two
Quadrangle is collectively referred to as a polygon, after two adjacent line segment extrudings, just forms one containing there are six tops by merging
Point e1e3e2e4e5e6Polygon.
3. the extruding of convex curve.Shown in the extruding of convex curve such as Fig. 4 (a), firstly, two endpoint h of connection convex curve1、
h2, extend ds length respectively to h11And h22, cross h11It is h11h22Vertical line, take a point h on vertical line12, h12To h11Distance be ds,
Similarly, h is crossed22It is h11h22Vertical line, take a point h on vertical line21, h21To h22Distance be ds, curve connect h12And h21, h12h21
It is just curve h11h22Curve after extruding.
4. the extruding of concave curve.Shown in the extruding of concave curve such as Fig. 4 (b), firstly, two endpoint g of connection convex curve1、
g2, cross g1It is g1g2Vertical line, take a point g on vertical line11, g11To g1Distance be ds, similarly, cross g2It is g1g2Vertical line, take vertical
A point g on line22, g22To g2Distance be ds, curve connect g11And g22, g11g22It is just curve g1g2Curve after extruding.
As shown in Fig. 5 (a), in the synthesis ken of USV pathfinder and photoelectric sensor be only able to detect barrier 1,
A part of barrier 2 and barrier 3, i.e. solid line indicate part, and barrier 3 some blocked by barrier 1, hinder
Hindering boundary point of the viewable portion of object 1 after extruding is l1And l2, boundary point of the viewable portion of barrier 2 after extruding
For l3And l4, boundary point of the viewable portion boundary of barrier 3 after extruding is l5And l6, connect Ol1、Ol2、Ol3、Ol4、Ol5
And Ol6, wherein l1、l2、l3And l4The referred to as visual point of contact of extruding, l5And l6The referred to as pseudo- visual point of contact of extruding, Ol1、Ol2、Ol3And Ol4
The referred to as visual tangent line of extruding, Ol5And Ol6The referred to as pseudo- visual tangent line of extruding.After USV travels a distance, as shown in Fig. 5 (b),
Viewpoint is located at O1When place, l7And l8It is viewpoint O1The visual point of contact of extruding, O1l7And O1l8It is the visual tangent line of its extruding.
The autoregression model of Obstacle Position prediction
According to above-mentioned analysis it is found that USV can obtain the speed and position letter of t moment dynamic barrier by sensor
Breath, and the position and direction of dynamic barrier subsequent time can then be gone out by autoregressive model prediction, it is assumed that k-th of dynamic hinders
Hinder object the position of t moment be Lk(t)=(xob,yob), the position sequence { L of barrierk(t), t=1,2 ... } it indicates, then may be used
To be formulated the n rank autoregression model of the position of k-th of dynamic barrier:
Wherein, e (t) is prediction error, factor alphaiFor 2 × 2 matrix.
The acceleration change of dynamic barrier can be made slow by shortening the time of sampling, therefore single order can be used
Acceleration { a of the autoregression model to dynamic barrierk(t), t=1,2 ... } it is modeled:
ak(t)=βk,tak(t-1)+w(t)
Wherein, w (t) is prediction error, βk,tFor Fast track surgery coefficient.
Assuming that the speed of t moment dynamic barrier is vk(t), then the relationship of its acceleration and position is as follows:
ak(t)=vk(t)-vk(t-1)
=| Lk(t)-Lk(t-1)|-|Lk(t-1)-Lk(t-2)|
=Lk(t)-2Lk(t-1)+Lk(t-2)
By predictable k-th of the dynamic barrier out of above-mentioned analysis in the position at t+1 moment are as follows:
Lk(t)=(2+ βk,t)Lk(t-1)+(1+2βk,t)Lk(t-2)+βk,tLk(t-3)+w(t)
Lk(t+1)=Lk(t)+vk(t)+βkak(t)
USV during the motion, every time to integrate portion, the specific item punctuate regional planning agency path in the ken, will be held in motion process
Row following steps:
(4.1) USV obtains the environmental information in the comprehensive ken by self-contained sensor；
(4.2) one is cooked up using particle swarm algorithm from starting point to specific item punctuate to pass through without the optimal road of collision prevention；
(4.3) USV will move to specific item punctuate from starting point along the path cooked up；
The self-renewing of particle swarm optimization algorithm:
(4.4) starting point, return step 1, until USV drives to global terminal are designated as with this specific item.
In each iteration, particle is evaluated by two " extreme values " and updates oneself: first is exactly that particle itself exists
Find during historical search without the optimal subpath points of collision prevention, referred to as individual extreme value, the other is what entire population was found
Without the optimal subpath points of collision prevention, referred to as global extremum.The self-renewing of particle such as formula:
Wherein, c1、c2For Studying factors, c under normal conditions1+c2≤ 4, r1、r2For the random number between [0,1], ω is used
Property weight, pidFor individual optimal value, gidFor global optimum, subscript " i ", " d " respectively indicate i-th of particle and i-th of particle
Dimension, wherein the value of the dimension d of particle is 2.
The parameter designing of particle swarm optimization algorithm
The value of ω is bigger, and the ability of searching optimum of particle is stronger, while also accelerating convergence speed of the algorithm, but asks
The optimal solution of topic is not easy to obtain, and the value of ω becomes smaller, and local search ability becomes strong, and the optimal solution of optimization problem is also easy to get, but
It is that convergence speed of the algorithm will be slack-off, or even falls into dead state.In particle swarm optimization algorithm iteration early period, in order to as far as possible
It was found that bigger solution space, particle needs stronger ability of searching optimum, and when to the later period, particle needs to reinforce part
Search capability, as far as possible search optimal solution.Therefore, with the increase of the number of iterations, inertia weight ω should constantly reduce.It takes
The value of inertia adjustable strategies Curve guide impeller ω based on tangent function, such as formula:
In formula, α=π/4, ωs=0.9, ωe=0.4, m are controlling elements, and value 0.5, k is current iteration number,
kmaxFor maximum number of iterations, ωkAs the increase of the number of iterations is from 0.9 linear decrease to 0.4.
At algorithm initial stage, the ability of searching optimum of particle is eager to excel, to detect global space as far as possible, after having arrived algorithm
Phase, particle local search will find globally optimal solution, therefore to reinforce the local search ability of particle, therefore,
C in iterative process1It should be gradually reduced, c2It should be gradually increased.It is taken based on the tune of the Serial regulation stragetic innovation design learning factor
Section, such as formula:
K is current iteration number, k in formulamaxFor maximum number of iterations, c1s=2.5, c2s=1.0, c1e=1.0, c2e=
2.5, c1As the increase of the number of iterations is from 2.5 decreases in non-linear to 1.0, c2Non-linear increase to 2.5 from 1.0.
The size of population scale PNUM is 80.
The starting point of known USV is (x0,y0), target point is (xn,yn), subpath points are (xi,yi), wherein i=0,1,
2 ... n-1, n, then fitness function is defined as:
The fitness function defined according to each corner of USV are as follows:
In conjunction with the above analysis, USV fitness function total during collision-avoidance planning is obtained:
F=λ1f1+λ2f2
Wherein λ1=0.8, λ2=20.
Can obtaining particle swarm optimization algorithm according to the above analysis, specific step is as follows:
Step 1: the parameter in initialization particle swarm algorithm: ωs、ωe、c1s、c1e、c2s、c2e, population quantity PNUM, most
Big the number of iterations kmax, enable k=0；
Step 2: the history optimal value of particle individual is set as by initialization particle populations, the i.e. speed of particle and position
Pbest, group's optimal value are set as gbest；
Step 3: in the evolution in the present age, the fitness function value f of each particle is calculated；
Step 4: the pbest and gbest of more new particle；
Step 5: the speed of more new particle and position；
Step 6: it if meeting termination condition, exports gbest and terminates, otherwise turn Step 5.
This invention builds emulation experiment environment by development platform of Qt, and according to the ginseng of pathfinder and photoelectric sensor
Number simulated behavior USV comprehensive survey ambient condition information, in emulation experiment, the size of simulated environment is 130 × 110, unit
For km, pathfinder can only detect the boundary information of barrier with photoelectric sensor, wherein the radius of investigation of pathfinder
For 10 kms, the radius of investigation of photoelectric sensor is 15 kms.This emulation experiment shares 3 USV, and initial velocity magnitude is equal
For 40 sections, analyze for convenience, the starting point and terminal of 3 USV be it is determining, the static-obstacle thing in environment includes island,
Such as round barrier and irregular slalom, the polygon obstacle manually placed, such as triangle, quadrangle and pentagon, dynamic hinders
Hindering object is a small-sized equilateral triangle, and the maximum number of iterations of particle swarm optimization algorithm is 200 times, population scale 80,
The calculating time of algorithm is 292ms.Collision-avoidance planning process of 3 USV under entire environment is as shown in Figure 7.
Claims (8)
1. a kind of USV cluster collision-avoidance planning method based on Modified particle swarm optimization algorithm, which is characterized in that mainly include following
Step:
Step 1: the comprehensive ken model of USV is established according to the parameter characteristic of pathfinder and photoelectric sensor；
Step 2: building coordinate system；
Step 3: constructing environment model；
Step 4: Rolling optimal strategy and Modified particle swarm optimization algorithm designed for USV cluster collision-avoidance planning；
Step 5: the information input of information and target point that comprehensive sensor is detected is obtained into Modified particle swarm optimization algorithm
Obtain the navigation of USV subsequent time and the adjustment instruction of the speed of a ship or plane.
2. a kind of USV cluster collision-avoidance planning method based on Modified particle swarm optimization algorithm according to claim 1, special
Sign is, three coordinate systems: global context coordinate system X are established in step 2GOGYG, local coordinate system X based on USVUOUYUAnd base
In the local coordinate system X of the comprehensive perception information of sensorRORYR；
Global coordinate system XGOGYGUsing east northeast coordinate system, the lower left corner is coordinate origin, and X-axis is directed toward direct north, and Y-axis is directed toward just
Dong Fangxiang；
Local coordinate system is divided into hull coordinate system and sensor coordinate system, and the origin of hull coordinate system is located at the center of gravity of USV, leads
Boat radar and photoelectric sensor are mounted at a of hull center, and X-axis and Y-axis are parallel to the X-axis and Y of USV local coordinate system
Axis.
3. a kind of USV cluster collision-avoidance planning method based on Modified particle swarm optimization algorithm according to claim 1, special
Sign is, constructing environment model in step 3 specifically includes the following steps:
(3.1) data acquire: the sensing range of pathfinder and photoelectric sensor is 360 °, it is assumed that is sensed every in 1 ° of angle
Device obtains and can only obtain a range information, and angle is changed to 360 ° from 0 ° counterclockwise, i.e. the every scanning one of sensor
Secondary just to obtain 361 groups of perception data points, these data points include two parameters, and first parameter is the angle between point and sensor
Degree, the distance of second parameter point to sensor；
(3.2) perception data point is extracted as line segment feature: firstly, being divided into point set by way of region division multiple
Then data point set in each region is carried out linear fit again, and then is extracted as line segment feature by region；
(3.3) static-obstacle thing extruding: extruding only is carried out to its part to the extruding of static-obstacle thing, extruding mode is mainly divided
For following three kinds of situations: the extruding of the extruding of single line section, the extruding of adjacent segments and curve；
(3.4) trajectory predictions of dynamic barrier: dynamic barrier is puffed to a circle first, then according to autoregression mould
Type predicts position and direction of the dynamic barrier under global coordinate system.
4. a kind of USV cluster collision-avoidance planning method based on Modified particle swarm optimization algorithm according to claim 1, special
Sign is that step 4 specifically includes: USV during the motion, every time to integrate portion, the specific item punctuate regional planning agency path in the ken,
Then using the window in each comprehensive ken of Modified particle swarm optimization algorithm optimization, the improvement of particle swarm optimization algorithm includes used
The designs of property weight and Studying factors, the selection of fitness function.
5. a kind of USV cluster collision-avoidance planning method based on Modified particle swarm optimization algorithm according to claim 4, special
Sign is: USV during the motion, every time to integrate portion, the specific item punctuate regional planning agency path in the ken, will be held in motion process
Row following steps:
(4.1) USV obtains the environmental information in the comprehensive ken by self-contained sensor；
(4.2) one is cooked up using particle swarm algorithm from starting point to specific item punctuate to pass through without the optimal road of collision prevention；
(4.3) USV will move to specific item punctuate from starting point along the path cooked up；
(4.4) starting point, return step 1, until USV drives to global terminal are designated as with this specific item.
6. a kind of USV cluster collision-avoidance planning method based on Modified particle swarm optimization algorithm according to claim 4, special
Sign is: being taken based on the value of the inertia weight ω of the inertia adjustable strategies Curve guide impeller of tangent function, formula are as follows:
In formula, α=π/4, m are controlling elements, and k is current iteration number, kmaxFor maximum number of iterations；
It is taken based on Serial regulation stragetic innovation design learning factor c1And c2Value, formula are as follows:
K is current iteration number, k in formulamaxFor maximum number of iterations, c1s=2.5, c2s=1.0, c1e=1.0, c2e=2.5；
The selection of fitness function:
The starting point of known USV is (x0,y0), target point is (xn,yn), subpath points are (xi,yi), wherein i=0,1,2 ... n-
1, n, then fitness function is defined as:
The fitness function defined according to each corner of USV are as follows:
In conjunction with above-mentioned formula, USV fitness function total during collision-avoidance planning is obtained:
F=λ1f1+λ2f2。
7. a kind of USV cluster collision-avoidance planning method based on Modified particle swarm optimization algorithm according to claim 4 or 6,
It is characterized in that, particle swarm optimization algorithm specifically includes the following steps:
Step 1: the parameter in initialization particle swarm algorithm: ωs、ωe、c1s、c1e、c2s、c2e, population quantity PNUM, greatest iteration
Number kmax, enable k=0；
Step 2: the history optimal value of particle individual is set as pbest by initialization particle populations, the i.e. speed of particle and position,
Group's optimal value is set as gbest；
Step 3: in the evolution in the present age, the fitness function value f of each particle is calculated；
Step 4: the pbest and gbest of more new particle；
Step 5: the speed of more new particle and position；
Step 6: it if meeting termination condition, exports gbest and terminates, otherwise turn Step 5.
8. a kind of USV cluster collision-avoidance planning method based on Modified particle swarm optimization algorithm according to claim 7, special
Sign is that particle is updated to following formula:
Wherein, c1、c2For Studying factors, c under normal conditions1+c2≤ 4, r1、r2For the random number between [0,1], ω is inertia power
Weight, pidFor individual optimal value, gidFor global optimum, subscript " i ", " d " respectively indicate i-th of particle and i-th of particle dimension
Number.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910032577.3A CN109582027B (en) | 2019-01-14 | 2019-01-14 | Improved particle swarm optimization algorithm-based USV cluster collision avoidance planning method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910032577.3A CN109582027B (en) | 2019-01-14 | 2019-01-14 | Improved particle swarm optimization algorithm-based USV cluster collision avoidance planning method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109582027A true CN109582027A (en) | 2019-04-05 |
| CN109582027B CN109582027B (en) | 2022-02-22 |
Family
ID=65915068
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910032577.3A Active CN109582027B (en) | 2019-01-14 | 2019-01-14 | Improved particle swarm optimization algorithm-based USV cluster collision avoidance planning method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109582027B (en) |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110632941A (en) * | 2019-09-25 | 2019-12-31 | 北京理工大学 | Trajectory generation method for target tracking of unmanned aerial vehicle in complex environment |
| CN111124762A (en) * | 2019-12-30 | 2020-05-08 | 航天科工网络信息发展有限公司 | Dynamic copy placing method based on improved particle swarm optimization |
| CN111142532A (en) * | 2020-01-03 | 2020-05-12 | 三一汽车制造有限公司 | Unmanned road roller group obstacle avoidance method and system based on 5G network communication |
| CN111413966A (en) * | 2020-03-12 | 2020-07-14 | 天津大学 | Progressive model prediction unmanned planning tracking cooperative control method |
| CN111505935A (en) * | 2020-04-30 | 2020-08-07 | 内蒙古工业大学 | Automatic guided vehicle control method and system |
| CN111563657A (en) * | 2020-04-10 | 2020-08-21 | 福建电子口岸股份有限公司 | Method for solving port tug scheduling through ant colony algorithm combined with multi-dimensional strategy |
| CN112213948A (en) * | 2020-10-23 | 2021-01-12 | 武汉理工大学 | Ship navigational speed cooperative control method based on MPC |
| CN112231793A (en) * | 2020-09-14 | 2021-01-15 | 重庆大学产业技术研究院 | Automatic obstacle avoidance arrangement method for reinforcing steel bars in crossed area of reinforced concrete member based on binary particle swarm optimization algorithm |
| CN112405547A (en) * | 2020-11-30 | 2021-02-26 | 湖南科技大学 | Swarm robot multi-target searching method under unknown environment |
| CN112434378A (en) * | 2020-12-09 | 2021-03-02 | 扬州中远海运重工有限公司 | Design method for drawing interactive control by shipbuilding enterprise line chart |
| CN112947447A (en) * | 2021-02-10 | 2021-06-11 | 大连理工大学 | Water surface unmanned ship autonomous navigation method based on synchronous planning-control strategy |
| CN113625780A (en) * | 2021-10-12 | 2021-11-09 | 北京大学 | Distributed unmanned cluster cooperative motion path planning method capable of avoiding deadlock |
| CN114237254A (en) * | 2021-12-20 | 2022-03-25 | 中国海洋大学 | Unmanned ship rudder direction PID control design method based on Smith predictor and particle swarm algorithm |
| CN114326702A (en) * | 2021-11-25 | 2022-04-12 | 江苏科技大学 | Model prediction control-based inland river unmanned ship collision avoidance method |
| CN114844403A (en) * | 2022-03-11 | 2022-08-02 | 西北工业大学 | Motor driver PID parameter self-tuning method based on improved particle swarm optimization |
| CN116448115A (en) * | 2023-04-07 | 2023-07-18 | 连云港杰瑞科创园管理有限公司 | Unmanned ship probability distance map construction method based on navigation radar and photoelectricity |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8285655B1 (en) * | 2004-02-03 | 2012-10-09 | Hrl Laboratories, Llc | Method for object recongnition using multi-layered swarm sweep algorithms |
| CN103336526A (en) * | 2013-06-20 | 2013-10-02 | 苏州经贸职业技术学院 | Robot path planning method based on coevolution particle swarm rolling optimization |
| CN105717929A (en) * | 2016-04-29 | 2016-06-29 | 中国人民解放军国防科学技术大学 | Planning method for mixed path of mobile robot under multi-resolution barrier environment |
-
2019
- 2019-01-14 CN CN201910032577.3A patent/CN109582027B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8285655B1 (en) * | 2004-02-03 | 2012-10-09 | Hrl Laboratories, Llc | Method for object recongnition using multi-layered swarm sweep algorithms |
| CN103336526A (en) * | 2013-06-20 | 2013-10-02 | 苏州经贸职业技术学院 | Robot path planning method based on coevolution particle swarm rolling optimization |
| CN105717929A (en) * | 2016-04-29 | 2016-06-29 | 中国人民解放军国防科学技术大学 | Planning method for mixed path of mobile robot under multi-resolution barrier environment |
Non-Patent Citations (4)
| Title |
|---|
| MAC,THI THOA等: "A hierarchical global path planning approach for mobile robots based on multi-objective particle swarm optimization", 《APPLIED SOFT COMPUTING》 * |
| 吕太之 等: "采用粒子群优化和B样条曲线的改进可视图路径规划算法", 《华侨大学学报(自然科学版)》 * |
| 吴高超: "基于粒子群算法的路径规划问题研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
| 槐创锋 等: "基于自回归模型的移动机器人路径规划", 《计算机测量与控制》 * |
Cited By (27)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110632941A (en) * | 2019-09-25 | 2019-12-31 | 北京理工大学 | Trajectory generation method for target tracking of unmanned aerial vehicle in complex environment |
| CN111124762A (en) * | 2019-12-30 | 2020-05-08 | 航天科工网络信息发展有限公司 | Dynamic copy placing method based on improved particle swarm optimization |
| CN111124762B (en) * | 2019-12-30 | 2023-11-14 | 航天科工网络信息发展有限公司 | Dynamic copy placement method based on improved particle swarm optimization |
| CN111142532A (en) * | 2020-01-03 | 2020-05-12 | 三一汽车制造有限公司 | Unmanned road roller group obstacle avoidance method and system based on 5G network communication |
| CN111413966B (en) * | 2020-03-12 | 2022-04-29 | 天津大学 | Progressive model prediction unmanned planning tracking cooperative control method |
| CN111413966A (en) * | 2020-03-12 | 2020-07-14 | 天津大学 | Progressive model prediction unmanned planning tracking cooperative control method |
| CN111563657A (en) * | 2020-04-10 | 2020-08-21 | 福建电子口岸股份有限公司 | Method for solving port tug scheduling through ant colony algorithm combined with multi-dimensional strategy |
| CN111563657B (en) * | 2020-04-10 | 2022-11-15 | 福建电子口岸股份有限公司 | Method for solving port tug scheduling through ant colony algorithm combined with multi-dimensional strategy |
| CN111505935A (en) * | 2020-04-30 | 2020-08-07 | 内蒙古工业大学 | Automatic guided vehicle control method and system |
| CN111505935B (en) * | 2020-04-30 | 2022-10-18 | 内蒙古工业大学 | Automatic guided vehicle control method and system |
| CN112231793A (en) * | 2020-09-14 | 2021-01-15 | 重庆大学产业技术研究院 | Automatic obstacle avoidance arrangement method for reinforcing steel bars in crossed area of reinforced concrete member based on binary particle swarm optimization algorithm |
| CN112213948A (en) * | 2020-10-23 | 2021-01-12 | 武汉理工大学 | Ship navigational speed cooperative control method based on MPC |
| CN112213948B (en) * | 2020-10-23 | 2022-06-10 | 武汉理工大学 | Ship navigational speed cooperative control method based on MPC |
| CN112405547A (en) * | 2020-11-30 | 2021-02-26 | 湖南科技大学 | Swarm robot multi-target searching method under unknown environment |
| CN112434378B (en) * | 2020-12-09 | 2024-02-02 | 扬州中远海运重工有限公司 | Design method for drawing interactive control of line form of shipbuilding enterprise |
| CN112434378A (en) * | 2020-12-09 | 2021-03-02 | 扬州中远海运重工有限公司 | Design method for drawing interactive control by shipbuilding enterprise line chart |
| CN112947447A (en) * | 2021-02-10 | 2021-06-11 | 大连理工大学 | Water surface unmanned ship autonomous navigation method based on synchronous planning-control strategy |
| CN112947447B (en) * | 2021-02-10 | 2023-11-24 | 大连理工大学 | Water surface unmanned ship autonomous navigation method based on synchronous planning-control strategy |
| CN113625780A (en) * | 2021-10-12 | 2021-11-09 | 北京大学 | Distributed unmanned cluster cooperative motion path planning method capable of avoiding deadlock |
| CN113625780B (en) * | 2021-10-12 | 2022-01-28 | 北京大学 | Distributed unmanned cluster cooperative motion path planning method capable of avoiding deadlock |
| CN114326702B (en) * | 2021-11-25 | 2023-09-22 | 江苏科技大学 | Model predictive control-based inland unmanned ship collision prevention method |
| CN114326702A (en) * | 2021-11-25 | 2022-04-12 | 江苏科技大学 | Model prediction control-based inland river unmanned ship collision avoidance method |
| CN114237254B (en) * | 2021-12-20 | 2024-01-09 | 中国海洋大学 | Unmanned rudder PID control design method based on Smith predictor and particle swarm optimization |
| CN114237254A (en) * | 2021-12-20 | 2022-03-25 | 中国海洋大学 | Unmanned ship rudder direction PID control design method based on Smith predictor and particle swarm algorithm |
| CN114844403A (en) * | 2022-03-11 | 2022-08-02 | 西北工业大学 | Motor driver PID parameter self-tuning method based on improved particle swarm optimization |
| CN116448115A (en) * | 2023-04-07 | 2023-07-18 | 连云港杰瑞科创园管理有限公司 | Unmanned ship probability distance map construction method based on navigation radar and photoelectricity |
| CN116448115B (en) * | 2023-04-07 | 2024-03-19 | 连云港杰瑞科创园管理有限公司 | Unmanned ship probability distance map construction method based on navigation radar and photoelectricity |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109582027B (en) | 2022-02-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109582027A (en) | A kind of USV cluster collision-avoidance planning method based on Modified particle swarm optimization algorithm | |
| CN108564202B (en) | Unmanned ship route optimization method based on environment forecast information | |
| CN110632931B (en) | Mobile robot collision avoidance planning method based on deep reinforcement learning in dynamic environment | |
| CN110782481A (en) | Unmanned ship intelligent decision method and system | |
| CN109933067A (en) | A kind of unmanned boat collision prevention method based on genetic algorithm and particle swarm algorithm | |
| CN104899590A (en) | Visual target tracking method and system for unmanned aerial vehicle | |
| CN110906935B (en) | Unmanned ship path planning method | |
| CN109682378A (en) | A kind of unmanned plane indoor positioning and multi-object tracking method based entirely on visual information | |
| Shan et al. | LiDAR-based stable navigable region detection for unmanned surface vehicles | |
| CN112819255B (en) | Multi-criterion ship route determining method and device, computer equipment and readable storage medium | |
| CN114077256A (en) | Overwater unmanned ship path planning method | |
| Zhu et al. | AUV dynamic obstacle avoidance method based on improved PPO algorithm | |
| Santos et al. | Tracking of multi-obstacles with laser range data for autonomous vehicles | |
| CN111611869B (en) | End-to-end monocular vision obstacle avoidance method based on serial deep neural network | |
| CN108983181A (en) | A kind of radar marine target detection system of gunz optimizing | |
| Samal et al. | Closed-loop approach to perception in autonomous system | |
| CN116466701A (en) | Unmanned surface vessel energy consumption minimum track planning method | |
| CN115454061B (en) | Robot path obstacle avoidance method and system based on 3D technology | |
| CN116300913A (en) | Unmanned ship multi-constraint local path planning method based on visual information | |
| CN115690343A (en) | Robot laser radar scanning and mapping method based on visual following | |
| Hu et al. | Visual-Based Moving Target Tracking With Solar-Powered Fixed-Wing UAV: A New Learning-Based Approach | |
| Wang et al. | Research on Dynamic Planning Method of Collision Avoidance Route Based on Ship Kinematics | |
| CN117872349B (en) | Maneuvering extension target tracking method and maneuvering extension target tracking system | |
| CN113885522B (en) | Obstacle avoidance method, device and equipment for water surface robot and readable storage medium | |
| CN110716570B (en) | Relative field acquisition method based on ship cluster situation for open and busy water area |
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 |