CN104809529A - Multi-ship encounter collision prevention method for bacterial foraging optimization - Google Patents
Multi-ship encounter collision prevention method for bacterial foraging optimization Download PDFInfo
- Publication number
- CN104809529A CN104809529A CN201510242549.6A CN201510242549A CN104809529A CN 104809529 A CN104809529 A CN 104809529A CN 201510242549 A CN201510242549 A CN 201510242549A CN 104809529 A CN104809529 A CN 104809529A
- Authority
- CN
- China
- Prior art keywords
- ship
- collision prevention
- situation
- collision
- bacterium
- 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.)
- Pending
Links
Landscapes
- Traffic Control Systems (AREA)
Abstract
The invention belongs to the technical field of planning of automatic collision prevention paths of ships, and mainly relates to a multi-ship encounter collision prevention method for bacterial foraging optimization. The multi-ship encounter collision prevention method comprises the following steps: establishing a required simulation interface of a ship simulation test, and determining parameters of a ship and each target chip in multi-ship encounter collision prevention; judging an encounter situation and analyzing collision risks; establishing a target function (an optimization algorithm target function) of the multi-ship encounter collision prevention method; determining a key ship which needs to prevent collision and risk based on a grey correlation analysis method: determining an ideal effect sequence: calling an improved bacterial foraging algorithm to optimize the multi-ship encounter collision prevention paths; completing the encounter collision prevention sailing, re-sailing and restoring original courses. According to the multi-ship encounter collision prevention method, the quality of the bacteria is integrally analyzed based on a mean value and variance of the bacteria, and a last-time trended target function value is combined for judging bacteria for copy operation, so that the convergence rate and the searching precision of the algorithm are improved, and the efficiency of generating a multi-ship encounter collision prevention strategy is improved.
Description
Technical field
The invention belongs to ship automatic avoidance collision Path Planning Technique field, relate generally to a kind of bacterium look for food optimize multi-ship situation collision prevention method.
Background technology
Ship automatic avoidance collision refers to that boats and ships ride the sea in process to become can meet situation with other ship shape, determine to need to take collision prevention action by Risk-Degree of Collision analysis, then crewman is based on existing International Rules of the Road, and between ship when information acquisition and the autonomous exchange of ship internal information, obtain ship collision prevention best mode by optimisation strategy, thus generate " economy " and the collision prevention of " safety " navigation path.Auxiliary crewman makes Decision of Collision Avoidance fast, is conducive to the generation reducing marine accident.
Wherein, Decision of Collision Avoidance problem in multi-ship situation situation, because International Regulations for Preventing Collisions at Sea does not provide concrete collision prevention instruction, thus bring certain difficulty to the decision-making of Ship avoiding-bump problem and evaluation, therefore the navigation and vehicle controL field of multi-ship situation is still a very tired scabrous problem.Scholar's research both domestic and external and summarize a lot of multi-ship situation collision prevention method in the last few years, the still water level Ship avoiding-bump simulated based on optimum control controls (plum Chaoyang, Xiong Yong, Cai's Wei. the still water level Ship avoiding-bump based on optimum control controls research. and transport science and techonologies .2014.263 (2): 159-164.), intelligence aided decision collision prevention (Lilina under multi-ship situation to situation, Xiongzhennan, Renqingsheng.Generation and optimization of intelligent decisions formulti-ship collision avoidance.Information and Control.2003, 32 (2): 189-192.), Ship avoiding-bump simulation test (the Xue Yanzhuo of collision prevention emphasis ship algorithm, Wei Yi, Sun Miao. the Ship avoiding-bump based on collision prevention emphasis ship algorithm is simulated. Maritime Affairs University Of Dalian journal .2014.40 (1): 13-16.), and the Intelligent Collision Avoidance process (Zhou Wei of many ships under close quarters situation, Li Lina, Xiong Zhennan, poplar is deified. the Intelligent Collision Avoidance decision-making of many ships immediate danger situation and emulation. and Collects The American University's journal (natural science edition) .2010.15 (5): 347-352.), establish multi-ship situation collision prevention model, the multi-ship situation collision prevention achieving simulation is analyzed, but in the analysis of comprehensive multi-ship situation situation, the foundation of multi-ship situation risk factor Grade Model, certain distance is still applied with test in the aspects such as the realization of effective Ship avoiding-bump method.Therefore should study the new approaches of multi-ship situation ship collision prevention intelligent decision, to the labour intensity and the psychological burden that alleviate driver, there is realistic meaning.
The bacterial foraging algorithm proposed by K.M.Passino for 2002 is a kind of first heuristic search algorithm of simulating Escherichia coli foraging behavior, not only there is the ability of the search of the very strong overall situation, and the optimization problem of many real worlds can be solved efficiently, Marvin's credit in 2013, Yang Jiaxuan etc. Single ship anti collision air route optimize that (Marvin's is shone based on this algorithm realization, Yang Jiaxuan. based on the collision prevention air route optimizing research of bacterial foraging algorithm. Maritime Affairs University Of Dalian journal .2013, 39 (2): 21-24.), but it is slow to there is speed of convergence in algorithm realization process, easily be absorbed in local precocious, the shortcoming not high with solving precision.
Summary of the invention
The object of the present invention is to provide a kind of bacterium look for food optimize multi-ship situation collision prevention method.
The object of the present invention is achieved like this:
(1) set up the Simulation Interface needed for boats and ships l-G simulation test, and determine my ship and each object ship parameter of multi-ship situation collision prevention;
(2) situation can be met judge and Risk-Degree of Collision analysis:
(2.1) determine that between I ship and each object ship, situation situation is met in two two Conferences;
(2.2) if determine identical multi-ship situation situation between I ship and each object ship, namely my ship is all end-on with each object ship, overtakes, intersects, so perform collision prevention action;
(2.3) if different multi-ship situation situation, so according to coming bearing, situation is met in the meeting of classifying different, according to the order that hazard degree analysis determination collision prevention performs;
(2.4) if after completing collision prevention action, the multi-ship situation analysis of situation repeated execution of steps (2.2), (2.3) is met in the meeting that will redefine between I ship and other object ships of the change due to my ship course;
(2.5) situation is regarded as two ship Meeting Situations, and predict the possible action of other ship according to this, thus form multiple two ship Meeting Situations;
(2.6) after the danger of my ship can meet condition ceases, return to original course line, complete and restore navigation, wherein the Risk-Degree of Collision of my ship and each object ship
u
tTfor time collision risk and U
dTfor space collision risk factor;
(3) objective function of multi-ship situation collision prevention method is set up: optimized algorithm objective function is:
J
total path=min (J
1+ J
2+ J
i... ..J
n),
J
total path=min (J
1)+min (J
2)+min (J
i) ... ..min (J
n) represent that all collision prevention operations my n bar collision prevention path summation of ship navigation rear is minimum value, be equivalent to the bee-line sum completed after solving each execution collision prevention task, thus complete the collision prevention navigational duty of my ship based on collision prevention shortest path, wherein, d
sfor this ship distance to go after collision prevention, d
rfor this ship distance to go in the stage of restoring navigation, then min (J
i) meet min (J
i)=min (d
s+ d
r), restore navigation if perform in the process of collision prevention operation not relate at i-th time, so the shortest collision prevention path should be min{d
s; Otherwise be min{d
s+ d
r, the speed of a ship or plane of this ship is v, and dodging steering angle is c
s, the steering angle c in stage of restoring navigation
r, the time of dodging rear navigation is t
a, d in objective function
s, d
rmeet following expression process: d
s=t
a* v, d
r=t
a* c*sin c
s/ | sin c
r|, wherein 0 °≤c
s≤ 30 ° ,-35 °≤c
r≤-35 °, t
cPA1≤ t
a≤ 40min, d
cPA1>=d
g, d
cPA2>=d
g, wherein, d
cPA1and d
cPA2be respectively after collision prevention and recover course stylish meet distance most recently; d
gsize sets according to space risk factor null boundary; The time of dodging rear navigation is t
aat least equal the time to closest point of approach taking collision prevention action new afterwards;
(4) based on the dangerous boats and ships of Grey Incidence Analysis determination emphasis collision prevention: determine ideal effect sequence:
X
0n=(X
01,X
02,X
0j....X
0n),
The effect of situation sequence X of i-th object ship
in=(X
i1, X
i2, X
ij... .X
in), and sequence length is identical,
Calculate X
onx
ingrey absolute correlation degree ε
oi:
s
0and s
ifor calculating the temporal operator of correlativity,
Determine the numbering of emphasis collision prevention boats and ships in object ship;
(5) call improvement bacterial foraging algorithm and optimize multi-ship situation collision prevention path:
(5.1) parameter, array, the initialization of position and the setting of initial step length is completed;
(5.2) position of initialization bacterium in two-dimensional space, calculate initial objective function J [l] [k] [j] [i]=livefun, wherein, l, k, j, i are migration number of times, number of copy times, trend number of times, bacterium number, arrange current best objective function Jlast=J;
(5.3) in two-dimensional space, overturn Delta=2* ((double) rand ())/RAND_MAX-1.0, produce random direction cov+=pow (Delta, 2.0);
(5.4) within the scope of the maximum step-length of regulation, move about according to new direction and produce new function of position
p
to restore navigation angle=P+C*Delta/sqrt (cov) and objective function J [l] [k] [j] [i]=livefun ();
(5.5) find the rear objective function optimal particle solution of this chemotactic circulation, find globally optimal solution P
gbestwith single optimum solution P
lbestcorresponding particle position Jbest=J, what upgrade each bacterium next time according to particle cluster algorithm stirs direction
thus determine the position that next time tends to new direction P=P+V and determines;
(5.6) replicate run: by the target function value in all for each bacterium trend paths being fitted to the form of curve, solve its average and variance, analyze the quality of this bacterium on the whole and the target function value after combining last trend is combined according to by size, sequence Quicksort (Jhealth, S), S is the number of bacterium, and judgement is used for the bacterium of front 50% of replicate run;
(5.7) according to migrating probability, migration operation being performed to the bacterium after all copying, obtaining the random reposition in two-dimensional space,
P (collision prevention)=60+30* (2.0* ((double) rand ())/RAND_MAX-1,
(2.0* ((double) rand ())/RAND_MAX-1, performs and circulates, until meet iterations and solving precision next time P (restoring navigation)=-45+15*;
(6) complete this collision prevention navigation, restore navigation, recover original course.
T is calculated in described step (3)
aprocess is:
(1) if d
cPA1< d
g, so t
a=40min;
(2) at d
cPA1>=d
gcondition under, if meet t
cPA1> 0, so t
cPA1=t
afor initial, increase, until meet d with certain time search step value
cPA2>=d
g, determine t
a;
(3) at d
cPA1>=d
gcondition under, if meet t
cPA1< 0, so initial value t
a=0.05*60, and could not d met
cPA2>=d
g, t
aincrease to search for step value sometime.
Beneficial effect of the present invention is:
For problems of the prior art, one aspect of the present invention is perfect, the process that multi-ship situation collision prevention method realizes, the collision prevention process realizing " circulation is dodged " dangerous boats and ships based on grey relational grade analysis is proposed, and the collision prevention result of real-time assessment collision prevention each time action, thus take appropriate measures, make whole multi-ship situation collision prevention process smooth, complete, reliable.On the other hand, by particle cluster algorithm optimize bacterium look for food trend operation in bacterium optimum stir direction, simultaneously in replicate run process, form the matched curve of bacterium target function value, the quality of this bacterium is analyzed on the whole and the target function value after combining last trend combines the bacterium judged for replicate run based on its average and variance, thus improve convergence of algorithm speed and search precision, improve the efficiency of multi-ship situation collision prevention strategy generating.
Accompanying drawing explanation
Fig. 1 is the overall flow figure of multi-ship situation automatic Collision Avoidance paths planning method of the present invention;
Fig. 2 is the process flow diagram that emphasis collision prevention ship is determined in gray relational decision-making analysis of the present invention;
The optimization bacterial foraging algorithm process flow diagram that Fig. 3 adopts for multi-ship situation collision prevention method of the present invention;
Fig. 4 is the l-G simulation test realizing multi-ship situation collision prevention based on optimization bacterial foraging algorithm;
Fig. 5 is that bacterium is looked for food and optimizes the convergence correlation curve figure of bacterial foraging algorithm for Single ship anti collision path planning;
Fig. 6 is that bacterium is looked for food and optimizes the convergence correlation curve figure of bacterial foraging algorithm for Ship avoiding-bump path planning.
The mode of concrete enforcement
Below in conjunction with accompanying drawing, the present invention is described in more detail
The present invention proposes a kind of bacterium look for food optimize multi-ship situation collision prevention method, belong to ship collision prevention routeing technical field.By my ship for dodging premised on ship, the method comprises: obtain target ship information, judge to meet situation judgement between boats and ships, set up multi-ship situation collision prevention objective function, determine circulation emphasis collision prevention boats and ships, formulate based on the ship collision prevention scheme optimizing bacterial foraging algorithm, perform circulation collision prevention action, I restores navigation by ship, continues navigation according to original course line.The on the one hand trend operation in bacterial foraging algorithm based on particle group optimizing in the present invention, improve convergence of algorithm speed, avoid being absorbed in local optimum, have adjusted the judge criterion for copying bacterium on the other hand, the bacterium having optimal objective function value the next generation can be ensured, improve the search precision of algorithm, and algorithm after this optimization is used in multi-ship situation automatic Collision Avoidance path planning, by the judgement of situation can be met, based on the objective function of collision prevention shortest path, the dangerous boats and ships that emphasis is dodged are determined successively in conjunction with grey correlation analysis algorithm, provide the optimum collision prevention parameter of a sequence, finally, validity and the superiority of proposed method is demonstrated based on dynamic ship collision prevention l-G simulation test and constringency performance simulation analysis.
A kind of bacterium look for food optimize multi-ship situation collision prevention method, overall realization flow for: utilize the object ship information that ARPS (ARPA) obtains, determine the quantity of target boats and ships, the status information such as course and the speed of a ship or plane, thus determine that situation is met in the meeting of many ships, under the risk of collision prerequisite that this ship of prediction is possible with many ships, the boats and ships of emphasis collision prevention are determined according to above-mentioned gray relational decision-making analysis, and based on optimization bacterial foraging algorithm (OBFO) algorithmization avoidingcollis ionscheme, this collision prevention action is performed after determining opportunity of collision prevention and amplitude, complete the process once dodging object ship, and judge to meet dangerous because whether this action creates new meeting, if my ship then will be adjusted, whether can restore navigation at the appointed time and remove newly-generated danger, if can not clear, the scheme that other object ships perform collision prevention again be had precedence over, if can my ship, move on after having planned route, and situation is met in the meeting rejudged between I ship and other object ship, re-start circulation, until not dangerous target boats and ships, after my ship completes and restores navigation, move on according to original route.Bacterium look for food optimize a multi-ship situation collision prevention method, specifically comprise following step:
Step 1, set up Simulation Interface needed for boats and ships l-G simulation test, and determine my ship and each object ship parameter of multi-ship situation collision prevention.
Step 2, can meet situation judge and Risk-Degree of Collision analysis.Multi-ship situation Study on Trend process is: 1. grasp two ships in " rule " and can meet the judgement of situation, determines that between I ship and each object ship, situation situation is met in two two Conferences; If 2. determine identical multi-ship situation situation between I ship and each object ship (my ship is all end-on with each object ship, overtakes, intersects), so according to operating experience and the avoidingcollis ionscheme execution collision prevention action of " rule " crewman; If 3. different multi-ship situation situation, so according to coming bearing, situation is met in the meeting of classifying different, according to the order that hazard degree analysis determination collision prevention performs; If 4. after completing collision prevention action, the change due to my ship course will redefine meeting between I ship and other object ships and meet situation and repeat 2. multi-ship situation analysis 3..5. situation is regarded as two ship Meeting Situations, and predict the possible action of other ship according to this, thus form multiple two ship Meeting Situations; 6., after the danger of my ship can meet condition ceases, return to original course line, complete and restore navigation.Risk-Degree of Collision is analyzed: the Risk-Degree of Collision (CRI Collision Risk Index) of my ship and each object ship, with distance to closest point of approach (DCPA), the shortlyest can meet the time (TCPA), boats and ships spacing (D), carrys out the relating to parameters such as ship position (B), ship's speed ratio (K).In order to above-mentioned factor can be considered, pass through based on time collision risk (U in the present invention
tT) and space collision risk factor (U
dT) determine between last single boats and ships Risk-Degree of Collision
and only have and work as U
tTand U
dTjust there is ship collision danger when entirely non-vanishing, according to Risk-Degree of Collision evaluation rank, select in the present invention to call when CRI=0.6 and optimize bacterial foraging algorithm and generate automatic Collision Avoidance scheme.
Step 3, set up the objective function of multi-ship situation collision prevention method.I primarily accomplishes " safety " collision prevention by ship in the process completing Ship avoiding-bump path planning, will consider " economy " collision prevention simultaneously, and the optimized algorithm objective function set up in the present invention is J
total path=min (J
1+ J
2+ J
i... ..J
n), J
total path=min (J
1)+min (J
2)+min (J
i) ... ..min (J
n) represent that all collision prevention operations my n bar collision prevention path summation of ship navigation rear is minimum value, be equivalent to the bee-line sum completed after solving each execution collision prevention task, thus complete the collision prevention navigational duty of my ship based on collision prevention shortest path, wherein, d
sfor this ship distance to go after collision prevention, d
rfor this ship distance to go in the stage of restoring navigation, then min (J
i) meet min (J
i)=min (d
s+ d
r), restore navigation if perform in the process of collision prevention operation not relate at i-th time, so the shortest collision prevention path should be min{d
s; Otherwise be min{d
s+ d
r.And the speed of a ship or plane supposing this ship is v, dodging steering angle is c
s, the steering angle c in stage of restoring navigation
r, the time of dodging rear navigation is t
a.D in so above-mentioned objective function
s, d
rmeet following expression process: d
s=t
a* v, d
r=t
a* c*sin c
s/ | sin c
r|, and above-mentioned parameter will meet following constraint condition: 0 °≤c
s≤ 30 ° ,-35 °≤c
r≤-35 °, t
cPA1≤ t
a≤ 40min, d
cPA1>=d
g, d
cPA2>=d
g, wherein, d
cPA1and d
cPA2be respectively after collision prevention and recover course stylish meet distance most recently; d
gsize sets according to space risk factor null boundary, according to concrete simulated environment require can also arrange lower; The time of dodging rear navigation is t
aat least should equal the time to closest point of approach taking collision prevention action new afterwards.
Step 4, based on the dangerous boats and ships of Grey Incidence Analysis determination emphasis collision prevention.When 1 target boats and ships with more than 2 or 2 meets, form risk of collision, and regularly my ship is born when giving way responsibility, correctly determines that main target ship is the basis of Right Moment for Collision Avoidance Action and collision avoidance behaviors.For this reason, how to set up a scientific and rational main target ship model, based on the basis of gray relational decision-making determination main target ship in the present invention, and extend the influence factor of gray relational decision-making Risk-Degree of Collision, for decision objective efficiently utilizes given data, the error determining risk factor is reduced with DCPA, TCPA, D, SDP (safe meeting distance), K (ship's speed ratio).Thus determine ideal effect sequence X
0n=(X
01, X
02, X
0j... .X
0n), the effect of situation sequence X of i-th object ship
in=(X
i1, X
i2, X
ij... .X
in), and sequence length is identical, by calculating X
onx
ingrey absolute correlation degree ε
oi:
(s
0and s
ifor calculating the temporal operator of correlativity,
) thus determine the numbering of emphasis collision prevention boats and ships in object ship, lay the foundation for performing collision prevention task.
Step 5, call improve bacterial foraging algorithm optimize multi-ship situation collision prevention path.In Models of Decision-making in Ship Collision Avoidance process, turn to collision prevention to be a kind of collision prevention method adopting frequency the highest, from the local consideration of collision prevention situation in the present invention, obtain a series of best ship turning collision prevention amplitude satisfied condition based on optimization bacterial foraging algorithm.
Step 5.1: complete parameter, array, the initialization of position and the setting of initial step length, in the present invention, initial step length is 0.05.
Step 5.2: the position of initialization bacterium in two-dimensional space, calculate initial objective function J [l] [k] [j] [i]=livefun (calculating according to initial parameter ship collision prevention shortest path target function value) (wherein, l, k, j, i are migration number of times, number of copy times, trend number of times, bacterium number), and to arrange current best objective function be Jlast=J (preserving this numerical value).
Step 5.3: overturn Delta=2* ((double) rand ())/RAND_MAX-1.0 in two-dimensional space, produces random direction cov+=pow (Delta, 2.0).
Step 5.4: within the scope of the maximum step-length of regulation, moves about according to new direction and produces new position (collision prevention and rudder angles of restoring navigation) function P
collision prevention angle=P+C*Delta/sqrt (cov), P
to restore navigation angle=P+C*Delta/sqrt (cov) (producing new position) and objective function J [l] [k] [j] [i]=livefun () (calculating target fitness value function).
Step 5.5: find the rear objective function optimal particle solution of this chemotactic circulation, find globally optimal solution P
gbestand list
Individual optimum solution P
lbestcorresponding particle position Jbest=J, upgrades next time according to particle cluster algorithm that each is thin
Bacterium stir direction V=w*V+C
1* R
1* (P
lbest-P)+C
2* R
2* (P
gbest-P), thus determine that next time becomes
To new direction P=P+V and the position determined, improve convergence of algorithm speed.
Step 5.6: replicate run.In order to the bacterium having optimal objective function value the next generation can be ensured, by the target function value in all for each bacterium trend paths being fitted to the form of curve, solve its average and variance, analyze the quality of this bacterium on the whole and the target function value after combining last trend is combined according to by size
Sequence Quicksort (Jhealth, S) (S is the number of bacterium), judgement is used for the bacterium of front 50% of replicate run, improves the search precision of algorithm.
Step 5.7: according to migrating probability, migration operation being performed to the bacterium after all copying, obtaining the random reposition in two-dimensional space.
P (collision prevention)=60+30* (2.0* ((double) rand ())/RAND_MAX-1,
P (restoring navigation)=-45+15* (2.0* ((double) rand ())/RAND_MAX-1, and then execution circulates, until meet iterations and solving precision next time.
Step 6, complete all collision preventions navigation, restore navigation, recover original course.
The present invention proposes a kind of bacterium look for food optimize multi-ship situation collision prevention method, specifically comprise following step as shown in Figure 1:
Step 1, foundation test combine and initialization simulation parameter.Set up Simulation Interface needed for boats and ships l-G simulation test as shown in Figure 4, multi-ship situation collision prevention test adopts the development platform of Visual Basic2010, interactively program design in operational process, user's intervention program at any time, the motion state of control break boats and ships.The modules that system realizes comprises main display interface, the judgement meeting situation, Decision of Collision Avoidance, risk factor calculating, the design of avoidingcollis ionscheme, the opportunity of restoring navigation determine.Under whole system is operated in real-time status, motion conditions and the collision prevention process of this ship and object ship can be shown intuitively.The situation that this ship is give-way vessel is only considered when can meet emulation for three ships during open waters visibility good in the present invention.Emulation basic parameter and the evaluate parameter of my ship and object ship are set to:
Course | The speed of a ship or plane | Longitude | Latitude | DCPA | TCPA | CRI | |
My ship | 0 | 8 | 0 | 0 | |||
Object ship 1 | 300 | 12 | 6 | 0 | 1.13 | 0.56 | 0.75 |
Object ship 2 | 250 | 17 | 6 | 6 | 0.61 | 0.4 | 1.00 |
Step 2, can meet situation judge and Risk-Degree of Collision analysis.Multi-ship situation Study on Trend process is: 1. grasp two ships in " rule " and can meet the judgement of situation, determines that between I ship and each object ship, situation situation is met in two two Conferences; If 2. determine identical multi-ship situation situation between I ship and each object ship (my ship is all end-on with each object ship, overtakes, intersects), so according to operating experience and the avoidingcollis ionscheme execution collision prevention action of " rule " crewman; If 3. different multi-ship situation situation, so according to coming bearing, situation is met in the meeting of classifying different, according to the order that hazard degree analysis determination collision prevention performs; If 4. after completing collision prevention action, the change due to my ship course will redefine meeting between I ship and other object ships and meet situation and repeat 2. multi-ship situation analysis 3..5. situation is regarded as two ship Meeting Situations, and predict the possible action of other ship according to this, thus form multiple two ship Meeting Situations; 6., after the danger of my ship can meet condition ceases, return to original course line, complete and restore navigation.Risk-Degree of Collision is analyzed: the Risk-Degree of Collision (CRI Collision Risk Index) of my ship and each object ship, with distance to closest point of approach (DCPA), the shortlyest can meet the time (TCPA), boats and ships spacing (D), carrys out the relating to parameters such as ship position (B), ship's speed ratio (K).In order to above-mentioned factor can be considered, pass through based on time collision risk (U in the present invention
tT) and space collision risk factor (U
dT) determine between last single boats and ships Risk-Degree of Collision
and only have and work as U
tTand U
dTjust there is ship collision danger when entirely non-vanishing, according to Risk-Degree of Collision evaluation rank, select in the present invention to call when CRI=0.6 and optimize bacterial foraging algorithm and generate automatic Collision Avoidance scheme.
Step 3, set up the objective function of multi-ship situation collision prevention method.I primarily accomplishes " safety " collision prevention by ship in the process completing Ship avoiding-bump path planning, will consider " economy " collision prevention simultaneously, and the optimized algorithm objective function set up in the present invention is J
total path=min (J
1+ J
2+ J
i... ..J
n), J
total path=min (J
1)+min (J
2)+min (J
i) ... ..min (J
n) represent that all collision prevention operations my n bar collision prevention path summation of ship navigation rear is minimum value, be equivalent to the bee-line sum completed after solving each execution collision prevention task, thus complete the collision prevention navigational duty of my ship based on collision prevention shortest path, wherein, d
sfor this ship distance to go after collision prevention, d
rfor this ship distance to go in the stage of restoring navigation, then min (J
i) meet min (J
i)=min (d
s+ d
r), restore navigation if perform in the process of collision prevention operation not relate at i-th time, so the shortest collision prevention path should be min{d
s; Otherwise be min{d
s+ d
r.And the speed of a ship or plane supposing this ship is v, dodging steering angle is c
s, the steering angle c in stage of restoring navigation
r, the time of dodging rear navigation is t
a.D in so above-mentioned objective function
s, d
rmeet following expression process: d
s=t
a* v, d
r=t
a* c*sinc
s/ | sinc
r|, and above-mentioned parameter will meet following constraint condition: 0 °≤c
s≤ 30 ° ,-35 °≤c
r≤-35 °, t
cPA1≤ t
a≤ 40min, d
cPA1>=d
g, d
cPA2>=d
g, wherein, d
cPA1and d
cPA2be respectively after collision prevention and recover course stylish meet distance most recently; d
gsize sets according to space risk factor null boundary, according to concrete simulated environment require can also arrange lower; The time of dodging rear navigation is t
aat least should equal the time to closest point of approach taking collision prevention action new afterwards.
Step 4, based on the dangerous boats and ships of Grey Incidence Analysis determination emphasis collision prevention.Wherein grey correlation hazard degree analysis process as shown in Figure 2.First the target ships quantity of decision-making will be determined, and in order to efficiently utilize given data in decision objective the present invention, reduce and determine the error of risk factor with DCPA, TCPA, D, SDP (safe meeting distance), K (ship's speed ratio) for decision objective, thus calculate the effect of situation sequence matrix of each target and the average item matrix of each target effect of situation sequence matrix, and then determine the effect of situation sequence X of i-th object ship
in=(X
i1, X
i2, X
ij... .X
in) and ideal effect sequence X
0n=(X
01, X
02, X
j0... .X
n0() wherein, n is for feeling target number) and two sequence lengths are identical, by calculating X
onx
ingrey absolute correlation degree ε
oi:
(s
0and s
ifor calculating the temporal operator of correlativity,
) thus determine the numbering of emphasis collision prevention boats and ships in object ship, lay the foundation for performing collision prevention task.
Step 5, call improve bacterial foraging algorithm optimize multi-ship situation collision prevention path.In Models of Decision-making in Ship Collision Avoidance process, collision prevention is turned to be a kind of collision prevention method adopting frequency the highest, from the local consideration of collision prevention situation in the present invention, obtain a series of best ship turning collision prevention amplitude satisfied condition as shown in Figure 3 based on optimization bacterial foraging algorithm.
Step 5.1: complete parameter, array, the initialization of position and the setting of initial step length, in the present invention, the initialization of bacterial foraging algorithm parameter comprises the dimension P=2 of search volume; Quantity S=26 for the bacterium searched for is set; The times N of trend
c=50, initial step length is 0.05, and travelling maximum step-length requires that general value is between 3 ~ 8, and search effect is relatively good, and emulating value is herein N
s=4; N
re=4 number of times copied; N
ed=2 migration number of times; The span of migration probability is generally between 0.05 ~ 0.3, and emulating value in literary composition is P
ed=0.25; C
1, C
2, R
1, R
2for the stray parameter of particle cluster algorithm.
Step 5.2: the position of initialization bacterium in two-dimensional space, calculate initial objective function J [l] [k] [j] [i]=livefun (calculating according to initial parameter ship collision prevention shortest path target function value) (wherein, l, k, j, i are migration number of times, number of copy times, trend number of times, bacterium number), and to arrange current best objective function be Jlast=J (preserving this numerical value).
Step 5.3: overturn Delta=2* ((double) rand ())/RAND_MAX-1.0 in two-dimensional space, produces random direction cov+=pow (Delta, 2.0).
Step 5.4: within the scope of the maximum step-length of regulation, moves about according to new direction and produces new position (collision prevention and rudder angles of restoring navigation) function P=P+C*Delta/sqrt (cov) (producing new position) and objective function J [l] [k] [j] [i]=livefun () (calculating target fitness value function)
Step 5.5: find the rear objective function optimal particle solution of this chemotactic circulation, find globally optimal solution P
gbestwith single optimum solution P
lbestcorresponding particle position Jbest=J, what upgrade each bacterium next time according to particle cluster algorithm stirs direction
select the position that next time tends to new direction P=P+V and determines, improve convergence of algorithm speed.
Step 5.6: replicate run.In order to the bacterium having optimal objective function value the next generation can be ensured, by the target function value in all for each bacterium trend paths being fitted to the form of curve, solve its average and variance, analyze the quality of this bacterium on the whole and the target function value after combining last trend is combined according to by size
Sequence Quicksort (Jhealth, S), judgement is used for the bacterium of front 50% of replicate run, improves the search precision of algorithm.
Step 5.7: copying rear execution migration operation according to migrating probability to all, obtaining the random reposition in two-dimensional space.P (collision prevention)=60+30* (2.0* ((double) rand ())/RAND_MAX-1, P (restoring navigation)=-45+15* (2.0* ((double) rand ())/RAND_MAX-1, and then execution circulates, until meet iterations and solving precision next time.
Step 6, complete all collision preventions navigation, restore navigation, recover original course.
Fig. 4 is that object ship and my ship intersection can meet situation, and the implementation procedure of two object ships dodged continuously by my ship, generates based on OBFO algorithm the scheme that multiple object ship dodged by my ship.Demonstrate the validity of the method in the present invention.
Fig. 5 and Fig. 6 is bacterial foraging algorithm (BFO) and optimizes the speed of convergence correlation curve that bacterial foraging algorithm (OBFO) solves objective function respectively in single ship and many ships automatic Collision Avoidance process, wherein, transverse axis represents iterations, the longitudinal axis represents target function value, can obtain from figure, in identical iterations situation, OBFO convergence of algorithm speed is obviously better than BFO algorithm, especially in many ships realize, therefore the superiority of OBFO algorithm is demonstrated, use it for Ship avoiding-bump and effectively can shorten the time generating ship collision prevention strategy, improve the efficiency of ship collision prevention.
Claims (2)
1. bacterium look for food optimize a multi-ship situation collision prevention method, it is characterized in that, comprise the following steps:
(1) set up the Simulation Interface needed for boats and ships l-G simulation test, and determine my ship and each object ship parameter of multi-ship situation collision prevention;
(2) situation can be met judge and Risk-Degree of Collision analysis:
(2.1) determine that between I ship and each object ship, situation situation is met in two two Conferences;
(2.2) if determine identical multi-ship situation situation between I ship and each object ship, namely my ship is all end-on with each object ship, overtakes, intersects, so perform collision prevention action;
(2.3) if different multi-ship situation situation, so according to coming bearing, situation is met in the meeting of classifying different, according to the order that hazard degree analysis determination collision prevention performs;
(2.4) if after completing collision prevention action, the multi-ship situation analysis of situation repeated execution of steps (2.2), (2.3) is met in the meeting that will redefine between I ship and other object ships of the change due to my ship course;
(2.5) situation is regarded as two ship Meeting Situations, and predict the possible action of other ship according to this, thus form multiple two ship Meeting Situations;
(2.6) after the danger of my ship can meet condition ceases, return to original course line, complete and restore navigation, wherein the Risk-Degree of Collision of my ship and each object ship
u
tTfor time collision risk and U
dTfor space collision risk factor;
(3) objective function of multi-ship situation collision prevention method is set up: optimized algorithm objective function is:
J
total path=min (J
1+ J
2+ J
i... ..J
n),
J
total path=min (J
1)+min (J
2)+min (J
i) ... ..min (J
n) represent that all collision prevention operations my n bar collision prevention path summation of ship navigation rear is minimum value, be equivalent to the bee-line sum completed after solving each execution collision prevention task, thus complete the collision prevention navigational duty of my ship based on collision prevention shortest path, wherein, d
sfor this ship distance to go after collision prevention, d
rfor this ship distance to go in the stage of restoring navigation, then min (J
i) meet min (J
i)=min (d
s+ d
r), restore navigation if perform in the process of collision prevention operation not relate at i-th time, so the shortest collision prevention path should be min{d
s; Otherwise be min{d
s+ d
r, the speed of a ship or plane of this ship is v, and dodging steering angle is c
s, the steering angle c in stage of restoring navigation
r, the time of dodging rear navigation is t
a, d in objective function
s, d
rmeet following expression process: d
s=t
a* v, d
r=t
a* c*sinc
s/ | sinc
r|, wherein 0 °≤c
s≤ 30 ° ,-35 °≤c
r≤-35 °, t
cPA1≤ t
a≤ 40min, d
cPA1>=d
g, d
cPA2>=d
g, wherein, d
cPA1and d
cPA2be respectively after collision prevention and recover course stylish meet distance most recently; d
gsize sets according to space risk factor null boundary; The time of dodging rear navigation is t
aat least equal the time to closest point of approach taking collision prevention action new afterwards;
(4) based on the dangerous boats and ships of Grey Incidence Analysis determination emphasis collision prevention: determine ideal effect sequence:
X
0n=(X
01,X
02,X
0j....X
0n),
The effect of situation sequence X of i-th object ship
in=(X
i1, X
i2, X
ij... .X
in), and sequence length is identical,
Calculate X
onx
ingrey absolute correlation degree ε
oi:
s
0and s
ifor calculating the temporal operator of correlativity,
Determine the numbering of emphasis collision prevention boats and ships in object ship;
(5) call improvement bacterial foraging algorithm and optimize multi-ship situation collision prevention path:
(5.1) parameter, array, the initialization of position and the setting of initial step length is completed;
(5.2) position of initialization bacterium in two-dimensional space, calculate initial objective function J [l] [k] [j] [i]=livefun, wherein, l, k, j, i are migration number of times, number of copy times, trend number of times, bacterium number, arrange current best objective function Jlast=J;
(5.3) in two-dimensional space, overturn Delta=2* ((double) rand ())/RAND_MAX-1.0, produce random direction cov+=pow (Delta, 2.0);
(5.4) within the scope of the maximum step-length of regulation, move about according to new direction and produce new function of position P
collision prevention angle=P+C*Delta/sqrt (cov), P
to restore navigation angle=P+C*Delta/sqrt (cov) and objective function J [l] [k] [j] [i]=livefun ();
(5.5) find the rear objective function optimal particle solution of this chemotactic circulation, find globally optimal solution P
gbestwith single optimum solution P
lbestcorresponding particle position Jbest=J, what upgrade each bacterium next time according to particle cluster algorithm stirs direction V=w*V+C
1* R
1* (P
lbest-P)+C
2* R
2* (P
gbest-P), thus determine the position that next time tends to new direction P=P+V and determines;
(5.6) replicate run: by the target function value in all for each bacterium trend paths being fitted to the form of curve, solve its average and variance, analyze the quality of this bacterium on the whole and the target function value after combining last trend is combined according to by size, sequence Quicksort (Jhealth, S), S is the number of bacterium, and judgement is used for the bacterium of front 50% of replicate run;
(5.7) according to migrating probability, migration operation being performed to the bacterium after all copying, obtaining the random reposition in two-dimensional space,
P (collision prevention)=60+30* (2.0* ((double) rand ())/RAND_MAX-1,
P (restoring navigation)=-45+15* (2.0* ((double) rand ())/RAND_MAX-1,
Perform and circulate, until meet iterations and solving precision next time;
(6) complete this collision prevention navigation, restore navigation, recover original course.
2. to look for food the multi-ship situation collision prevention method optimized according to a kind of bacterium shown in claim 1, it is characterized in that: calculate t in described step (3)
aprocess is:
(1) if d
cPA1< d
g, so t
a=40min;
(2) at d
cPA1>=d
gcondition under, if meet t
cPA1> 0, so t
cPA1=t
afor initial, increase, until meet d with certain time search step value
cPA2>=d
g, determine t
a;
(3) at d
cPA1>=d
gcondition under, if meet t
cPA1< 0, so initial value t
a=0.05*60, and could not d met
cPA2>=d
g, t
aincrease to search for step value sometime.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510242549.6A CN104809529A (en) | 2015-05-13 | 2015-05-13 | Multi-ship encounter collision prevention method for bacterial foraging optimization |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510242549.6A CN104809529A (en) | 2015-05-13 | 2015-05-13 | Multi-ship encounter collision prevention method for bacterial foraging optimization |
Publications (1)
Publication Number | Publication Date |
---|---|
CN104809529A true CN104809529A (en) | 2015-07-29 |
Family
ID=53694338
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201510242549.6A Pending CN104809529A (en) | 2015-05-13 | 2015-05-13 | Multi-ship encounter collision prevention method for bacterial foraging optimization |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN104809529A (en) |
Cited By (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109190170A (en) * | 2018-08-02 | 2019-01-11 | 哈尔滨工程大学 | A kind of Collision Risk Index modeling method based on cloud models theory |
CN109976342A (en) * | 2019-03-22 | 2019-07-05 | 智慧航海(青岛)科技有限公司 | A kind of intelligent ship multi-ship situation collision prevention method of News Search |
CN110164186A (en) * | 2019-06-18 | 2019-08-23 | 武汉理工大学 | Ship passes through confluence waters auxiliary collision avoidance system and method |
CN110164185A (en) * | 2019-06-05 | 2019-08-23 | 大连海事大学 | A kind of ship collision prevention system based on AIS |
CN110400491A (en) * | 2019-06-10 | 2019-11-01 | 北京海兰信数据科技股份有限公司 | A kind of Open sea area multiple target auxiliary Decision of Collision Avoidance method and decision system |
CN110648556A (en) * | 2019-10-29 | 2020-01-03 | 青岛科技大学 | Ship collision avoidance method based on ship collision avoidance characteristic |
CN110703759A (en) * | 2019-10-29 | 2020-01-17 | 青岛科技大学 | Ship collision prevention processing method for multi-ship game |
CN111260963A (en) * | 2020-03-11 | 2020-06-09 | 武汉理工大学 | Multi-ship meeting collision avoidance method |
CN111538332A (en) * | 2020-05-07 | 2020-08-14 | 湖南国天电子科技有限公司 | Automatic track planning method for unmanned ship |
CN111831048A (en) * | 2020-06-18 | 2020-10-27 | 广东工业大学 | Optimization method for photovoltaic array |
CN112100917A (en) * | 2020-09-14 | 2020-12-18 | 中国船级社 | Intelligent ship collision avoidance simulation test system and method based on expert confrontation system |
CN112580261A (en) * | 2020-12-23 | 2021-03-30 | 武汉理工大学 | Multi-ship collision avoidance decision method and device |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103471592A (en) * | 2013-06-08 | 2013-12-25 | 哈尔滨工程大学 | Multi-unmanned aerial vehicle route planning method based on bee colony collaborative foraging algorithm |
US20140100777A1 (en) * | 2008-06-30 | 2014-04-10 | Autonomous Solutions, Inc. | Vehicle dispatching method and system |
-
2015
- 2015-05-13 CN CN201510242549.6A patent/CN104809529A/en active Pending
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20140100777A1 (en) * | 2008-06-30 | 2014-04-10 | Autonomous Solutions, Inc. | Vehicle dispatching method and system |
CN103471592A (en) * | 2013-06-08 | 2013-12-25 | 哈尔滨工程大学 | Multi-unmanned aerial vehicle route planning method based on bee colony collaborative foraging algorithm |
Non-Patent Citations (3)
Title |
---|
LIU HONGDAN等: "The Application Research with Particle Swarm Bacterial Foraging Intelligent Algorithm in Ship Collision Avoidance", 《INTERNATIONAL CONFERENCE ON MECHATRONICS, CONTROL AND ELECTRONIC ENGINEERING(MCE2014),DOI:10.2991/MCE-14.2014.15》 * |
刘静等: "灰色关联决策在重点避让船选择中的应用研究", 《船舶工程》 * |
黄跃华 等: "基于灰色综合关联度的船舶碰撞事故因子权重分析", 《天津职业院校联合学报》 * |
Cited By (18)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109190170A (en) * | 2018-08-02 | 2019-01-11 | 哈尔滨工程大学 | A kind of Collision Risk Index modeling method based on cloud models theory |
CN109976342A (en) * | 2019-03-22 | 2019-07-05 | 智慧航海(青岛)科技有限公司 | A kind of intelligent ship multi-ship situation collision prevention method of News Search |
CN110164185A (en) * | 2019-06-05 | 2019-08-23 | 大连海事大学 | A kind of ship collision prevention system based on AIS |
CN110400491A (en) * | 2019-06-10 | 2019-11-01 | 北京海兰信数据科技股份有限公司 | A kind of Open sea area multiple target auxiliary Decision of Collision Avoidance method and decision system |
CN110164186A (en) * | 2019-06-18 | 2019-08-23 | 武汉理工大学 | Ship passes through confluence waters auxiliary collision avoidance system and method |
CN110648556A (en) * | 2019-10-29 | 2020-01-03 | 青岛科技大学 | Ship collision avoidance method based on ship collision avoidance characteristic |
CN110703759A (en) * | 2019-10-29 | 2020-01-17 | 青岛科技大学 | Ship collision prevention processing method for multi-ship game |
CN110703759B (en) * | 2019-10-29 | 2022-08-09 | 青岛科技大学 | Ship collision prevention processing method for multi-ship game |
CN111260963B (en) * | 2020-03-11 | 2021-02-26 | 武汉理工大学 | Multi-ship meeting collision avoidance method |
CN111260963A (en) * | 2020-03-11 | 2020-06-09 | 武汉理工大学 | Multi-ship meeting collision avoidance method |
CN111538332B (en) * | 2020-05-07 | 2021-01-05 | 湖南国天电子科技有限公司 | Automatic track planning method for unmanned ship |
CN111538332A (en) * | 2020-05-07 | 2020-08-14 | 湖南国天电子科技有限公司 | Automatic track planning method for unmanned ship |
CN111831048A (en) * | 2020-06-18 | 2020-10-27 | 广东工业大学 | Optimization method for photovoltaic array |
CN111831048B (en) * | 2020-06-18 | 2022-08-19 | 广东工业大学 | Optimization method for photovoltaic array |
CN112100917A (en) * | 2020-09-14 | 2020-12-18 | 中国船级社 | Intelligent ship collision avoidance simulation test system and method based on expert confrontation system |
CN112100917B (en) * | 2020-09-14 | 2023-12-22 | 中国船级社 | Expert countermeasure system-based intelligent ship collision avoidance simulation test system and method |
CN112580261A (en) * | 2020-12-23 | 2021-03-30 | 武汉理工大学 | Multi-ship collision avoidance decision method and device |
CN112580261B (en) * | 2020-12-23 | 2022-10-14 | 武汉理工大学 | Multi-ship collision avoidance decision method and device |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN104809529A (en) | Multi-ship encounter collision prevention method for bacterial foraging optimization | |
Chen et al. | A knowledge-free path planning approach for smart ships based on reinforcement learning | |
CN109540151A (en) | A kind of AUV three-dimensional path planning method based on intensified learning | |
Tsou et al. | Decision support from genetic algorithms for ship collision avoidance route planning and alerts | |
CN110083155B (en) | Machine learning method for realizing ship anthropomorphic intelligent collision avoidance decision | |
CN109597417A (en) | A kind of more USV multi-agent synergy collision-avoidance planning methods based on collision prevention criterion | |
WO2019076044A1 (en) | Mobile robot local motion planning method and apparatus and computer storage medium | |
CN111063218A (en) | Ship collision avoidance decision method | |
CN108319293A (en) | A kind of UUV Realtime collision free planing methods based on LSTM networks | |
CN108279692A (en) | A kind of UUV dynamic programming methods based on LSTM-RNN | |
Wang et al. | Local path optimization method for unmanned ship based on particle swarm acceleration calculation and dynamic optimal control | |
Wang et al. | Cooperative collision avoidance for unmanned surface vehicles based on improved genetic algorithm | |
Zheng et al. | A Decision‐Making Method for Ship Collision Avoidance Based on Improved Cultural Particle Swarm | |
Xu et al. | Ship Trajectory online prediction based on BP neural network algorithm | |
CN111338356A (en) | Multi-target unmanned ship collision avoidance path planning method for improving distributed genetic algorithm | |
Porres et al. | Scenario-based testing of a ship collision avoidance system | |
CN113032896B (en) | Anti-collision auxiliary decision-making method based on preference of ship driver | |
CN112612290A (en) | Underwater vehicle three-dimensional multi-task path planning method considering ocean currents | |
Vagale et al. | Evaluation of path planning algorithms of autonomous surface vehicles based on safety and collision risk assessment | |
Yang et al. | Improved reinforcement learning for collision-free local path planning of dynamic obstacle | |
Wu et al. | Deep reinforcement learning with dynamic window approach based collision avoidance path planning for maritime autonomous surface ships | |
CN104182613A (en) | Method for building ship electric power plant fault diagnosis petri net model based on rough set | |
CN115790608B (en) | AUV path planning algorithm and device based on reinforcement learning | |
CN113239520B (en) | Near-water three-dimensional underwater topography environment modeling method | |
Xiao et al. | Artificial force fields for multi-agent simulations of maritime traffic and risk estimation |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
EXSB | Decision made by sipo to initiate substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20150729 |
|
RJ01 | Rejection of invention patent application after publication |