A real-time path planning method for dynamic continuous monitoring of multiple robots based on synthetic pheromone model

Abstract

The present invention provides a multi-robot dynamic continuous monitoring real-time path planning method based on a synthetic pheromone model, which solves the shortcomings of the existing multi-robot dynamic road network continuous monitoring method, which has high computational complexity and cannot maintain a stable monitoring effect, so as to realize the coordinated continuous monitoring of a road network containing complex obstacles by multiple robots with detection capabilities. It includes: 1) identifying individual positions and calculating monitoring uncertainty; 2) global position declaration; 3) pheromone synthesis; 4) pheromone correction; 5) using synthetic pheromones to calculate the monitoring benefit value of the robot to the adjacent monitoring point in the undirected weighted graph; 6) determining the moving direction and generating monitoring path points.

Description

Technical Field

[0001] The invention belongs to the technical field of continuous monitoring of mobile robots, and in particular relates to a real-time path planning method for dynamic continuous monitoring of multiple robots based on a synthetic pheromone model. Background Art

[0002] How to successfully detect randomly occurring events in a known environment is an interesting and widely discussed topic. This topic comes from many real-world application scenarios, such as detecting possible accidents in factory inspections, detecting randomly occurring fires in forest firefighting, detecting sudden changes in ambient temperature or humidity in precision agriculture, and detecting randomly occurring biological populations in scientific investigations. A known environment means that the area in the environment where events are likely to occur is determinable. The way to capture 100% of the events is to place enough static sensors in the area, but this approach is obviously uneconomical, and in most outdoor or field environments, it is not realistic to deploy enough static sensors. One of the reasonable solutions is to use multiple robots with limited detection capabilities to continuously monitor the known area so that random events can be detected in a timely manner when they occur.

[0003] Existing multi-robot dynamic road network continuous monitoring methods are not suitable for online collaborative planning tasks because of their high computational complexity, at least O(N 4 ), which is a well-known symbol in the field of algorithms; secondly, the probability-based decision-making mechanism cannot identify the local optimal phenomenon, resulting in unstable results even in multiple calculations of the same task scenario. This deficiency increases the time for parking, turning, and U-turns during online planning, and cannot maintain a stable monitoring effect.

[0004] Based on this, it is necessary to design a new path planning method to achieve dynamic and continuous monitoring of multiple robots. Summary of the invention

[0005] The purpose of the present invention is to solve the shortcomings of the existing multi-robot dynamic road network continuous monitoring method, which has high computational complexity and cannot maintain a stable monitoring effect, and to provide a multi-robot dynamic continuous monitoring real-time path planning method based on a synthetic pheromone model to achieve collaborative continuous monitoring of a road network containing complex obstacles by multiple robots with detection capabilities.

[0006] To achieve the above purpose, the technical solution provided by the present invention is:

[0007] A multi-robot dynamic continuous monitoring real-time path planning method based on a synthetic pheromone model, which is special in that it includes the following steps:

[0008] 1) Identify individual locations and calculate monitoring uncertainty

[0009] 1.1) Read the environment model;

[0010] The environment model consists of a cell group consisting of a static observation obstacle Q and a road network G consisting of weighted monitoring points and feasible edges.<Q,G> ;

[0011] Q={q1,q2,…,q n} is a set of polygons, G = (V, E, D, W) is an undirected weighted graph;

[0012] V={v1,v2,…,v N} is a set of monitoring points, (x i ,y i ) indicates v i ∈V, N represents the number of monitoring points, E={e1,e2,…,e M} is the set of feasible edges, M is the number of feasible edges, D={d1,d2,…,d M} represents the edge length set, W = {σ1,σ2,…,σ N} is the monitoring weight set;

[0013] 1.2) Use the environment model to calculate the relative distance d from the monitoring point to the robot center s The piecewise probability function is used to construct the monitoring accuracy model of the robot for the monitoring points on the road network;

[0014] 1.3) Calculate monitoring uncertainty using the monitoring accuracy model, where monitoring uncertainty refers to the unreliability of the judgment results of all robots on the event status of a certain monitoring point;

[0015] 1.4) Determine whether each robot has reached the monitoring point in the environment model. If it has, execute step 2). If it has not, execute step 2) after it has arrived.

[0016] The above step 1.4) is not limited to being performed after step 1.3), and can be performed in parallel with step 1.2) or 1.3);

[0017] 2) Global location declaration

[0018] Each robot declares the location it has just left, is about to pass, and plans to go to;

[0019] 3) Pheromone synthesis

[0020] 3.1) Calculation of pheromone release using monitoring uncertainty

[0021]

[0022] in, is the pheromone released at the i-th point, indicating the information collected at the monitoring point, V i O is the complete monitoring point set of the ith monitoring point, ||V i O || is O vi The number of monitoring points contained in u j (t) is O vi The uncertainty of the jth point at time t, u max is the global maximum uncertainty at the current moment;

[0023] 3.2) Using pheromone release And the position information in the global position declaration calculates the diffusion information and undergo pheromone mutation;

[0024] Among them, the calculation of diffusion pheromone Specifically:

[0025]

[0026] The diffused pheromone refers to the maximum value of the pheromone diffused to a monitoring point by all adjacent monitoring points, V i A is the neighbor set of the ith monitoring point, that is, the set of monitoring points connected to the ith monitoring point by an edge. PU(i,j) is the pheromone transmitted from the jth point to the ith point, which is calculated as:

[0027]

[0028] in, represents the synthetic pheromone of the jth neighboring node, C d (i,j) is the diffusion loss function, β is the absorption coefficient, I i (t) is the absorption factor of the ith monitoring point at time t, p m >0 is the minimum diffusion;

[0029] Diffusion loss function C d (i, j), which represents the numerical loss of the synthetic pheromone when it diffuses between the i-th and j-th monitoring points;

[0030]

[0031] Among them, c L is the loss coefficient, which is a man-made variable used to control the loss rate. is the diffusion cost, which is used to control the local loss intensity, The initial value of ||D i,j ||, assuming any vu is the robot moving speed, t lim is the expected time, σ j is the monitoring weight of the jth monitoring point;

[0032] Absorption coefficient β = min(u k (t):k=1,2,...,N^u k (t)≠0), absorption factor I j (t):

[0033]

[0034] Among them, V c represents the set of all monitoring points occupied by the robot, v j represents the jth monitoring point;

[0035] The specific pheromone mutations are:

[0036] Order any v R i and all its neighbors Diffusion cost v Ri With v Li Diffusion cost Among them, E Lr ∈E L ={E L1 , E L2 ,…,E LR} is the monitoring point v Rr and v Lr The connected edges, η and κ are the cost control quantities;

[0037] Order any And i>r, OP i The release of pheromone p i r =υ, at the same time, let any The release of pheromone p j r =γ, and satisfies Among them, υ and γ are the synergistic strength control quantities;

[0038] 3.3) Using pheromone release and diffusion pheromones Synthetic pheromones

[0039]

[0040] 4) Pheromone Correction

[0041] Find the coordinates of the largest synthetic pheromone in step 3.3) as the location where the robot plans to go;

[0042] If the location the robot plans to go to satisfies d c <d<d L , modify the uncertainty of the position where the robot plans to go to to 1, and return to step 3); if not satisfied, execute step 5);

[0043] Where d is the distance from the robot’s current location to the location it plans to go to; d c Represents the robot's observation radius, d L represents the detection limit of the robot;

[0044] 5) Using synthetic pheromones Calculate the monitoring benefit value of the robot to the adjacent monitoring points in the undirected weighted graph

[0045]

[0046] Among them, A j is the monitoring benefit value of the robot going to the jth monitoring point;

[0047] 6) Determine the moving direction and generate monitoring path points

[0048] 6.1) Calculate the path to the monitoring point with the maximum monitoring benefit value

[0049] Calculate the direction θ(t) toward the monitoring point

[0050]

[0051] And use this to calculate the direction of travel (p x ,p y ), where p x =cosθ(t),p y =sinθ(t) respectively represent the coordinate increment of the trajectory;

[0052] 6.2) Generate monitoring path points

[0053] Let the robot move along the forward direction for a sampling time interval dt, and obtain the coordinates of the trajectory point at the next moment (x N ,y N ), where x N ,y N Satisfy x N =x t +p x *dt,y N =y t +p y *dt;

[0054] Further, in step 1.2), the monitoring accuracy model is:

[0055]

[0056] Among them, f cr is the recognition success rate of the rth UGV for events occurring at different distances, d s represents the distance from a point in the environment to the center of the unmanned vehicle, l s Represents a virtual connection line between a point in the environment and the center of the unmanned vehicle, d c represents the observation radius of the unmanned vehicle, d L represents the detection limit of the unmanned vehicle, g(d s ) is a monotone non-increasing function of any form, satisfying g(d L )=0,

[0057] Furthermore, in step 1.3), the monitoring uncertainty is:

[0058]

[0059] Among them, u i (t) is the coverage uncertainty of point i, σ i represents the coverage weight of point i, t k Indicates the current time, t lvi represents the time when the i-th point in the undirected graph was last fully recognized, t lim is the maximum allowed interval time, a i is the monitoring point v i The total number of times that u is fully monitored. i When (t) = 0, a i The value of is increased by 1.

[0060] Furthermore, in step 1.4), take the rth robot at the current time t k The position coordinates (x t ,y t ,z t ), determine whether it has reached one of the coverage points i, and record the coordinates (x i ,y i ,z i ), if the coverage point is reached, then

[0061] (x i -x t ) 2 +(y i -y t ) 2 ≤(v u ·dt / 2) 2

[0062] Then i is the position point of the robot; where dt is the sampling time interval, which is constant.

[0063] Furthermore, in step 2), the global variable v is used Ri ∈V L and v Li ∈V R Record the point that each robot just left and the point that it will pass through, and declare the global variable OP i ∈OP records the location where each robot plans to go.

[0064] The advantages of the present invention are:

[0065] 1. This invention proposes a planning method for the problem of continuous monitoring of multiple robots on a road network. Based on the principle of pheromone diffusion, a position cooperation technology suitable for online planning between multiple robots and a technology to avoid local optimality between multiple robots are designed to achieve continuous monitoring path planning. Only one variable is used to achieve global state estimation, local optimal avoidance and global position coordination. No special cooperation method is required. The monitoring effect is indirectly controlled by pheromones. At the same time, its computational complexity is only O(N 2 ), which greatly shortens the parking calculation time and meets the real-time requirements of online collaboration.

[0066] 2. The method of the present invention uses global information transmission, and does not need to be based on a probability-based decision-making mechanism. Instead, it directly identifies local optimal phenomena through pheromones and responds, thereby maintaining a stable and high-precision monitoring effect. Since the present invention uses relative information as a decision variable, it is particularly suitable for multiple robots in a real large-scale complex topology network to perform online continuous monitoring tasks. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] Figure 1 It is a flow chart of the trajectory generation method of the present invention;

[0068] Figure 2 The model used in the embodiment of the present invention;

[0069] Figure 3 The trajectories and coverage of all robots planned in the embodiments of the present invention;

[0070] Figure 4 FIG. 1 is the relationship between the monitoring uncertainty of the monitoring point and time in the embodiment of the present invention. DETAILED DESCRIPTION

[0071] The present invention is further described in detail below with reference to the accompanying drawings and specific embodiments:

[0072] like Figure 1As shown, the multi-robot dynamic continuous monitoring real-time path planning method based on the synthetic pheromone model of the present invention has the following specific implementation steps:

[0073] Step 1: Calculation of individual position identification and monitoring uncertainty:

[0074] Step 1.1, read the environment model: Figure 2 As shown in the figure, the environment consists of a cell group consisting of a static observation obstacle Q and a road network G consisting of weighted monitoring points and feasible edges.<Q,G> Q={q1,q2,…,q n} is a set of polygons, G = (V, E, D, W) is an undirected weighted graph. V = {v1, v2, ..., v N} is the set of monitoring points, N is the number of monitoring points, E={e1,e2,…,e M} is the set of feasible edges, M is the number of feasible edges, D={d1,d2,…,d M} represents the edge length set, W = {σ1,σ2,…,σ N} is the monitoring weight set.

[0075] 1.2) Use the environment model to calculate the relative distance d from the monitoring point to the robot center s The piecewise probability function of the robot is used to construct the monitoring accuracy model of the monitoring points on the road network, where d c =10m,d L =20m:

[0076]

[0077] Among them, f cr is the recognition success rate of the rth UGV for events occurring at different distances, d s represents the distance from a point in the environment to the center of the unmanned vehicle, l s Represents a virtual connection line between a point in the environment and the center of the unmanned vehicle, d c represents the observation radius of the unmanned vehicle, d L Represents the detection limit of the unmanned vehicle.

[0078] Step 1.3: Monitor uncertainty calculation according to formula (2), where σ i =1,t lim =1000s:

[0079]

[0080] Among them, u i (t) is the coverage uncertainty of point i, σ i represents the coverage weight of point i, t k Indicates the current time, t lvirepresents the time when the i-th point in the undirected graph was last fully recognized, t lim is the maximum allowed interval time, a i is the monitoring point v i The total number of times that u is fully monitored. i When (t) = 0, a i The value of is increased by 1;

[0081] Step 1.4: Determine whether each robot has reached the monitoring point in the environment model: take v UGV =1m / s, dt=1s, at the current time t k The position coordinates (x t ,y t ), determine whether one of the monitoring points i has been reached, and record the coordinates (x i ,y i ), if the monitoring point is reached, it is satisfied

[0082] (x i -x t ) 2 +(y i -y t ) 2 ≤0.25 (3)

[0083] Then i is the position of the robot. If it arrives, execute step 2). If it does not arrive, execute step 2 after it arrives.

[0084] Step 2: Global location declaration

[0085] Each robot declares the location it has just left, the location it is about to pass, and the location it plans to go to. Ri ∈V L and v Li ∈V R Record the point that each robot just left and the point that it will pass through, and declare the global variable OP i ∈OP records the location where each robot plans to go.

[0086] Step 3: Pheromone synthesis:

[0087] Step 3.1: Calculate the release of pheromones using monitoring uncertainty

[0088]

[0089] in, is the pheromone released at the i-th point, indicating the information collected at the monitoring point, V i O is the complete monitoring point set of the ith monitoring point, ||V i O|| is O vi The number of monitoring points contained in u j (t) is O vi The uncertainty of the jth point at time t, u max is the global maximum uncertainty at the current moment;

[0090] Step 3.2: Using pheromone release And the position information in the global position declaration calculates the diffusion information and undergo pheromone mutation;

[0091]

[0092] Among them, V i A is the neighbor set of the ith monitoring point, that is, the set of monitoring points connected to the ith monitoring point by an edge. PU(i,j) is the pheromone transmitted from the jth point to the ith point. m =0.01, β=0.05, calculated as:

[0093]

[0094] in, represents the synthetic pheromone of the jth neighboring node, C d (i,j) is the diffusion loss function, I i (t) is the absorption factor of the ith monitoring point at time t.

[0095] Diffusion loss function C d (i,j), which represents the numerical loss of the synthetic pheromone when it diffuses between the i-th and j-th monitoring points. L =0.25, yes

[0096]

[0097] in, is the diffusion cost, which is used to control the local loss intensity, The initial value of ||D i,j ||, assuming any

[0098] Absorption coefficient β = min(u k (t):k=1,2,...,N∧u k (t)≠0), absorption factor I j (t):

[0099]

[0100] Among them, V cDenote the set of all monitoring points occupied by robots as \(v\). j Denote the \(j\)-th monitoring point.

[0101] Specifically, pheromone mutation is as follows:

[0102] Take \(\eta = 500\), \(\kappa = 1000\), \(\upsilon = 0\), \(\gamma = 0.1\). For the cost control quantity, let any \(v\) Ri and all its neighbors the diffusion cost \(v\) Ri and \(v\) Li the diffusion cost where \(E\) Lr \(\in E\) L \(=\{E\) L1 , \(E\) L2 , \(\cdots\), \(E\) LR}\) is the edge connecting the monitoring points \(v\) Rr and \(v\) Lr ; Let any and \(i > r\), the pheromone \(p\) i released by \(OP\) i r \(= 0\). At the same time, let the pheromone \(p\) released by any j r \(= 0.1\).

[0103] Step 3.3: Use the released pheromone and the diffused pheromone to synthesize pheromone

[0104]

[0105] Step 4: Pheromone correction

[0106] Find the coordinates with the maximum synthesized pheromone in Step 3.3 as the location that the robot plans to go to (which can also be called the main diffusion source); when the main diffusion source satisfies \(10m < d < 20m\), where \(d\) is the distance from the main diffusion source to the robot, modify the uncertainty of the main diffusion source to 1 and return to Step 3); if not satisfied, execute Step 5);

[0107] Step 5: Use the synthesized pheromone to calculate the monitoring benefit value of the robot to the adjacent monitoring points in the undirected weighted graph

[0108]

[0109] where \(A\) j is the monitoring benefit value of the robot going to the \(j\)-th monitoring point.

[0110]

[0111] Step 6: Determine the moving direction and generate monitoring path points

[0112] Step 6.1: Calculate the path to the monitoring point with the maximum monitoring benefit value

[0113] The monitoring benefit value is the uncertainty reduction value that can be obtained from the current position to a certain monitoring point; the maximum monitoring benefit value is the maximum value of all monitoring benefit values ​​at the current moment;

[0114] Calculate the direction θ(t) towards the monitoring point

[0115]

[0116] And use this to calculate the direction of travel (p x ,p y ), where p x =cosθ(t),p y =sinθ(t) respectively represent the coordinate increment of the trajectory;

[0117] Step 6.2: Generate monitoring path points

[0118] Let the robot move along the forward direction for a sampling time interval dt, and obtain the coordinates of the trajectory point at the next moment (x N ,y N ,z N ). where x N ,y N ,z N satisfy

[0119] x N =x t +p x ,y N =y t +p y ,z N =z t +p z (11)

[0120] like Figure 3 , showing the trajectory and coverage of the robot planned online using the method of the embodiment of the present invention within a 3000s task cycle, Figure 4The ordinate represents the monitoring uncertainty of the undirected graph at a certain moment. It can be seen that in this embodiment, the monitoring uncertainty reaches a stable state over time, and each position is monitored, indicating that the method is effective for the collaboration between multiple robots. At the same time, the single running time of the embodiment is 0.013s, which is much smaller than the single running time of 0.506s of the prior art. The average uncertainty during the entire task is 0.072, which is approximately equal to 0.071 of the prior art. It can be seen that in a simulation environment, the present invention can achieve a monitoring effect comparable to that of the prior art with an absolutely superior computing speed, thereby greatly reducing the hardware cost, reducing the parking waiting time, and improving the monitoring effect in the actual environment.

[0121] The above description is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and these modifications or substitutions should be included in the protection scope of the present invention.

Claims

1. A multi-robot dynamic continuous monitoring real-time path planning method based on a synthetic pheromone model, characterized in that: The following steps are involved: 1) Identify individual locations and calculate monitoring uncertainty 1.1) Read the environment model; The environment model consists of a cell group consisting of a static observation obstacle Q and a road network G consisting of weighted monitoring points and feasible edges.<Q,G> ; Q={q1,q2,…,q n } is a set of polygons, G = (V, E, D, W) is an undirected weighted graph; V={v1,v2,…,v N } is a set of monitoring points, (x i ,y i ) indicates v i ∈V, N represents the number of monitoring points, E={e1,e2,…,e M } is the set of feasible edges, M is the number of feasible edges, D={d1,d2,…,d M } represents the edge length set, W = {σ1,σ2,…,σ N } is the monitoring weight set; 1.2) Use the environment model to calculate the relative distance d from the monitoring point to the robot center s The piecewise probability function is used to construct the monitoring accuracy model of the robot for the monitoring points on the road network; 1.3) Calculate monitoring uncertainty using the monitoring accuracy model; 1.4) Determine whether each robot has reached the monitoring point in the environment model. If it has, execute step 2). If it has not, execute step 2) after it has arrived. 2) Global location declaration Each robot declares the location it has just left, is about to pass, and plans to go to; 3) Pheromone synthesis 3.1) Calculation of pheromone release using monitoring uncertainty in, is the pheromone released at the i-th point, V i O is the complete monitoring point set of the ith monitoring point, ||V i O || is O vi The number of monitoring points contained in u j (t) is O vi The uncertainty of the jth point at time t, u max is the global maximum uncertainty at the current moment; 3.2) Using pheromone release And the position information in the global position declaration calculates the diffusion information and undergo pheromone mutation; Among them, the calculation of diffusion pheromone Specifically: Among them, V i A is the neighbor set of the ith monitoring point, that is, the set of monitoring points connected to the ith monitoring point by an edge. PU(i,j) is the pheromone transmitted from the jth point to the ith point, which is calculated as: in, represents the synthetic pheromone of the jth neighboring node, C d (i,j) is the diffusion loss function, β is the absorption coefficient, I i (t) is the absorption factor of the ith monitoring point at time t, p m >0 is the minimum diffusion; Diffusion loss function C d (i, j), which represents the numerical loss of the synthetic pheromone when it diffuses between the i-th and j-th monitoring points. Among them, c L is the loss coefficient, which is a man-made variable used to control the loss rate. is the diffusion cost, which is used to control the local loss intensity, The initial value of ||D i,j ||, assuming any v u is the robot moving speed, t lim is the expected time, σ j is the monitoring weight of the jth monitoring point; Absorption coefficient β = min(u k (t):k=1,2,...,N∧u k (t)≠0), absorption factor I j (t): Among them, V c represents the set of all monitoring points occupied by the robot, v j represents the jth monitoring point; The specific pheromone mutations are: Order any v Ri With all its neighbors Diffusion cost v Ri With v Li Diffusion cost Among them, E Lr ∈E L ={E L1 , E L2 ,…,E LR } is the monitoring point v Rr and v Lr The connected edges, η and κ are cost control quantities; Order any And i>r, OP i The release of pheromone p i r =υ, at the same time, let any The release of pheromone p j r =γ, and satisfies Among them, υ and γ are the synergistic strength control quantities; 3.3) Using pheromone release and diffusion pheromones Synthetic pheromones 4) Pheromone Correction Find the coordinates of the largest synthetic pheromone in step 3.3) as the location where the robot plans to go; If the location the robot plans to go to satisfies d c <d<d L , modify the uncertainty of the position where the robot plans to go to to 1, and return to step 3); if not satisfied, execute step 5); Where d is the distance from the robot’s current location to the location it plans to go to; d c Represents the robot's observation radius, d L represents the detection limit of the robot; 5) Using synthetic pheromones Calculate the monitoring benefit value of the robot to the adjacent monitoring points in the undirected weighted graph Among them, A j is the monitoring benefit value of the robot going to the jth monitoring point; 6) Determine the moving direction and generate monitoring path points 6.1) Calculate the path to the monitoring point with the maximum monitoring benefit value Calculate the direction θ(t) toward the monitoring point And use this to calculate the direction of travel (p x ,p y ), where p x =cosθ(t),p y =sinθ(t) respectively represent the coordinate increment of the trajectory; 6.2) Generate monitoring path points Let the robot move along the forward direction for a sampling time interval dt, and obtain the coordinates of the trajectory point at the next moment (x N ,y N ), where x N ,y N Satisfy x N =x t +p x *dt,y N =y t +p y *dt.

2. The multi-robot dynamic continuous monitoring real-time path planning method based on the synthetic pheromone model according to claim 1 is characterized by: In step 1.2), the monitoring accuracy model is: Among them, f cr is the recognition success rate of the rth UGV for events occurring at different distances, d s represents the distance from a point in the environment to the center of the unmanned vehicle, l s Represents a virtual connection line between a point in the environment and the center of the unmanned vehicle, d c represents the observation radius of the unmanned vehicle, d L represents the detection limit of the unmanned vehicle, g(d s ) is a monotone non-increasing function of any form, satisfying g(d L )=0, 3. The multi-robot dynamic continuous monitoring real-time path planning method based on the synthetic pheromone model according to claim 2 is characterized by: In step 1.3), the monitoring uncertainty is: Among them, u i (t) is the coverage uncertainty of point i, σ i represents the coverage weight of point i, t k Indicates the current time, t lvi represents the time when the i-th point in the undirected graph was last fully recognized, t lim is the maximum allowed interval time, a i is the monitoring point v i The total number of times that u is fully monitored. i When (t) = 0, a i The value of is increased by 1.

4. The multi-robot dynamic continuous monitoring real-time path planning method based on the synthetic pheromone model according to any one of claims 1 to 3, characterized in that: In step 1.4), take the rth robot at the current time t k The position coordinates (x t ,y t ), determine whether it has reached one of the coverage points i, and record the coordinates (x i ,y i ), if the coverage point is reached, then (x i -x t ) 2 +(y i -y t ) 2 ≤(v u ·dt / 2) 2 Then i is the position of the robot; where v u is the moving speed of the robot, and dt is the sampling time interval.

5. The multi-robot dynamic continuous monitoring real-time path planning method based on the synthetic pheromone model according to claim 4 is characterized by: In step 2), use the global variable v Ri ∈V L and v Li ∈V R Record the point that each robot just left and the point that it will pass through, and declare the global variable OP i ∈OP records the location where each robot plans to go.

Images