Multi-robot target position allocation method, medium and electronic equipment

Abstract

The present invention discloses a multi-robot target position allocation method, medium and electronic device. The method includes: modeling the sensor perception range as an integer multiple of the grid area; initializing the grid pheromone; performing position allocation: obtaining the position of the robot and its adjacent vertices, selecting the adjacent vertex with the largest pheromone as the pre-allocated position; if the distance between any two robots is less than or equal to the preset distance, updating the pre-allocated position with each previous position; otherwise: if the number of adjacent vertices is greater than 1, updating the pre-allocated position with the adjacent vertices except the previous position; otherwise, updating the pre-allocated position with the only adjacent vertex; jumping out of the self-loop when the robot is in a self-loop; reducing the two-point oscillation when the robot is in a two-point oscillation; reducing the L-shaped oscillation when the robot is in an L-shaped oscillation; using the pre-allocated position as the target position; updating the position of the robot; updating all grid pheromones. According to the method of the present invention, it is possible to deal with the problem of inefficient planning.

Description

Technical Field

[0001] The present invention relates to the technical field of trajectory planning, and in particular to a multi-robot target position allocation method, a computer-readable storage medium and an electronic device. Background Art

[0002] Scenes such as urban subways, underground parking lots, and air-raid shelters have limited space and complex channels. It is necessary to continuously monitor various state changes in limited spaces and obtain information including but not limited to people, objects, and the environment in a timely manner. Performing this task usually requires a lot of manpower. The traditional solution is to use a sensor network arranged in the environment with a certain density to cover the entire task space, but this static monitoring method often has problems such as fixed sensing range, sensing blind spots, and cost-effectiveness. Therefore, a multi-robot system is constructed, and each robot carries a sensor to move, which can dynamically cover the task area and obtain global information, thereby greatly improving the cost-effectiveness and attention to key areas.

[0003] Affected by the crisscrossing geographical features of limited space, improving the collaborative efficiency of multi-robot continuous monitoring has become a key issue that needs to be solved urgently, especially when the robot is in a narrow passage or a complex intersection, its field of view is blocked by the passage wall, and the field of view varies with the robot's spatial position. Existing research has not paid enough attention to this type of space, and most monitoring models idealize the robot's field of view as a complete and constant area. If this model is directly applied to the above narrow space, it is difficult to quantify the obstruction of the robot's field of view, which directly affects the coverage effect and collaborative efficiency. Summary of the invention

[0004] The present invention aims to solve at least one of the technical problems existing in the prior art. To this end, the present invention proposes a multi-robot target position allocation method, which can effectively deal with the planning inefficiency problem when the robot's field of view is blocked.

[0005] The present invention also proposes a computer-readable storage medium for implementing the multi-robot target position allocation method.

[0006] The present invention also provides an electronic device for implementing the multi-robot target position allocation method.

[0007] According to a multi-robot target position allocation method of an embodiment of a first aspect of the present invention, the method is used to allocate target positions of multiple robots in a task area, the robots are provided with sensors for accessing the task area, the task area is provided with multiple channels for the robots to walk, the widths of different channels are different from each other, the sensing range of the sensor is greater than the width of at least part of the channels, and the method comprises:

[0008] The task area is evenly divided into multiple grids on a two-dimensional plane to establish a two-dimensional grid map, the maximum side length of the grid is less than or equal to the minimum width of the channel, the robot is regarded as a particle on the two-dimensional grid map, the vertex of the grid located in the channel is taken as the target position of the robot, and the sensing range of the sensor is calculated according to the following formula: To model:

[0009]

[0010] In the formula, a is the number of grids covered by the sensor on the two-dimensional grid map; is the area of ​​a single grid;

[0011] Assigning values ​​to all the grids to initialize the pheromones of the grids, considering the grids that are within the sensing range of the sensor and are not blocked as being visited by the sensor, and performing pheromone attenuation on the visited grids;

[0012] Calculating the pheromone of the vertex based on the pheromones of the grid surrounding the vertex;

[0013] To perform location-allocation:

[0014] Obtaining the position of each robot and its adjacent vertices, and selecting the adjacent vertex with the largest pheromone as the pre-allocated position;

[0015] If the distance between any two of the robots is less than or equal to the preset distance, the pre-assigned position is updated with the last position of each of the two robots; otherwise:

[0016] If the number of the adjacent vertices of the robot is greater than 1, updating the pre-assigned position with the adjacent vertices except the previous position;

[0017] Otherwise, updating the pre-allocated position with the only adjacent vertex;

[0018] If the current execution position allocation is w and the current position of the robot, the pre-assigned position and the first ( w -1)th, w -2) When the positions of the robots cover the four vertices of the rectangle during the secondary allocation, the pre-allocated positions are updated with the adjacent vertices other than the four vertices of the rectangle by jumping out of the self-loop;

[0019] If the current execution position allocation is w The pre-assigned position is the same as the previous position and the current position of the robot is the same as the first (w -2) If the positions of the second allocation are the same, then reducing the two-point oscillation is performed to update the pre-allocated position with the adjacent vertices except the previous position;

[0020] If the current execution position allocation is w and the current position of the robot, the pre-assigned position and the first ( w -1)th, w -2) times, ( w -3) the positions in the secondary allocation jointly cover the three vertices of the L-shape, and if the number of the adjacent vertices is greater than 1, reducing the L-shape oscillation is performed to update the pre-allocated positions with the adjacent vertices other than the three vertices of the L-shape;

[0021] Taking the pre-assigned position as the target position and moving the robot toward the target position;

[0022] updating the position of the robot when the robot reaches the target position;

[0023] The pheromones of all the grids are updated, and the pheromone decay is performed on the visited grids.

[0024] The multi-robot target position allocation method according to the embodiment of the present invention can effectively deal with the problem of inefficient planning when the robot's field of vision is blocked.

[0025] In addition, the multi-robot target position allocation method according to the embodiment of the present invention also has the following additional technical features:

[0026] According to some embodiments of the present invention, the executing jump out of the self-loop further comprises:

[0027] The adjacent vertex with the largest pheromone is selected as the pre-assigned position.

[0028] According to some embodiments of the present invention, the performing the reduction of two-point oscillations further comprises:

[0029] One of the adjacent vertices is randomly selected as the pre-assigned position.

[0030] According to some embodiments of the present invention, the performing the reduction of L-shaped oscillations further comprises:

[0031] The adjacent vertex with the smallest number of visits is selected as the pre-allocated position.

[0032] According to some embodiments of the present invention, when a portion of the grid is within a sensing range of the sensor and is not blocked, the grid is regarded as being accessed by the sensor.

[0033] According to some embodiments of the present invention, grids at different distances from the same robot have different attenuation levels.

[0034] According to some embodiments of the present invention, the pheromone of the grid is updated according to the following formula P h :

[0035]

[0036] In the formula, k Indicates that the grid is located at k The robot is within the sensing range of the sensor and is not blocked, and the pheromone is attenuated. k Second-rate;

[0037] When the grid is within the sensing range of the sensor and is not blocked, α Attenuation effect ( α (0,1)); otherwise, α Sedimentation α >1).

[0038] According to some embodiments of the present invention, the pheromone of the vertex is the average value of the pheromones of the four grids around it:

[0039]

[0040] In the formula, P h1 , P h2 , P h3 , P h4 They are the pheromones of the four grids around the vertex respectively.

[0041] According to the computer-readable storage medium of the second aspect of the present invention, a computer program is stored thereon, and when the computer program is executed by a processor, the multi-robot target position allocation method as described in the first aspect of the present invention is implemented.

[0042] The computer-readable storage medium according to the embodiment of the present invention can effectively address the problem of inefficient planning when the robot's field of vision is blocked.

[0043] According to an embodiment of the third aspect of the present invention, an electronic device includes a processor and a memory, wherein the processor is connected to the memory, and the memory is used to store a computer program. When the computer program is executed by the processor, the multi-robot target position allocation method as described in the embodiment of the first aspect of the present invention is implemented.

[0044] The electronic device according to the embodiment of the present invention can effectively deal with the problem of inefficient planning when the robot's field of vision is blocked.

[0045] Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 is a schematic diagram of the robot's field of vision being obstructed when continuous monitoring is performed using existing technology;

[0047] Figure 2 is a grid map of a multi-robot target position allocation method according to an embodiment of the present invention;

[0048] Figure 3 is a schematic diagram of the field of view of a robot according to an embodiment of the present invention;

[0049] Figure 4 is a robot vision modeling diagram according to an embodiment of the present invention;

[0050] Figure 5 Each grid α The effect diagram when the values ​​are the same;

[0051] Figure 6 is a schematic diagram of adjacent conditions according to an embodiment of the present invention;

[0052] Figure 7 is a schematic diagram of jumping out of a self-loop according to an embodiment of the present invention;

[0053] Figure 8 is a flow chart of conflict resolution and oscillation reduction according to an embodiment of the present invention;

[0054] Fig. 9 4 is a diagram showing the simulation effect of the target position allocation method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0055] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and cannot be understood as limiting the present invention.

[0056] This application is based on the inventor's discovery and understanding of the following facts and problems:

[0057] In the task of continuous monitoring in narrow spaces, the robot's field of view is obstructed and presents an irregular shape. Typical continuous monitoring algorithms often treat the robot's field of view as equivalent to a single grid range, which is suitable for continuous monitoring tasks in open spaces. If it is directly applied to narrow spaces, it will lead to inefficient robot trajectory planning. Figure 1 As shown, space D There are narrow passages and complex intersections in the scene. The robot’s field of view in a typical continuous monitoring algorithm is equivalent to a dotted box. D R When the robot is in the channel, its vision is blocked by the channel wall, causing a blind area. D vb (black filled area), blind area of ​​vision D vb The position, shape and size of the robot are related to its spatial position in the environment.

[0058] If the robot's field of view is still equivalent to D R , the blind spot of vision is ignored D vb , the coverage effect is not good; if the robot's field of view is too conservatively reduced to the minimum spacing of the channel (the single gray box grid in the figure), the number of nodes assigned to the robot must be increased to allow the robot to cover the task area once, resulting in redundant track coverage. Therefore, in this scenario, the obstruction of the robot's field of view changes with the change of the robot's position, which is difficult to quantify in existing perception models.

[0059] In addition, factors such as wall obstructions, tortuous passages, and weak light in narrow spaces have a greater impact on robot sensors, which makes the robot's long-distance detection accuracy within its field of view weaker than its close-range detection accuracy. Existing monitoring models rarely take this influence into account, resulting in the robot's perception of the mission area being inconsistent with reality.

[0060] The following describes a multi-robot target position allocation method according to a first embodiment of the present invention with reference to the accompanying drawings.

[0061] According to a multi-robot target position allocation method according to an embodiment of the present invention, the method is used to allocate target positions of multiple robots in a task area, the robots are provided with sensors for monitoring the task area, the task area is provided with multiple channels for the robots to walk, the widths of different channels are different from each other, the sensing range of the sensor is greater than the width of at least part of the channels, and the method includes:

[0062] The task area is evenly divided into multiple grids on the two-dimensional plane to establish a two-dimensional grid map. The maximum side length of the grid is less than or equal to the minimum width of the channel. The robot is regarded as a particle on the two-dimensional grid map, and the vertex of the grid located in the channel is regarded as the target position of the robot. The sensor's perception range is calculated according to the following formula: To model:

[0063] (1)

[0064] In the formula, a is the number of grids covered by the sensor on the two-dimensional grid map; is the area of ​​a single grid;

[0065] Assign values ​​to all grids to initialize the pheromone of the grids, regard the grids that are within the sensing range of the sensor and are not blocked as being visited by the sensor, and perform pheromone attenuation on the visited grids;

[0066] Compute the pheromone of a vertex based on the pheromones of the grid surrounding it;

[0067] To perform location-allocation:

[0068] Get the position of each robot and its adjacent vertices, and select the adjacent vertex with the largest pheromone as the pre-assigned position;

[0069] If the distance between any two robots is less than or equal to the preset distance, the pre-assigned positions are updated with the last positions of the two robots; otherwise:

[0070] If the number of adjacent vertices of the robot is greater than 1, the pre-assigned position is updated with adjacent vertices except the previous position;

[0071] Otherwise, update the pre-assigned position with the only adjacent vertex;

[0072] If the current execution position allocation is w times, and the robot's current position, pre-assigned position and ( w -1)th, w -2) When the positions of the robots cover the four vertices of the rectangle during the second allocation, the robot will jump out of the self-loop and update the pre-allocated positions with the adjacent vertices other than the four vertices of the rectangle;

[0073] If the current execution position allocation is w If the pre-assigned position is the same as the previous position and the current position of the robot is the same as the position at the (w-2)th assignment, the two-point oscillation reduction is performed to update the pre-assigned position with the adjacent vertices except the previous position;

[0074] If the current execution position allocation is wtimes, and the robot's current position, pre-assigned position and ( w -1)th, w -2) times, ( w -3) The positions during the secondary allocation cover the three vertices of the L-shape together. If the number of adjacent vertices is greater than 1, the L-shape oscillation reduction is performed to update the pre-allocated positions with the adjacent vertices other than the three vertices of the L-shape;

[0075] The pre-assigned position is used as the target position, and the robot is moved toward the target position;

[0076] Update the robot's position when it reaches the target position;

[0077] Updates the pheromones of all grids and performs pheromone decay on the visited grids.

[0078] Specifically, the environmental and robot constraints for continuous monitoring in narrow spaces are as follows.

[0079] exist Figure 2 Grid map shown D Established on xO Coordinate system, the dots, rectangles and triangles represent the initial positions of the three robots, and the solid black lines represent obstacles. F g is the set of free vertices not occupied by obstacles, C g is the set of vertices occupied by obstacles.

[0080] robot r k At the mesh vertex p ( x , y ) can only move to its unoccupied adjacent vertices (i.e., the four neighbors):

[0081] (2)

[0082] In the formula, u k For robots r k At the point p The target location; Yes p The four neighborhoods of is from x Direction or y Distance in direction p The nearest vertex. The velocities of multiple robots are of the same magnitude.

[0083] In order to analyze the obstruction of the robot's field of view, we first quantify the robot's theoretical field of view. Assuming that the detection distance of the depth camera in the model is within 3 unit lengths and the detection range is a fan-shaped area, when quantifying the field of view, its effective detection distance is equivalent to 2 unit lengths, and the theoretical field of view is equivalent to Figure 3 The size of the rectangular area is 4×2, so in formula (1) a =8.

[0084] When continuously monitoring in a confined space, the robot’s field of view changes depending on the robot’s position in the space. Figure 3 The thick solid line in the figure represents obstructions such as walls, the dots represent the robot's position, and the arrows connecting the dots represent the robot's field of view. The robot's field of view when unobstructed is , the range of vision when obstructed The white grid area in the dotted box is the blind area of ​​the visual field. It is the grey grid area.

[0085] (3)

[0086] Figure 4 The triangle in the figure represents the robot, whose speed is in the upward direction. The dotted line with an arrow represents the scanning line of the sensor. In the left sub-figure, when any of the 8 thick solid line segments is occupied by a wall, it constitutes a visual field obstruction. The 8 sub-figures (a) to (h) on the right show the visual field obstruction caused by the wall at different positions. The gray area is the visual field blind spot. To avoid missing the blind spot, the entire grid where the gray triangle is located is regarded as the visual field blind spot. Therefore, the visual field blind spot for:

[0087] (4)

[0088] in G 1. G 2. G 3. G 4. G 5. G 8 represent grids ①, ②, ③, ④, ⑤, and ⑧ respectively.

[0089] Viewable area for:

[0090] (5)

[0091] Figure 4 (a) and Figure 4 (b) The middle field of view is ①~③, ⑤~⑧, Figure 4 (c) and Figure 4 (d) The middle field of view is ②~⑧, Figure 4(e) The middle visual field area is ②~④, ⑥~⑧, Figure 4 (f) The visual field area is ③~⑧, Figure 4 (g) The middle field of view is ①, ②, ⑤~⑧, Figure 4 (h) The middle visual field areas are ①~③, ⑤~⑦.

[0092] When there are two or more wall obstructions, the blind area is the union of the above wall obstructions, and the visual field is the intersection of the above wall obstructions. When the robot's speed direction is down, left, and right, the change in the visual field is determined based on the above analysis.

[0093] Update the pheromone of the grid according to the following formula P h :

[0094] (6)

[0095] In the formula, k Indicates that the grid is located at k The robot is within the sensing range of the sensor and is not blocked, and the pheromone decays k Second-rate.

[0096] When the grid is within the sensor's sensing range and is not blocked, α Attenuation effect ( α (0,1)); otherwise, α Sedimentation α >1). When the robot's field of vision is not blocked, the pheromones of the eight grids are all attenuated. If there is a wall blocking the robot's field of vision, the pheromones of the grids in the field of vision are attenuated, and the pheromones of the grids in the blind area are deposited.

[0097] The pheromone of a vertex is the average of the pheromones of the four surrounding grids:

[0098] (7)

[0099] In the formula, P h1 , P h2 , P h3 , P h4 They are the pheromones of the four grids around the vertex.

[0100] When the attenuation coefficient of each grid α When the values ​​are the same, the local grids may have the same pheromones, making the robot trajectory planning inefficient. Figure 5 Attenuation αThe value of 0.9 is used to analyze the pheromone at the initial position of the robot. The solid arrow indicates that the robot is moving to the right. The pheromone values ​​of multiple grids located in the same robot field of view are all 0.9. According to formula (7), the pheromones of the three adjacent vertices at the current position of the robot are 0.95, 0.95, and 0.9 respectively. The large pheromone value at a certain position indicates that the access degree here is low, and it needs to be visited first. Therefore, the vertex with a pheromone value of 0.95 (called the first step position) is selected. At this time, the initial position becomes the adjacent vertex of the first step position; after updating the pheromone of each vertex according to formula (6) and formula (7), the initial position is the adjacent vertex with the largest pheromone at the first step position, so it is chosen to return to the initial position. The trajectory is as shown in dotted line ① and dotted line ②, resulting in inefficient trajectory planning.

[0101] In addition, the sensor has a lower detection accuracy at a farther location, so the pheromone attenuation at this location is lower, and the attenuation coefficient α Therefore, for grids at different positions within the robot's field of view, the pheromone attenuation coefficient α The distance settings are as follows:

[0102] (8)

[0103] ( x g , y y ) is the grid position; ( x , y ) is the current position of the robot, selected based on experience , , .

[0104] Narrow spaces are complex, and robots are prone to conflict and oscillation when performing continuous monitoring tasks. It is not easy to obtain a global optimal solution using existing methods. Conflicts between two robots include being assigned to the same target position and exchanging each other's current positions. Robot oscillation refers to the robot's trajectory cycling back and forth in a local area between consecutive iterations. The present invention designs a conflict resolution strategy and a vibration reduction method to improve the global performance indicators of the algorithm.

[0105] (1) Determination of conflict resolution conditions

[0106] To ensure access efficiency, conflict resolution is triggered when certain conditions occur. The trigger condition is that the robots are close to each other and the Manhattan distance between the two robots is defined as d Less than or equal to D means the robots are adjacent to each other, such as Figure 6 (a) The distance between the two robots in different directions is less than or equal to D, and Figure 6 (b) and Figure 6(c) The distance in the same direction is less than or equal to D. The proximity judgment condition in the figure is D=3. d The calculation formula is:

[0107] (9)

[0108] ( x 1, y 1) and ( x 2, y 2) are the coordinates of the two robots respectively.

[0109] (2) Analysis and determination of oscillation types

[0110] In a narrow space, the robot will produce three types of oscillations: two-point oscillation, L-shaped oscillation, and self-circulation. Assume that the iteration step is w The current position of the robot is s , the robot plans the trajectory , s w-1 The iteration step is ( w -1) the position of the robot, for( w -m) of the robot.

[0111] ① Two-point oscillation judgment conditions

[0112] Figure 6 In (a), the robot oscillates between two adjacent vertices. w Pre-allocated position and Same, robot current position s and The same condition determines that the robot is in two-point oscillation:

[0113] (10)

[0114] (11)

[0115] ②L-shaped oscillation judgment conditions

[0116] Robots in Figure 6 (b) The three adjacent vertices form an "L"-shaped trajectory. and same, and Same as s and Similarly, determine that the robot is in L-shaped oscillation:

[0117] (12)

[0118] (13)

[0119] (14)

[0120] or, s and same, and At the same time, it is determined that the robot is in L-shaped oscillation.

[0121] ③Self-loop determination conditions

[0122] Self-loop means that the robot's position is always Figure 6 One of the four adjacent vertices in (c), the trajectory shape is a rectangle, and the arrows in the figure indicate the position order. Construct a rectangle to determine if the robot is in a self-loop:

[0123] (15)

[0124] (16)

[0125] (17)

[0126] (18)

[0127] Based on the analysis and determination of conflict resolution conditions and oscillation types, the following conflict resolution strategies and vibration reduction methods are designed.

[0128] (i) Turnaround strategy

[0129] When a robot r 1. r 2 When they are close to each other, they adopt the back-turning strategy. r 1. r 2's target position so that they are farthest apart:

[0130] (4.18)

[0131] (4.19)

[0132] (4.20)

[0133] N f1 yes r 1's adjacent vertex set; N f2 yes r 2's adjacent vertex set; The iteration step is w hour p 1i Point and p 2j The distance between yes r 1 Target position u 1. r 2 Target location u The maximum distance between 2; Is the allocation r 1. r 2 is the utility function of the target position.

[0134] When the maximum value is achieved of( p 1i , p 2j ) is multiple pairs, find the vertex with the largest pheromone among these vertices (such as p 11 ) as the corresponding robot ( r 1) target position, another robot ( r 2) The target position is the vertex that is paired with it (such as p 21 In order to make the planned trajectory behave like a random trajectory, when the pre-selected pair of vertices (such as ( p 11 , p 21 )) When robots collide, randomly select the maximum value The remaining pairs of vertices are used as target positions.

[0135] When the maximum value is achieved There is only one pair of vertices ( p 11 , p 21 ) and make the robots collide, if N f1 The number of vertices is greater than N f2 The number of vertices in the p 11 ,choose N f1 The point with the largest pheromone among the remaining vertices is u 1, u 2= p 21 ;like N f1 The number of vertices is less than or equal to N f2 The number of vertices in the p 21 ,chooseN f2 The point with the largest pheromone among the remaining vertices is u 2, u 1= p 11 .

[0136] (ii) No-return strategy

[0137] When the robots are not adjacent to each other, N f For a robot with a vertex number of 1, its target position can only be that vertex. N f For robots with more than 1 vertices, the no-turn-back strategy is adopted: N f Remove For equal vertices, select the point with the largest pheromone among the remaining vertices as the target position u .

[0138] (iii) Break out of the self-loop

[0139] Figure 7 middle p 1. p 2. p 3. p 4 are the four vertices of the white self-circulating rectangle, and the current position of the robot is p 1. Only select points p 5 or points p 6 can jump out of the self-loop, so choose p 5 、p The 6 points with the largest pheromones are used as target locations.

[0140] (iv) Reducing two-point oscillation

[0141] When the robot oscillates at two points, N f Exclude , randomly select one of the remaining adjacent vertices as the target position:

[0142] (4.21)

[0143] In the formula, random() represents a random selection operation.

[0144] (v) Reduce L-shaped oscillations

[0145] When the robot oscillates in an L-shape, N f If the number of vertices is greater than 1, then N f Remove the track Among the vertices in the vertices, select the point with the smallest number of visits among the remaining vertices, that is,

[0146] (4.22)

[0147] (4.23)

[0148] The iteration step is w The minimum number of visits when express q i The number of visits.

[0149] Based on conflict resolution strategies and vibration reduction methods, design Figure 8 The algorithm flow chart is shown in the figure. The priority allocation is designed as follows.

[0150] S1: When the trigger conditions between robots are met When , each robot selects the target position according to the back-tracking strategy.

[0151] S2, when the robots When N f If the number of vertices in is equal to 1, a unique adjacent vertex is assigned; if N f If the number of vertices is greater than 1, the target position is selected according to the no-look-back strategy.

[0152] S3, when the self-loop determination condition is met, the execution jumps out of the self-loop.

[0153] S4, when the two-point oscillation determination condition is met, reducing the two-point oscillation is executed.

[0154] S5, when the L-shaped oscillation determination condition is met, reducing the L-shaped oscillation is executed.

[0155] Fig. 9 A simulation effect diagram of the multi-robot target position allocation method in Gazebo according to an embodiment of the present invention is given. Fig. 9 (a) is the initial position, Fig. 9 (b) is the effect diagram when the area is covered once.

[0156] The multi-robot target position allocation method according to the embodiment of the present invention can effectively address the problem of inefficient planning when the robot's field of vision is blocked, allowing the robot to complete the continuous monitoring task in a narrow space.

[0157] According to the computer-readable storage medium of the second aspect of the present invention, a computer program is stored thereon, and when the computer program is executed by a processor, the multi-robot target position allocation method of the first aspect of the present invention is implemented.

[0158] The computer-readable storage medium according to the embodiment of the present invention can effectively address the problem of inefficient planning when the robot's field of vision is blocked.

[0159] According to an embodiment of the third aspect of the present invention, the electronic device includes a processor and a memory, the processor and the memory are connected, the memory is used to store a computer program, and when the computer program is executed by the processor, the multi-robot target position allocation method as in the embodiment of the first aspect of the present invention is implemented.

[0160] The electronic device according to the embodiment of the present invention can effectively deal with the problem of inefficient planning when the robot's field of vision is blocked.

[0161] Other structures and operations of the electronic device according to the embodiment of the present invention are known to those skilled in the art and will not be described in detail here.

[0162] In the description of the present invention, it is to be understood that the terms "first" and "second" are used for descriptive purposes only and are not to be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, unless otherwise specified, "plurality" means two or more.

[0163] In the description of the present invention, "a first feature" or "a second feature" may include one or more of the features. A first feature "above" or "below" a second feature may include direct contact between the first and second features, or may include contact between the first and second features through another feature between them instead of direct contact. A first feature "above", "above" or "above" a second feature may include the first feature being directly above or obliquely above the second feature, or may simply mean that the first feature is higher in level than the second feature.

[0164] It should be noted that, in the description of the present invention, the description with reference to the terms "one embodiment", "some embodiments", "specific embodiments", "examples" or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0165] Although the embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the claims and their equivalents.

Claims

1. A multi-robot target position allocation method, characterized in that: The method is used to assign target positions of multiple robots in a task area, wherein the robots are provided with sensors for accessing the task area, wherein the task area is provided with multiple passages for the robots to walk, wherein the widths of different passages are different from each other, and the sensing range of the sensors is greater than the widths of at least some of the passages, and the method comprises: The task area is evenly divided into multiple grids on a two-dimensional plane to establish a two-dimensional grid map, the maximum side length of the grid is less than or equal to the minimum width of the channel, the robot is regarded as a particle on the two-dimensional grid map, the vertex of the grid located in the channel is taken as the target position of the robot, and the sensing range of the sensor is calculated according to the following formula: To model: In the formula, a is the number of grids covered by the sensor on the two-dimensional grid map; is the area of ​​a single grid; Assigning values ​​to all the grids to initialize the pheromones of the grids, considering the grids that are within the sensing range of the sensor and are not blocked as being visited by the sensor, and performing pheromone attenuation on the visited grids; Calculating the pheromone of the vertex based on the pheromones of the grid surrounding the vertex; To perform location-allocation: Obtaining the position of each robot and its adjacent vertices, and selecting the adjacent vertex with the largest pheromone as the pre-allocated position; If the distance between any two of the robots is less than or equal to the preset distance, the pre-assigned position is updated with the last position of each of the two robots; otherwise: If the number of the adjacent vertices of the robot is greater than 1, updating the pre-assigned position with the adjacent vertices except the previous position; Otherwise, updating the pre-allocated position with the only adjacent vertex; If the current execution position allocation is w and the current position of the robot, the pre-assigned position and the first ( w -1)th, w -2) When the positions of the robots cover the four vertices of the rectangle during the secondary allocation, the pre-allocated positions are updated with the adjacent vertices other than the four vertices of the rectangle by jumping out of the self-loop; If the current execution position allocation is w The pre-assigned position is the same as the previous position and the current position of the robot is the same as the first ( w -2) If the positions of the second allocation are the same, then reducing the two-point oscillation is performed to update the pre-allocated position with the adjacent vertices except the previous position; If the current execution position allocation is w and the current position of the robot, the pre-assigned position and the first ( w -1)th, w -2) times, ( w -3) the positions in the secondary allocation jointly cover the three vertices of the L-shape, and if the number of the adjacent vertices is greater than 1, reducing the L-shape oscillation is performed to update the pre-allocated positions with the adjacent vertices other than the three vertices of the L-shape; Taking the pre-assigned position as the target position and moving the robot toward the target position; updating the position of the robot when the robot reaches the target position; The pheromones of all the grids are updated, and the pheromone decay is performed on the visited grids.

2. The multi-robot target position allocation method according to claim 1, characterized in that: The execution jumps out of the self-loop further comprises: The adjacent vertex with the largest pheromone is selected as the pre-assigned position.

3. The multi-robot target position allocation method according to claim 1, characterized in that: The performing of reducing two-point oscillations further comprises: One of the adjacent vertices is randomly selected as the pre-assigned position.

4. The multi-robot target position allocation method according to claim 1, characterized in that: The performing of reducing L-shaped oscillations further comprises: The adjacent vertex with the smallest number of visits is selected as the pre-allocated position.

5. The multi-robot target position allocation method according to claim 1, characterized in that: When a portion of the grid is within the sensing range of the sensor and is not blocked, the grid is considered to be visited by the sensor.

6. The multi-robot target position allocation method according to claim 1, characterized in that: The grids with different distances from the same robot have different attenuation levels.

7. The multi-robot target position allocation method according to claim 1, characterized in that: Update the pheromone of the grid according to the following formula P h : In the formula, k Indicates that the grid is located at k The robot is within the sensing range of the sensor and is not blocked, and the pheromone is attenuated. k Second-rate; When the grid is within the sensing range of the sensor and is not blocked, α Attenuation effect, α (0,1); otherwise, α The sedimentation effect, α >1.

8. The multi-robot target position allocation method according to claim 1, characterized in that: The pheromone of the vertex is the average of the pheromones of the four grids around it: In the formula, P h1 , P h2 , P h3 , P h4 They are the pheromones of the four grids around the vertex respectively.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the multi-robot target position allocation method as described in any one of claims 1 to 8 is implemented.

10. An electronic device, characterized in that: The electronic device includes a processor and a memory, the processor is connected to the memory, the memory is used to store a computer program, and when the computer program is executed by the processor, the multi-robot target position allocation method according to any one of claims 1 to 8 is implemented.

Images