A multi-optical intelligent agent coverage control method in a dynamic environment

Abstract

The present application relates to multi-optical intelligent agent coverage control technology, in order to solve the problem of weak autonomy and low observation efficiency of existing search method, a multi-optical intelligent agent coverage control method in dynamic environment is proposed, comprising: arranging n optical intelligent agents in the target area to form an optical network, calculating the initial angle corresponding to the pointing point of a single optical intelligent agent according to the pointing point coordinates of a single optical intelligent agent and the current station point coordinates of a single optical intelligent agent; Vinoy segmentation is carried out on the target area, and the initial allocation result of each optical intelligent agent Vinoy subregion is calculated; according to the distribution of the moving obstacle in the target area, the angle cost required for each optical intelligent agent pointing point to rotate to the target point in the target area is obtained, and then the generalized Vinoy centroid is obtained; the control law of the optical intelligent agent is designed, the gradient descent method is used to update the initial angle of each optical intelligent agent pointing point, and the autonomous planning and control of multi-optical intelligent agent coverage in dynamic environment are completed.

Description

Technical Field

[0001] This invention relates to optimal coverage control technology for multiple photoelectric agents, specifically to a coverage control method for multiple photoelectric agents in a dynamic environment. Background Technology

[0002] An optoelectronic intelligent agent is a system that combines optoelectronic technology and intelligent technology. It is mainly used to realize the perception, recognition and control of objects. Specifically, the optoelectronic intelligent agent realizes the perception and change of ambient light through optoelectronic sensors and intelligent algorithms, thereby making automatic adjustments and controls.

[0003] In large-scale sky surveys, photoelectric agents can be mounted on telescopes to form observation equipment. This equipment has advantages such as long observation distance, strong operability, and the ability to communicate with each other, and has become an important means of astronomical observation. However, due to the limitations of their own observation aperture and imaging time, it takes a long time to complete regional monitoring using a single photoelectric agent, and the observation range is limited. Therefore, multiple photoelectric agents need to cooperate with each other to conduct a comprehensive search of the region.

[0004] In dynamic observation environments, when multiple observation devices perform observation tasks, they usually adopt the method of manually dividing the observation area to realize the joint search operation of multiple photoelectric intelligent agents. However, this search method has weak autonomy, and the observation devices cannot quickly and effectively observe the dynamically changing environment, resulting in low observation efficiency and wasted observation resources. At the same time, this search method lacks a certain degree of autonomous control capability, and cannot reasonably plan the search area or conduct efficient and comprehensive coverage search of areas with target information. Summary of the Invention

[0005] This invention addresses the problems of weak autonomy, low observation efficiency, and inability to perform efficient and comprehensive coverage searches of areas containing target information in existing multi-photoelectric intelligent agent joint search methods. It proposes a multi-photoelectric intelligent agent coverage control method for dynamic environments.

[0006] The design concept of this invention is as follows: mathematical models are established based on the characteristics of the photoelectric intelligent agent and the observation scenario, and the generalized Vino centroid of the dynamic observation area is calculated by combining the covering control theory. The control law of the photoelectric intelligent agent is designed, and the system is eventually stabilized at the generalized Vino centroid, thereby realizing the autonomous planning and control of the multi-photoelectric intelligent agent system in a dynamic environment and improving the search efficiency and autonomy of the system.

[0007] To achieve the above objectives, the technical solution proposed by this invention is as follows:

[0008] A multi-photovoltaic intelligent agent coverage control method for dynamic environments, characterized by the following steps:

[0009] S1. Deploy n photoelectric intelligent agents to form a photoelectric network within the target area. Calculate the initial angle corresponding to the pointing point of a single photoelectric intelligent agent based on the coordinates of the pointing point of a single photoelectric intelligent agent and the coordinates of the current deployment site of a single photoelectric intelligent agent.

[0010] S2. Perform Vino segmentation on the target region according to the initial angle corresponding to the pointing point of each photoelectric agent, and calculate the initial allocation result of the Vino sub-region of each photoelectric agent.

[0011] S3. Based on the distribution of moving obstacles in the target area, obtain the angle cost required for each photoelectric agent's pointing point to rotate to the target point in the target area through the coverage control objective function of each photoelectric agent, and then calculate the generalized Vino centroid of the Vino sub-region of each photoelectric agent.

[0012] S4. Based on the generalized Vino centroid of each photoelectric agent's Vino sub-region, design the control law for the photoelectric agent. Use the gradient descent method to update the initial angle of each photoelectric agent's pointing point, so that each photoelectric agent rotates to the specified position until each photoelectric agent is in the optimal distribution state, thereby completing the autonomous planning and control of multi-photoelectric agent coverage in a dynamic environment.

[0013] Furthermore, step S1 specifically includes:

[0014] S1.1, Define the set of pointing points of n photoelectric intelligent agents arranged in the target area as follows:

[0015] S1.2 Establish a three-dimensional coordinate system and obtain the coordinates of the pointing point of a single photoelectric intelligent agent as: p i =[x i ,y i ,z i ] T The current deployment coordinates of a single photoelectric intelligent agent are: p i0 =[x i0 ,y i0 ,z i0 ] T i = 1, 2, ..., n;

[0016] S1.3. Based on the coordinates of the pointing point of a single photoelectric agent and the current deployment site coordinates of the single photoelectric agent obtained in step S1.2, calculate and obtain the initial angle corresponding to the pointing point of the single photoelectric agent.

[0017] Where, p i =[x i ,y i ,z i ] T Chinese x iLet y be the X-coordinate of the point pointed to by the i-th photoelectric intelligent agent. i Let z be the Y-coordinate of the point pointed to by the i-th photoelectric intelligent agent. i p is the Z-coordinate of the point pointed to by the i-th photoelectric intelligent agent; i0 =[x i0 ,y i0 ,z i0 ] T Chinese x i0 Let y be the X-coordinate of the current deployment site of the i-th photoelectric intelligent agent. i0 Let z be the Y-coordinate of the current deployment site of the i-th photoelectric intelligent agent. i0 Let Z be the Z-coordinate of the current deployment site of the i-th photoelectric intelligent agent.

[0018] Furthermore, in step S1, the coordinates of the pointing point of a single photoelectric agent, the coordinates of the current deployment site of a single photoelectric agent, and the initial angle corresponding to the pointing point of a single photoelectric agent satisfy the following equation:

[0019] When x i >x i0 ,y i >y i0 or x i >x i0 ,y i <y i0 hour,

[0020]

[0021] When x i <x i0 ,y i >y i0 hour,

[0022]

[0023] When x i <x i0 ,y i <y i0 hour,

[0024]

[0025] in, Let be the azimuth angle of the i-th photoelectric intelligent agent. With the Z-axis as the rotation axis, 0 degrees is when the pointing point points to the positive direction of the X-axis, and the positive direction of rotation is the counterclockwise direction along the Z-axis. Let be the pitch angle of the i-th photoelectric agent, with direction following the right-hand rule; when the coordinates of the pointing point of the i-th photoelectric agent coincide with the X-axis, then the azimuth angle of the photoelectric agent is... and pitch angle All are located at the zero position.

[0026] Furthermore, step S2 specifically includes:

[0027] Based on the set of pointing points P of n photoelectric agents, the target region is divided into Vino segments to obtain the initial allocation result of the Vino sub-region of each photoelectric agent, so that each Vino sub-region contains a pointing point of a photoelectric agent.

[0028] When performing Vino segmentation, the initial allocation result of each optoelectronic intelligent agent's Vino sub-region is obtained using the following formula:

[0029]

[0030] Among them, V i Let θ be the Vino subregion of the i-th photoelectric intelligent agent, D be the target region, q be the target point in the target region, q∈D, and θ q Let be the pointing angle when the i-th photoelectric intelligent agent points to the target point q. Let be the current pointing angle of the j-th photoelectric intelligent agent, and N be the set of individual photoelectric intelligent agents;

[0031] The target point q in the target area and its corresponding pointing angle θ q The following conversion relationships exist:

[0032] θ q =π i (q), where π i It is a relational function.

[0033] Furthermore, step S3 specifically includes:

[0034] S3.1. Establish a moving obstacle model within the target area. The moving obstacle model is a Gaussian function. Calculate the environmental information function φ(q,t) within the target area using the following formula:

[0035]

[0036] Where t is the current time, T1 is the time constant, and y t 0 represents the y-axis coordinate of the mean of the environmental information function, and z represents the mean of the environmental information function. t 0 represents the z-axis coordinate of the mean of the environmental information function, and T = 5 is set. t 0 = 1, z t 0 = 2;

[0037] S3.2 Calculate the coverage control objective function H(P,t) for each photoelectric intelligent agent using the following formula, and obtain the minimum value of the coverage control objective function:

[0038]

[0039] in, The measurement cost function represents the direction of a single photoelectric agent pointing to point p. i The angle cost required to rotate to the target point q;

[0040] The The measurement cost function satisfies the following equation:

[0041]

[0042] S3.3. Based on the coverage control objective function of each photoelectric agent, calculate the generalized Vino centroid of the Vino sub-region of each photoelectric agent using the following formula.

[0043]

[0044] in, Let the mass of the i-th optoelectronic intelligent agent's Vino subregion be denoted as .

[0045] Furthermore, step S4 specifically includes:

[0046] S4.1 The control law of the designed optoelectronic intelligent agent satisfies the following equation:

[0047]

[0048] The control model for the i-th photoelectric intelligent agent is:

[0049]

[0050] Among them, u i For the i-th photoelectric intelligent agent control model, Let be the angular acceleration of the i-th photoelectric intelligent agent, ranging from 0 to 20° / s. 2 ;

[0051] The derivation process is as follows:

[0052]

[0053] Where, k p k is the proportionality coefficient. v are the differential coefficients, and k p k v All are greater than zero. Let be the angular velocity of the i-th photoelectric intelligent agent, ranging from 0 to 20° / s;

[0054] S4.2. Use the gradient descent method to find the derivative of the control objective function H(P,t);

[0055] The formula for calculating the derivative of the control objective function H(P,t) is as follows:

[0056]

[0057] The first derivative of the control objective function H(P,t) is calculated:

[0058]

[0059]

[0060] because Both are bounded, that is, when k v When large enough, it satisfies

[0061] Wherein: S i Let ε be the energy function, and ε be the Lyapunov function;

[0062] S4.3. Update the initial angle of each photoelectric agent's pointing point according to the control law of the photoelectric agent in step S4.1, so that each photoelectric agent rotates to the specified position. At that time, each photoelectric intelligent agent is in an optimal distribution state; when If the situation is not ideal, return to step S2 to re-perform the Vino segmentation until all photoelectric agents are in the optimal distribution state, thereby completing the autonomous planning and control of multi-photoelectric agent coverage in a dynamic environment.

[0063] The beneficial effects of this invention are:

[0064] [1] The present invention provides a multi-photoelectric intelligent agent coverage control method in a dynamic environment. By using Vino segmentation to reasonably divide a large target area, each photoelectric intelligent agent avoids cloud cover and is assigned to the best scanning area, so as to achieve comprehensive and coordinated coverage of the target area. The photoelectric intelligent agent control algorithm is designed to allocate resources to the best observation area, which effectively improves the search efficiency of the multi-photoelectric intelligent agents, avoids missed scans and repeated searches, improves the search efficiency and autonomy of the system, saves search time, and ensures the effectiveness and reliability of search data information.

[0065] [2] This invention provides an effective coverage control method for multi-photovoltaic intelligent systems by performing collaborative coverage and observation of the target area through Vino segmentation. It enables real-time control of environments with dynamic cloud cover, reduces human intervention, improves observation efficiency, and realizes large-scale sky survey missions in fields such as environmental monitoring, target orbit determination, and situational awareness. It has strong practicality.

[0066] [3] This invention uses multiple photoelectric intelligent agents as a second-order system, performs Vino segmentation on the target pointing region, solves the generalized Vino centroid of the Vino region, constructs the control objective function at the cost of the rotation range of the photoelectric intelligent agents, designs the control law of the photoelectric intelligent agents by a gradient descent method, updates the pointing points of each photoelectric intelligent agent, and achieves efficient coverage of the observable area in the partially occluded environment of the cloud layer. Attached Figure Description

[0067] Figure 1 This is a flowchart illustrating an embodiment of a multi-photoelectric intelligent agent coverage control method under dynamic conditions according to the present invention;

[0068] Figure 2 This is a schematic diagram of a single photoelectric intelligent agent pointing point in an embodiment of the present invention;

[0069] Figure 3 This is a schematic diagram showing the result of coordinated coverage control of a pentagonal convex planar region by 50 photoelectric intelligent agents in an embodiment of the present invention.

[0070] in:

[0071] (a) is a schematic diagram of coverage control at t = 1s;

[0072] (b) is a schematic diagram of coverage control at t = 10s;

[0073] (c) is a schematic diagram of coverage control at t = 20s;

[0074] (d) is a schematic diagram of coverage control at t = 30s;

[0075] (e) is a schematic diagram of coverage control at t = 40s;

[0076] (f) is a schematic diagram of coverage control at t = 50s;

[0077] (g) is a schematic diagram of coverage control at t = 60s;

[0078] (h) is a schematic diagram of coverage control at t = 70s. Detailed Implementation

[0079] like Figure 1 As shown, a multi-photoelectric intelligent agent coverage control method in a dynamic environment includes the following steps:

[0080] S1. Deploy n photoelectric intelligent agents to form a photoelectric network within the target area. Calculate the initial angle corresponding to the pointing point of a single photoelectric intelligent agent based on the coordinates of the pointing point of a single photoelectric intelligent agent and the coordinates of the current deployment site of a single photoelectric intelligent agent.

[0081] S1.1, Define the set of pointing points of n photoelectric intelligent agents arranged in the target area as follows:

[0082] S1.2, such as Figure 2 As shown, a three-dimensional coordinate system is established, and the coordinates of the pointing point of a single photoelectric intelligent agent are obtained as: p i =[x i ,y i ,z i ] T The current deployment coordinates of a single photoelectric intelligent agent are: p i0 =[x i0 ,y i0 ,z i0 ] T i = 1, 2, ..., n;

[0083] S1.3 Calculate and obtain the initial angle corresponding to the pointing point of a single photoelectric intelligent agent according to step S1.2.

[0084] The coordinates of the pointing point of a single photoelectric agent, the coordinates of the current deployment site of a single photoelectric agent, and the initial angle corresponding to the pointing point of a single photoelectric agent satisfy the following equation:

[0085] When x i >x i0 ,y i >y i0 or x i >x i0 ,y i <y i0 hour,

[0086]

[0087] When x i <x i0 ,y i >y i0 hour,

[0088]

[0089] When x i <x i0 ,y i <y i0 hour,

[0090]

[0091] in, Let be the azimuth angle of the i-th photoelectric intelligent agent. With the Z-axis as the rotation axis, 0 degrees is when the pointing point points to the positive direction of the X-axis, and the positive direction of rotation is the counterclockwise direction along the Z-axis. Let be the pitch angle of the i-th photoelectric agent, with direction following the right-hand rule, and T be the time constant; when the pointing point coordinates of the i-th photoelectric agent coincide with the X-axis, then the azimuth angle of the photoelectric agent is... and pitch angle All are located at the zero position.

[0092] p i =[x i ,y i ,z i ] T Chinese x i Let y be the X-coordinate of the point pointed to by the i-th photoelectric intelligent agent. i Let z be the Y-coordinate of the point pointed to by the i-th photoelectric intelligent agent. i p is the Z-coordinate of the point pointed to by the i-th photoelectric intelligent agent; i0 =[x i0 ,y i0 ,z i0 ] T Chinese x i0 Let y be the X-coordinate of the current deployment site of the i-th photoelectric intelligent agent. i0 Let z be the Y-coordinate of the current deployment site of the i-th photoelectric intelligent agent. i0 Let Z be the Z-coordinate of the current deployment site of the i-th photoelectric intelligent agent.

[0093] S2. Based on the observation effect of each photoelectric agent's pointing point, the target area is divided into Vino segments, and the initial allocation result of the Vino sub-region of each photoelectric agent is calculated.

[0094] Based on the set of pointing points P of n photoelectric agents, the target region is divided into Vino segments to obtain the initial allocation result of the Vino sub-region of each photoelectric agent, so that each Vino sub-region contains a pointing point of a photoelectric agent.

[0095] When performing Vino segmentation, the initial allocation result of each optoelectronic intelligent agent's Vino sub-region is obtained using the following formula:

[0096]

[0097] Among them, V i Let θ be the Vino subregion of the i-th photoelectric intelligent agent, D be the target region, q be the target point in the target region, q∈D, and θ q Let be the pointing angle when the i-th photoelectric intelligent agent points to the target point q. Let be the current pointing angle of the j-th photoelectric intelligent agent, and N be the set of individual photoelectric intelligent agents;

[0098] The target point q in the target area and its corresponding pointing angle θ q The following conversion relationships exist:

[0099] θ q =π i (q), where π i It is a relational function.

[0100] S3. Based on the distribution of moving obstacles within the target area, quantify the initial allocation results through the coverage control objective function of each photoelectric agent, and calculate the generalized Vino centroid of the Vino sub-region of each photoelectric agent in step S2.

[0101] S3.1. Establish a moving obstacle model within the target area. The moving obstacle model is a Gaussian function. Calculate the environmental information function φ(q,t) within the target area using the following formula:

[0102]

[0103] Where t is the current time, y t 0 represents the y-axis coordinate of the mean of the environmental information function, and z represents the mean of the environmental information function. t 0 represents the z-axis coordinate of the mean of the environmental information function, and T = 5 is set. t 0 = 1, z t 0 = 2;

[0104] S3.2 Calculate the coverage control objective function H(P,t) for each photoelectric intelligent agent using the following formula, and obtain the minimum value of the coverage control objective function:

[0105]

[0106] in, The measurement cost function represents the direction of a single photoelectric agent pointing to point p. i The angle cost required to rotate to the target point q;

[0107] The The measurement cost function satisfies the following equation:

[0108]

[0109] S3.3. Based on the coverage control objective function of each photoelectric agent, calculate the generalized Vino centroid of the Vino sub-region of each photoelectric agent using the following formula.

[0110]

[0111] in, Let the mass of the i-th optoelectronic intelligent agent's Vino subregion be denoted as .

[0112] S4. Based on the generalized Vino centroid of each photoelectric agent's Vino sub-region, design the control law for the photoelectric agent. Use the gradient descent method to update the initial angle of each photoelectric agent's pointing point, so that each photoelectric agent rotates to the specified position until each photoelectric agent is in the optimal distribution state, thereby completing the autonomous planning and control of multi-photoelectric agent coverage in a dynamic environment.

[0113] S4.1 The control law of the designed optoelectronic intelligent agent satisfies the following equation:

[0114]

[0115] The control model for the i-th photoelectric intelligent agent is:

[0116]

[0117] Among them, u i This is the control model for the i-th optoelectronic intelligent agent; Let be the angular acceleration of the i-th photoelectric intelligent agent, ranging from 0 to 20° / s. 2 ;

[0118] The derivation process is as follows:

[0119]

[0120] Where, k p k is the proportionality coefficient. v are the differential coefficients, and k p k v All are greater than zero. Let be the angular velocity of the i-th photoelectric intelligent agent, ranging from 0 to 20° / s;

[0121] S4.2. Use the gradient descent method to find the derivative of the control objective function H(P,t);

[0122] The formula for calculating the derivative of the control objective function H(P,t) is as follows:

[0123]

[0124] The first derivative of the control objective function H(P,t) is calculated:

[0125]

[0126] because Both are bounded, that is, when k v When large enough, it satisfies

[0127] Wherein: S i Let ε be the energy function, and ε be the Lyapunov function;

[0128] S4.3. Update the initial angle of each photoelectric agent's pointing point according to the control law of the photoelectric agent in step S4.1, so that each photoelectric agent rotates to the specified position. At that time, each photoelectric intelligent agent is in an optimal distribution state; when If the situation is not ideal, return to step S2 to re-perform the Vino segmentation until all photoelectric agents are in the optimal distribution state, thereby completing the autonomous planning and control of multi-photoelectric agent coverage in a dynamic environment.

[0129] According to the aforementioned multi-electro-optical agent coverage control method, a multi-agent system composed of 50 photoelectric agents is used to collaboratively cover a pentagonal convex planar area. Within this target area, the x-axis coordinates of the photoelectric agent pointing points range from [-1.6, 2], the y-axis range is [-1.9, 1.9], and the z-axis range is 10. The coordinates of the current photoelectric agent deployment point are [-3 + 0.01*(i-5*X), -3 + 0.01*...].

[0130] [(i%5+1),0], the unit is km.

[0131] Where: X represents the largest integer that makes 5X>i, % represents the remainder, that is, the photoelectric intelligent agents are deployed in a square array, and the distance between adjacent intelligent agents in the same row or column is 10m, the mathematical model of information distribution in the dynamic environment is φ(q,t); the maximum speed is 20° / s, and the maximum acceleration is 20° / s2; such as Figure 3 As shown in (a)-(h), the optimal distribution state of the pointing points of each photoelectric intelligent agent is obtained at t=1s, 10s, 20s, 30s, 40s, 50s, 60s and 70s respectively, and the autonomous planning and control of multi-photoelectric intelligent agent coverage in dynamic environment is completed.

Claims

1. A multi-photoelectric intelligent agent coverage control method in a dynamic environment, characterized in that, Includes the following steps: S1. Deploy n photoelectric intelligent agents to form a photoelectric network within the target area. Calculate the initial angle corresponding to the pointing point of a single photoelectric intelligent agent based on the coordinates of the pointing point of a single photoelectric intelligent agent and the coordinates of the current deployment site of a single photoelectric intelligent agent. S2. Perform Vino segmentation on the target region according to the initial angle corresponding to the pointing point of each photoelectric agent, and calculate the initial allocation result of the Vino sub-region of each photoelectric agent. S3. Based on the distribution of moving obstacles in the target area, obtain the angle cost required for each photoelectric agent's pointing point to rotate to the target point in the target area through the coverage control objective function of each photoelectric agent, and then calculate the generalized Vino centroid of the Vino sub-region of each photoelectric agent. S4. Based on the generalized Vino centroid of each photoelectric agent's Vino sub-region, design the control law for the photoelectric agent. Use the gradient descent method to update the initial angle of each photoelectric agent's pointing point, so that each photoelectric agent rotates to the specified position until each photoelectric agent is in the optimal distribution state, thereby completing the autonomous planning and control of multi-photoelectric agent coverage in a dynamic environment.

2. The multi-photoelectric intelligent agent coverage control method in a dynamic environment according to claim 1, characterized in that, Step S1 is as follows: S1.1, Define the set of pointing points of n photoelectric intelligent agents arranged in the target area as follows: S1.2 Establish a three-dimensional coordinate system and obtain the coordinates of the pointing point of a single photoelectric intelligent agent as: p i =[x i ,y i ,z i ] T The current deployment coordinates of a single photoelectric intelligent agent are: p i0 =[x i0 ,y i0 ,z i0 ] T i = 1, 2, ..., n; S1.

3. Based on the coordinates of the pointing point of a single photoelectric agent and the coordinates of the current deployment site of a single photoelectric agent obtained in step S1.2, calculate and obtain the initial angle θ corresponding to the pointing point of a single photoelectric agent. pi .

3. The multi-photoelectric intelligent agent coverage control method in a dynamic environment according to claim 2, characterized in that: In step S1.3, the coordinates of the pointing point of a single photoelectric agent, the coordinates of the current deployment site of a single photoelectric agent, and the initial angle corresponding to the pointing point of a single photoelectric agent satisfy the following equation: When x i >x i0 ,y i >y i0 or x i >x i0 ,y i <y i0 hour, When x i <x i0 ,y i >y i0 hour, When x i <x i0 ,y i <y i0 hour, in, Let be the azimuth angle of the i-th photoelectric intelligent agent. Let be the pitch angle of the i-th photoelectric agent; when the pointing point coordinates of the i-th photoelectric agent coincide with the X-axis, then the azimuth angle of the photoelectric agent is... and pitch angle All are located at the zero position.

4. The multi-photoelectric intelligent agent coverage control method in a dynamic environment according to claim 3, characterized in that, Step S2 is as follows: Based on the set of pointing points P of n photoelectric agents, the target region is divided into Vino segments to obtain the initial allocation result of the Vino sub-region of each photoelectric agent, so that each Vino sub-region contains a pointing point of a photoelectric agent. When performing Vino segmentation, the initial allocation result of the Vino sub-region for each photoelectric agent is obtained by the following formula: Among them, V i Let θ be the Vino subregion of the i-th photoelectric intelligent agent, D be the target region, q be the target point in the target region, q∈D, and θ q Let be the pointing angle when the i-th photoelectric intelligent agent points to the target point q. Let be the current pointing angle of the j-th photoelectric intelligent agent, and N be the set of individual photoelectric intelligent agents; The target point q in the target area and its corresponding pointing angle θ q The following conversion relationships exist: θ q =π i (q), where π i It is a relational function.

5. The multi-photoelectric intelligent agent coverage control method in a dynamic environment according to claim 4, characterized in that, Step S3 is as follows: S3.

1. Establish a moving obstacle model within the target area. The moving obstacle model is a Gaussian function. Calculate the environmental information function φ(q,t) within the target area using the following formula: Where t is the current time, T1 is the time constant, and y t0 The z-axis coordinates of the mean of the environmental information function are: t0 Let T1 = 5, and let y be the z-axis coordinate of the mean of the environmental information function. t0 =1,z t0 =2; S3.2 Calculate the coverage control objective function H(P,t) for each photoelectric agent using the following formula: Where, f(θ) pi ,θ q ) is the measurement cost function, representing the direction of a single photoelectric agent pointing to point p. i The angle cost required to rotate to the target point q; The f(θ) pi ,θ q The measurement cost function satisfies the following equation: S3.

3. Based on the coverage control objective function of each photoelectric agent, calculate the generalized Vino centroid of the Vino sub-region of each photoelectric agent using the following formula. in, Let the mass of the i-th optoelectronic intelligent agent's Vino subregion be denoted as .

6. The multi-photoelectric intelligent agent coverage control method in a dynamic environment according to claim 5, characterized in that, Step S4 is as follows: S4.1 The control law of the designed optoelectronic intelligent agent satisfies the following equation: The control model for the i-th photoelectric intelligent agent is: Among them, u i For the i-th photoelectric intelligent agent control model, Let be the angular acceleration of the i-th photoelectric intelligent agent, ranging from 0 to 20° / s. 2 ; k p k is the proportionality coefficient. v are the differential coefficients, and k p k v All are greater than zero. Let be the angular velocity of the i-th photoelectric intelligent agent, ranging from 0 to 20° / s; S4.

2. Use the gradient descent method to find the derivative of the control objective function H(P,t); The formula for calculating the derivative of the control objective function H(P,t) is as follows: The first derivative of the control objective function H(P,t) is calculated: because Both are bounded, that is, when k v When large enough, it satisfies Wherein: S i Let ε be the energy function, and ε be the Lyapunov function; S4.

3. Update the initial angle of each photoelectric agent's pointing point according to the control law of the photoelectric agent in step S4.1, so that each photoelectric agent rotates to the specified position. At that time, each photoelectric intelligent agent is in an optimal distribution state; when If the situation is not ideal, return to step S2 to re-perform the Vino segmentation until all photoelectric agents are in the optimal distribution state, thereby completing the autonomous planning and control of multi-photoelectric agent coverage in a dynamic environment.

Images