CN114518754A - Multi-agent pursuit problem modeling and trapping strategy generation method - Google Patents
Multi-agent pursuit problem modeling and trapping strategy generation method Download PDFInfo
- Publication number
- CN114518754A CN114518754A CN202210104867.6A CN202210104867A CN114518754A CN 114518754 A CN114518754 A CN 114518754A CN 202210104867 A CN202210104867 A CN 202210104867A CN 114518754 A CN114518754 A CN 114518754A
- Authority
- CN
- China
- Prior art keywords
- chaser
- escaper
- target
- agent
- pursuit
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 51
- 238000004364 calculation method Methods 0.000 claims abstract description 8
- 230000004888 barrier function Effects 0.000 claims abstract description 6
- 238000000638 solvent extraction Methods 0.000 claims abstract description 5
- 230000003068 static effect Effects 0.000 claims description 12
- 238000010586 diagram Methods 0.000 claims description 3
- 230000007547 defect Effects 0.000 abstract description 2
- 150000001875 compounds Chemical class 0.000 description 2
- 230000005484 gravity Effects 0.000 description 2
- 230000003042 antagnostic effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000001788 irregular Effects 0.000 description 1
- 238000005192 partition Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course or altitude of land, water, air, or space vehicles, e.g. automatic pilot
- G05D1/02—Control of position or course in two dimensions
- G05D1/021—Control of position or course in two dimensions specially adapted to land vehicles
- G05D1/0259—Control of position or course in two dimensions specially adapted to land vehicles using magnetic or electromagnetic means
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course or altitude of land, water, air, or space vehicles, e.g. automatic pilot
- G05D1/02—Control of position or course in two dimensions
- G05D1/021—Control of position or course in two dimensions specially adapted to land vehicles
- G05D1/0276—Control of position or course in two dimensions specially adapted to land vehicles using signals provided by a source external to the vehicle
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N5/00—Computing arrangements using knowledge-based models
- G06N5/04—Inference or reasoning models
- G06N5/042—Backward inferencing
Abstract
In order to solve the defects that the existing trapping strategy cannot truly reflect the actual pursuit situation and the problem that the existing trapping strategy considering the environment with obstacles is difficult to solve when the intelligent agent is large in scale, the invention provides a modeling and trapping strategy generation method for the pursuit problem of multiple intelligent agents. When modeling is carried out on the multi-agent pursuit problem, the invention comprehensively considers the situations of obstacles and exits in a real scene; each chaser adjusts the enclosure target point in real time according to the situation change of the escaper, and receives the repulsive force from the barrier while enclosing, the closer the distance to the barrier, the larger the repulsive force, thereby avoiding the barrier and enclosing the escaper, and the chaser is particularly suitable for the enclosure task in a real complex scene; and carrying out Voronoi partitioning on the game environment, wherein each chaser takes a Voronoi unit of a minimum escaper as a target, the factors needed to be considered for decision making are less, the calculation is only carried out in a low-dimensional configuration space of a single intelligent agent, and the solution is simple.
Description
Technical Field
The invention relates to a modeling and trapping strategy generation method for a multi-agent pursuit problem.
Background
The multi-agent pursuit problem is that in a pursuit united system consisting of a plurality of mobile robots, the pursuit of one escaper or a united body consisting of a plurality of escapers is completed by applying corresponding motion strategies to each pursuer. The behavior between the chaser and the escaper is antagonistic, and for the constantly changing chasing situation, each intelligent agent must know the dynamically changing environment in real time so as to judge the current chasing situation, reasonably process the real-time information and finally make a decision accurately and timely. As a typical problem for researching the confrontation and cooperation of multi-agents, the pursuit problem is a problem of real-time dynamic system cooperative game, and a plurality of key technologies are applied to the industrial field and are attracted by people.
In studying the enclosure of a single escaper by multiple chasers within a closed bounded area, Zhengyuan Zhou et al propose an enclosure strategy based on minimizing the escaper's generalized Voronoi cell area, simplifying the high dimensional problem, where each chaser can share state information, compute its own strategy input independently, and reduce the capture time by improving cooperativity. The strategy is verified to ensure that the chaser finishes the enclosure catching of the escaper within a limited time, and a new technical scheme is provided for solving the problem of the chasing game. In recent years, the application scenarios of the multi-agent pursuit problem are increasing, and higher requirements are not provided for the existing pursuit algorithm from the confrontation of unmanned aerial vehicles to the pursuit of space vehicles and spacecrafts, so that the unmanned aerial vehicle pursuit algorithm has the requirements of better obstacle avoidance, high expansibility, flexibility and closer to the actual environment (with obstacles and exits). For the game problem in the environment with obstacles, the prior art mostly combines a target distribution algorithm and a classical differential game algorithm, and finds an optimal track by integrating backwards from a terminal condition according to a set performance index function, thereby obtaining an optimal trapping strategy of a chaser. When the scale of the intelligent agent is large, the method has high state space dimension and is difficult to solve, and the problem of dimension disaster is easily caused.
Disclosure of Invention
The invention provides a modeling and trapping strategy generation method for a multi-agent pursuit problem, aiming at overcoming the defects that the actual pursuit situation cannot be truly reflected because the actual environment of an obstacle and an exit is not considered in the existing trapping strategy and solving the technical problem that the existing trapping strategy considering the environment of the obstacle is difficult when the scale of an agent is large. The invention expands the existing pursuit escape problem to a scene closer to the reality, can apply the pursuit escape algorithm to a more detailed scene by considering the condition of exits and obstacles, and can reflect the actual pursuit escape condition more truly. The provided enclosure strategy generation method can realize the enclosure task of the chaser in the real environment.
The technical scheme of the invention is as follows:
a multi-agent pursuit problem modeling and containment strategy generation method is characterized by comprising the following steps:
step 1: modeling for multi-agent pursuit problem
Step 1.1: building a gaming environment
1.1.1 define a bounded unclosed region Ω
Defining a bounded unclosed region omega with n on its boundaryexpAn outlet with n in region omegabarTaking any point in an area omega as a coordinate origin, taking a horizontal rightward direction as an x-axis positive direction and a vertical x-axis direction as a y-axis positive direction, and establishing a global coordinate system xOy;
position of each outlet on the boundary of region omega
Position of stationary obstacles in region omega
The width of each outlet on the boundary of region Ω is denoted as { Dk|k＝1,···,nexpArea of each static obstacle in region Ω is denoted as { S }w|w＝1,···,nbarInfluence of each stationary obstacle is halfLet a diameter be { ρw|w＝1,···,nbarThe influence range of each static obstacle is that the center of the obstacle is taken as a round point and the influence radius rho iswIs the circular domain of radii, affecting the radius ρwArtificially setting, wherein the value of the safety distance r is ensured to be equal to the value of the safety distance r, each circular area can completely cover the obstacle, and the distance from any point on the boundary of the obstacle to the boundary of the circular area is greater than the safety distance rsA safe distance rsThe device is set according to actual requirements, and meanwhile, the positions of all the outlets are not in the influence range of all the static obstacles.
1.1.2 defining parameters of Agents
Defining a plurality of intelligent bodies, dividing the intelligent bodies into two types of chasers and evacuees, and setting the chasers P ═ P i1, ·, N }, escaper E ═ E ·j1, ·, M }, i.e., the number of chasers is N, the number of escapers is M, and the location x of each agent isp∈Ω，xeE.g. omega, the distance from the initial position of each escaper to any outlet is specified to be larger than the escape distance reBy adjusting the escape distance reThe value can change the difficulty of escaping from the escaper; meanwhile, in the pursuit process, each intelligent agent completely knows the position information of the exit in the non-closed area omega, the position information of the static obstacle and the position information of each intelligent agent, namely, the process is a game under the complete information. The equation of motion for each agent is shown in equation (1):
in the formula (I), the compound is shown in the specification,
initial positions, u, of chaser and fleeer, respectivelyi，ujSpeed control inputs for chasers and fleets, respectively, each having a maximum rate of motion vp,max，ve,maxAnd v isp,max≥ve，max；
Step 1.2: setting decision mode
The judgment of each state in the pursuit escape game is defined as follows:
when a certain escaper is far from the chaser at any distance dijAre all less than the capture distance dminOr when the escaper collides with the boundary of the region omega, the escaper is considered to be successfully captured; dminSetting according to actual requirements;
when a certain escaper reaches any one of the outlets of the region omega or passes through a certain outlet, the escaper is considered to be successful in escaping;
if the escapers in the region omega are successfully captured or successfully escaped, the pursuit game is considered to be ended;
step 1.3: setting an escaper policy
In order to ensure the universality of the hunting strategy of the chaser, namely, the hunting strategy of the invention can capture the escaper no matter how the escaper moves, so the invention does not make special requirements on the movement of the escaper, and only makes the following provisions:
1) the escaper can identify and avoid the obstacle;
2) the escaper should escape from the enclosure of the chaser as much as possible;
3) on the basis of realizing the two requirements, the escaper should move to the outlet as far as possible to realize escape;
step 2: generating a containment strategy
Step 2.1: enclosure task allocation
Using the position coordinates of each agent in the region in the global coordinate system xOy as the generatrix of the Voronoi diagram, generating the Voronoi unit of each agent, and aiming at a certain chaser pi，
If the escaper exists in the adjacent Voronoi unit, the escaper nearest to the chaser is the target of the enclosure capture;
if there is no escaper in the adjacent Voronoi cell, the chaser piThe escaper closest to the escaper in the region omega is taken as a target for enclosure capture;
the respective chasers in the region Ω thus available enclose the target accordingly.
Step 2.2: determining an enclosure target point
Calculating chaser piTo the target ejIntercept factor fij，
When f isijWhen p is greater than or equal to 0, catch upiIs a target object ejThe location of the nearest outlet;
when f isij<0, chasing person piThe target point is determined by a Voronoi partitioning method;
step 2.3: determining direction and rate of travel of chaser
Step 2.4: pursuit and escape game
And (3) each chaser moves for one time unit in the advancing direction of the chaser to obtain the position coordinate of the next moment, and the step 2.1 is returned until the chaser is judged to finish the chaser game according to the judgment mode of the step 1.2.
Further, in step 2.2 above, the chaser p is calculated according to the following formulaiTo the target ejCoefficient of interception
In the formula, k is a distance e from the current trapping targetjThe number of the nearest exit is,
is a chaser piThe distance to the k-th outlet is,
for current trapping of target ejDistance to the kth outlet.
Further, the method for determining the target point by the Voronoi partition method in step 2.2 is specifically as follows: if it is caught piAnd an enclosure target ejIf there is a boundary between Voronoi cells, then the chaser piThe target point of (1) is the middle point of the boundary of the two Voronoi units; if it is caught piAnd an enclosure target ejThe Voronoi cell of (1) has no boundary, then the chaser piIs a trapping target pjIs located.
Further, the method for determining the traveling direction of the chaser in the step 2.3 comprises the following steps: calculating chaser piThe resultant force of the attraction force and the repulsion force is applied, and the direction of the resultant force is the chaser piThe direction of travel of.
Further, the method for determining the traveling direction of the chaser in the step 2.3 is specifically as follows:
2.3.1 calculation of chaser piIs subjected to an attractive force from the target point
Fatt(pi)＝ξρ(pi,qgoal) (3)
Where xi is a gravitational gain coefficient, ρ (p)i,qgoal) Is a chaser piDistance from its target point, the direction of the attractive force being directed by the chaser piAt a position pointing to the target point.
2.3.2 calculation of chaser piSubject to repulsion from w obstacles
Where η is the repulsive gain coefficient, ρwThe radius of influence of the w-th obstacle,
is a chaser piThe direction of the repulsive force is directed to the chaser p from the position of the w-th barrieri。
2.3.3 calculation of chaser piResultant force of attraction force and repulsion force
In the formula, nbarThe attraction force and the repulsion force are vector superposition for the number of static obstacles, and the resultant force F (p)i) Is the direction of (1) as the chaser piThe direction of travel of.
Further, step 2.3 sets that each chaser is traveling at a maximum rate of motion, i.e. the chaser is moving at a maximum speed of motion
The invention has the beneficial effects that:
1. when modeling the multi-agent pursuit problem, the invention comprehensively considers the situations of obstacles and exits in the real scene, is more practical compared with the traditional pursuit problem model, and can apply the research on the pursuit algorithm to more detailed scenes.
2. According to the enclosure strategy generation method provided by the invention, each chaser can independently determine the enclosure task, and the self-planning is dynamically adjusted according to the change of the position information of each agent in the game process, so that the cooperativity among the chasers is improved, the completion of the whole task is further accelerated, and the enclosure strategy generation method is particularly suitable for the enclosure task of multiple chasers to multiple escapers.
3. According to the enclosure strategy generation method provided by the invention, each chaser adjusts the enclosure target point in real time according to the situation change of the escaper, and receives the repulsive force from the barrier while enclosing the escaper, and the closer the chaser is to the barrier, the larger the repulsive force is, so that the obstacle can be avoided and the escaper can be enclosed, and the enclosure strategy generation method is particularly suitable for the enclosure task in a real complex scene.
4. The capture strategy generation method provided by the invention has the advantages that the game environment is subjected to Voronoi partitioning, each chaser takes the Voronoi unit of the minimum escaper as a target, the factors needed to be considered for decision making are few, and a certain chaser can obtain the required motion strategy only by knowing the position information of each intelligent agent and the position information of the obstacle and the exit in the environment, so that the calculation is only carried out in the low-dimensional configuration space of a single intelligent agent, but not in the high-dimensional combined state space of all the intelligent agents, and the solution is simple.
Drawings
FIG. 1 is a flow chart of a modeling and containment strategy generation method of the present invention.
FIG. 2 is a first process of the multi-agent pursuit game of the present invention.
FIG. 3 is a second process of the multi-agent pursuit game of the present invention.
FIG. 4 is a third process of the multi-agent pursuit game of the present invention.
FIG. 5 is a graph showing the minimum distance change from the escaper to the chaser according to the present invention.
Detailed Description
The invention is further described below with reference to the accompanying drawings.
The invention provides a multi-agent pursuit problem modeling and trapping strategy generation method, wherein the modeling method comprises the following steps: constructing a game environment, setting a judgment mode and setting a runner strategy; the method for generating the trapping strategy comprises the following steps: allocating an enclosure task, determining an enclosure target point, determining the traveling direction and the traveling speed of the chaser, playing a game and judging whether the game is finished.
Step 1: modeling for multi-agent pursuit problem
Step 1.1: building a gaming environment
Given a certain square region Ω (in other embodiments, it may also be another shape region, such as a circular, polygonal, or irregular shape region, and the method steps involved in the following are not changed), the side length is 3km, the left-lower vertex of the region Ω is taken as the origin of coordinates, the horizontal-right direction is taken as the positive direction of the x-axis, and the vertical-x-axis direction is taken as the positive direction of the y-axis, so as to establish the global coordinate system xOy. The boundary of the region omega is provided with 4 outlets, 7 static obstacles are arranged in the region omega, the positions of the outlets and the static obstacles in the region omega are shown in fig. 2, the five-pointed star on the boundary in fig. 2 represents the outlets, the black filled region inside the boundary represents the static obstacles, and the region surrounded by the dotted circle around the static obstacles is the influence range of the static obstacles. And width of each outlet { Dk0.05km | k ═ 1, ·,4}, the area of each obstacle { S ·1＝S2＝0.08km2，S3＝S4＝S5＝S6＝S7＝0.02km2H, safety distance rs0.15km, each staticInfluence radius [ rho ] of obstacle stoppingw0.35km | w ═ 1, ·,7}, and the exit positions are not within the influence range of the obstacle.
The number of the chasers N is 4, the number of the escapers M is 3, namely, the chasers P is { P i1, ·,4}, escaper E ═ E ·j1, ·,3}, the positions of the agents are shown in fig. 2, wherein X1-X4 represent chasers, X5-X7 represent escapes, the motion tracks of the agents are marked in the figure, and the distance from the initial position of each escape to any exit is greater than the escape distance re0.2 km. Meanwhile, each agent has complete knowledge of the exit on the boundary of the non-closed region Ω under the global coordinate system xOy, the position information of the stationary obstacles in the region Ω, and the position information of each agent. The equation of motion for each agent is as follows:
in the formula (I), the compound is shown in the specification,
initial positions, u, of chaser and fleeer, respectivelyi，ujSpeed control inputs for chasers and fleets, respectively, each having a maximum rate of motion vp,max，ve,maxIn this embodiment, the maximum movement rates of the chaser and the fleeer are vp,max＝0.02km/s，ve,max＝0.02km/s。
Step 1.2: setting decision mode
When a certain escaper is far from the chaser at any distance dijAre all less than the capture distance dminWhen the number is 0.04km or the escaper collides with the boundary, the escaper is considered to be successfully captured;
when a certain escaper arrives at any one of the exits or passes through a certain exit, the escaper is regarded as successful in escaping.
If the evacuees in region Ω are either successfully captured or have successfully escaped, the catch-up game is considered to be over.
Step 1.3: setting an escaper policy
The escaper strategy meets the following three requirements:
1) the escaper can identify and avoid the obstacle;
2) the escaper should escape from the enclosure of the chaser as much as possible;
3) on the basis of realizing the two requirements, the escaper should move to the outlet as far as possible to realize escape;
the embodiment combines an artificial potential field method to set a escaper strategy, and the specific method is as follows:
step 1.3.1 determining the escape target Point
When a certain escaper ejWhen the escape person is not surrounded by the chaser, namely when the escape person has no chaser and an exit, the escape person selects the exit with the closest distance in the direction of the escape person as a target point;
when the escaper ejWhen surrounded by a chaser, the escaper should move in a direction away from the chaser nearest to the escaper, and the target point at the moment is a point in the direction with a length equal to the distance between the escaper and the chaser nearest to the escaper.
Step 1.3.2 determining the direction and rate of flight of the escaper
The direction of travel of the escaper is determined using the following method:
in this embodiment, a certain escaper ejIs subjected to an attractive force from the target point
Fatt(ej)＝ξρ(ej,qgoal)
In the formula, the gravity gain coefficient ξ is 0.7, ρ (e)j,qgoal) For escaping person ejDistance from its target point, the direction of the attraction being determined by the escaper ejAt a position pointing to the target point.
A certain escaper ejSubject to repulsion from w obstacles
Where the repulsive gain coefficient η is 0.3, ρwThe radius of influence of the w-th obstacle,
for escaping person ejThe direction of the repulsive force is directed to the escaper e from the position of the w-th obstaclej。
A certain escaper ejResultant force of attraction force and repulsion force
In the formula, the number n of stationary obstaclesbarEach escaper can obtain a resultant force F (e) by the above formulaj) And thus its direction of travel can be determined.
Rate of flight setting for the escaper:
setting each fleeer to travel at a maximum rate of movement, ve＝ve,max＝0.02km/s。
Step 2: generating a containment strategy
Step 2.1: enclosure task allocation
The position coordinates of each agent in the region omega are used as the generatrix of the Voronoi diagram, Voronoi units of each agent are generated, and a certain chaser piIf the escaper exists in the adjacent Voronoi unit, the escaper nearest to the chaser is the target of the enclosure capture; if there is no escaper in the adjacent Voronoi cell, the chaser piThe escaper closest to itself in the region Ω should be taken as the target of the enclosure. The respective chasers in the region Ω thus available enclose the target accordingly. The catching-up game process is shown in fig. 2, 3 and 4, and taking fig. 2 as an example, the catching-up persons X1, X2 and X4 use the escaper X7 as the escaperThe target of enclosure, i.e., the chaser X3, is the escape X5 target of enclosure.
Step 2.2: determining an aim point for an enclosure
Each chaser calculates the interception coefficient f of the enclosure target through the following formulaij，
In the formula, k is a distance e from the current trapping targetjThe number of the nearest outlet is given,
is a pursuit of the person piThe distance to the k-th outlet is,
for current trapping of target ejDistance to the kth outlet.
When a chaser piIntercept coefficient fijWhen the value is more than or equal to 0, the chaser piShould be a target point e from the trapping targetjThe nearest exit coordinate;
when a chaser piIntercept factor fij<0, the chaser piThe target point (b) should be determined by a Voronoi partitioning method, specifically: if it is caught piAnd an enclosure target ejIf there is a boundary between Voronoi cells, then the chaser piThe target point of (1) is the middle point of the boundary of the two Voronoi units; if it is caught piAnd an enclosure target ejThe Voronoi cell of (1) has no boundary, then the chaser piTarget point of (a) is an enclosure target ejIs located.
Taking fig. 3 as an example, each chaser is connected with the target point thereof through a straight line, and the interception coefficient f of the chaser X1 to the enclosure target X515＝-11.3s<0, and the two Voronoi cells have a boundary, so the target point of the chaser X1 is the midpoint of the boundary between the Voronoi cell and the Voronoi cell surrounding the target X5, as indicated in the figure; interception coefficient f of chaser X2 to enclosure target X626＝-10.1s<0, and the two Voronoi cells have a boundary, so the target point of the chaser X2 is the midpoint of the boundary between the Voronoi cell and the Voronoi cell surrounding the target X6, as indicated in the figure; interception coefficient f of chaser X3 to enclosure target X535＝44.5s>0, and the exit nearest to the enclosure target X5 is the lower exit, so the target point of the chaser X3 is the lower exit; interception coefficient f of chaser X4 to enclosure target X646＝0.5s>0, and the exit nearest to the enclosure target X6 is the left exit, so the target point of the chaser X4 is the left exit.
Step 2.3: determining direction and rate of travel of chaser
The traveling direction of the chaser is determined by the following method:
a chaser p can be calculated from the following formulaiIs subjected to an attractive force from the target point
Fatt(pi)＝ξρ(pi,qgoal)
In the formula, the gravity gain coefficient ξ is 0.7, ρ (p)i,qgoal) Is a chaser piDistance from its target point, the direction of the attraction being determined by the chaser piAt a position pointing to the target point.
Pursuing the person piSubject to repulsion from w obstacles
Wherein the repulsive force gain coefficient eta is 0.3,
is a chaser piDistance from the w-th obstacle, pwThe direction of the repulsive force is directed to the chaser p from the position of the w-th barrieri。
Catch person piResultant force of attraction force and repulsion force
In the formula, the number n of stationary obstaclesbarEach chaser can obtain the resultant force F (p) by the above formula (7)i) And thus its direction of travel can be determined.
Travel rate setting of chaser:
each chaser travelling at maximum rate of movement during the game, i.e.
Step 2.4: and (3) each chaser moves for a time unit in the advancing direction of the chaser to obtain the position coordinate of the next moment, the step 2.1 is returned until the chaser escape game is ended, and whether the chaser escape game is ended is judged according to the judgment mode set in the step 1.2.
As can be seen from the motion tracks of the chasers in fig. 4, in the process of the pursuit escape game, the chasers do not collide with the obstacle, so that the requirement of obstacle avoidance is met; the escapers exit the game as long as the catching condition is met, and as can be seen from the last section of each curve in fig. 5, the minimum distance from the escaper to the catcher is smaller than the catching distance d before the catching game is finishedminThe capturing condition set in the step 1.2 can be met when the speed is 0.04km, so that each chaser can complete the enclosure capturing of the escaper under the established pursuit game model, and the method provided by the invention is suitable for the enclosure capturing task in the real environment.
Claims (6)
1. A multi-agent pursuit problem modeling and containment strategy generation method is characterized by comprising the following steps:
step 1: modeling for multi-agent pursuit problem
Step 1.1: building a gaming environment
1.1.1 define a bounded unclosed region Ω
Defining a region omega with n on its boundaryexpAn outlet with n thereinbarA stationary barrier, and each exit is not located at each stationary barrierWithin the range of influence of the obstacle; n isexp≥1，nbar≥1；
Taking any point in the region omega or on the boundary as a coordinate origin, taking the horizontal rightward direction as the positive direction of an x axis and the vertical x axis as the positive direction of a y axis, and establishing a global coordinate system xOy;
1.1.2 defining parameters of Agents
Defining a plurality of intelligent agents, wherein each intelligent agent is in an area omega or on a boundary, and each intelligent agent knows each outlet, a static obstacle and position information of each intelligent agent under a global coordinate system xOy in the pursuing process;
the intelligent agent is divided into two types of chasers and escapers, wherein the number of the chasers is N, the number of the escapers is M, and the distance from the initial position of each escaper to any outlet is greater than the escape distance re；N≥1，M≥1，reSetting according to actual requirements;
the maximum movement rate of the chaser is greater than or equal to the maximum movement rate of the escaper;
step 1.2: setting decision mode
The decision method is as follows:
when a certain escaper is far from the chaser at any distance dijAre all less than the capture distance dminOr when the escaper collides with the boundary of the region omega, the escaper is considered to be successfully captured; dminSetting according to actual requirements;
when a certain escaper reaches or passes through any one of the outlets of the region omega, the escaper is regarded as successful in escaping;
if the escapers in the region omega are successfully captured or escaped, the pursuit game is considered to be ended;
step 1.3: setting an escaper policy
The escaper strategy is as follows:
1) the escaper can identify and avoid the obstacle;
2) the escaper should escape from the enclosure of the chaser as much as possible;
3) on the basis of realizing the two requirements, the escaper should move to the outlet as far as possible to realize escape;
step 2: generating a containment strategy
Step 2.1: enclosure task allocation
Using the position coordinates of each agent in the region in the global coordinate system xOy as the generatrix of the Voronoi diagram, generating the Voronoi unit of each agent, and aiming at a certain chaser pi，
If the escaper exists in the adjacent Voronoi unit, the escaper nearest to the chaser is the target of the enclosure capture;
if there is no escaper in the adjacent Voronoi cell, the chaser piTaking the escaper closest to the escaper in the region omega as a capture target;
step 2.2: determining an enclosure target point
Calculating chaser piTo the target ejIntercept factor fij，
When f isijWhen p is greater than or equal to 0, catch upiIs a target object ejThe location of the nearest exit;
when f isijWhen < 0, the chaser piThe target point of (2) is determined by a Voronoi division method;
step 2.3: determining a direction of travel and a rate of travel of a chaser
Step 2.4: pursuit and escape game
And (3) each chaser moves for one time unit in the advancing direction of the chaser to obtain the position coordinate of the next moment, and the step 2.1 is returned until the chaser is judged to finish the chaser game according to the judgment mode of the step 1.2.
2. The multi-agent pursuit problem modeling and containment strategy generation method of claim 1, characterized by: in step 2.2 the chaser p is calculated according to the following formulaiTo the target ejCoefficient of interception
In the formula, k is a distance e from the current trapping targetjThe number of the nearest exit is,
is a chaser piThe distance to the k-th outlet is,
for current trapping of target ejDistance to kth outlet, vp，maxMaximum rate of movement of chaser, ve，maxIs the maximum rate of movement of the player.
3. The multi-agent pursuit problem modeling and containment strategy generation method of claim 2, characterized by: the method for determining the target point by the Voronoi partitioning method in the step 2.2 specifically comprises the following steps: if it is caught piAnd an enclosure target ejIf there is a boundary between Voronoi cells, then the chaser piThe target point of (1) is the middle point of the boundary of the two Voronoi units; if it is caught piAnd an enclosure target ejThe Voronoi cell of (1) has no boundary, then the chaser piTarget point of (a) is an enclosure target ejIs located.
4. The multi-agent pursuit problem modeling and containment strategy generation method of claim 1, characterized by: the method for determining the traveling direction of the chaser in the step 2.3 comprises the following steps: calculating chaser piThe resultant force of the attraction force and the repulsion force is applied, and the direction of the resultant force is the chaser piThe direction of travel of.
5. The multi-agent pursuit problem modeling and hunting strategy generation method of claim 4, wherein:
step 2.3 the method for determining the traveling direction of the chaser specifically comprises the following steps:
2.3.1 calculation of chaser piIs subjected to an attractive force from the target point
Fatt(pi)＝ξρ(pi，qgoal)
In the formula, xi is the gravityGain factor, p (p)i，qgoal) Is a chaser piDistance from its target point, the direction of the attraction being determined by the chaser piThe position points to the target point;
2.3.2 calculation of chaser piSubject to repulsion from w obstacles
Where η is the repulsive gain coefficient, ρwThe radius of influence of the w-th obstacle,
is a chaser piThe direction of the repulsive force is directed to the chaser p from the position of the w-th barrieri；
2.3.3 calculation of chaser piResultant force of attraction force and repulsion force
In the formula, nbarThe attraction force and the repulsion force are vector superposition for the number of static obstacles, and the resultant force F (p)i) Is the direction of (1) as the chaser piThe direction of travel of.
6. The multi-agent pursuit problem modeling and containment strategy generation method of claim 1, characterized by: in step 2.3 it is set that each chaser is travelling at maximum rate of movement.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210104867.6A CN114518754B (en) | 2022-01-28 | Multi-agent escape problem modeling and trapping strategy generation method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210104867.6A CN114518754B (en) | 2022-01-28 | Multi-agent escape problem modeling and trapping strategy generation method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114518754A true CN114518754A (en) | 2022-05-20 |
| CN114518754B CN114518754B (en) | 2024-04-23 |
Family
ID=
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116430865A (en) * | 2023-04-17 | 2023-07-14 | 北方工业大学 | Multi-machine collaborative trapping method under uncertain probability framework |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20040073368A1 (en) * | 2002-05-10 | 2004-04-15 | Hector Gonzalez-Banos | Real-time target tracking of an unpredictable target amid unknown obstacles |
| CN109085754A (en) * | 2018-07-25 | 2018-12-25 | 西北工业大学 | A kind of spacecraft neural network based is pursued and captured an escaped prisoner game method |
| WO2019194628A1 (en) * | 2018-04-06 | 2019-10-10 | 엘지전자 주식회사 | Mobile robot and control method for same |
| CN113552872A (en) * | 2021-01-28 | 2021-10-26 | 北京理工大学 | Pursuit and escape game decision method for chaser at different speeds |
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20040073368A1 (en) * | 2002-05-10 | 2004-04-15 | Hector Gonzalez-Banos | Real-time target tracking of an unpredictable target amid unknown obstacles |
| WO2019194628A1 (en) * | 2018-04-06 | 2019-10-10 | 엘지전자 주식회사 | Mobile robot and control method for same |
| CN109085754A (en) * | 2018-07-25 | 2018-12-25 | 西北工业大学 | A kind of spacecraft neural network based is pursued and captured an escaped prisoner game method |
| CN113552872A (en) * | 2021-01-28 | 2021-10-26 | 北京理工大学 | Pursuit and escape game decision method for chaser at different speeds |
Non-Patent Citations (1)
| Title |
|---|
| 张旭;贾磊磊;陈群;: "关于多机器人围捕协调路径策略研究", 计算机仿真, no. 06, 15 June 2016 (2016-06-15), pages 362 - 366 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116430865A (en) * | 2023-04-17 | 2023-07-14 | 北方工业大学 | Multi-machine collaborative trapping method under uncertain probability framework |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108549407B (en) | Control algorithm for multi-unmanned aerial vehicle cooperative formation obstacle avoidance | |
| CN108827312B (en) | Cooperative game path planning method based on neural network and artificial potential field | |
| CN108594853B (en) | Unmanned aerial vehicle formation control method | |
| CN111609864B (en) | Multi-policeman cooperative trapping task allocation and path planning method under road network constraint | |
| Li et al. | Path planning of unmanned aerial vehicle based on improved gravitational search algorithm | |
| CN108680063A (en) | A kind of decision-making technique for the dynamic confrontation of extensive unmanned plane cluster | |
| CN112917486B (en) | Automatic planning method for intelligent spraying path of ship outer plate based on unmanned aerial vehicle | |
| CN109521794A (en) | A kind of multiple no-manned plane routeing and dynamic obstacle avoidance method | |
| CN109542115B (en) | Multi-unmanned aerial vehicle formation flight obstacle avoidance control strategy making method | |
| Bhattacharya et al. | A cell decomposition approach to visibility-based pursuit evasion among obstacles | |
| CN113655810B (en) | Unmanned aerial vehicle obstacle avoidance method and system based on speed potential field | |
| CN107703970B (en) | Unmanned plane cluster is around method for tracing | |
| CN106716282B (en) | Method, control device and the control equipment that a kind of pair of target is controlled | |
| CN113359756B (en) | Method for realizing real-time planning of obstacle avoidance path of omnidirectional mobile robot based on grid method | |
| CN113189992B (en) | Group intelligent obstacle avoidance method based on local orientation | |
| CN114442628A (en) | Mobile robot path planning method, device and system based on artificial potential field method | |
| CN114518754A (en) | Multi-agent pursuit problem modeling and trapping strategy generation method | |
| CN116203989A (en) | Multi-unmanned aerial vehicle cooperative target searching method and system based on particle swarm optimization | |
| CN114518754B (en) | Multi-agent escape problem modeling and trapping strategy generation method | |
| CN112651486A (en) | Method for improving convergence rate of MADDPG algorithm and application thereof | |
| CN116225065A (en) | Unmanned plane collaborative pursuit method of multi-degree-of-freedom model for multi-agent reinforcement learning | |
| CN113900449B (en) | Multi-unmanned aerial vehicle track planning method and device, unmanned aerial vehicle and storage medium | |
| CN114779828B (en) | Unmanned aerial vehicle cluster topological control and intelligent anti-collision method based on heterogeneous formation datum points | |
| Khachumov | Solution of the problem of group pursuit of a target under perturbations (spatial case) | |
| CN115470706A (en) | Hierarchical reinforcement learning framework training method for multi-target decision |
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 |