Multi-agent cooperative search and target exploration simulation system and method of use thereof

Abstract

The application discloses a kind of multi-agent cooperative search and target exploration simulation system and its use method, it is related to marine environment detection technical field.The simulation system includes: regional division module, for dividing the task area into multiple sub-regions of different shape and size according to the ability of agent;Comb-like line search module, for global path planning to the task area;Local path re-planning module, for re-planning agent path to avoid obstacle;Task allocation module, for allocating target based on greedy algorithm;And target exploration module, for approaching exploration to the target allocated along the specified angle, and improved artificial potential field method is used to avoid obstacle in exploration stage to dangerous area.This application significantly reduces the failure rate of obstacle avoidance by dynamically adjusting the obstacle avoidance strategy.This application significantly improves the efficiency of target allocation by improving the greedy algorithm.

Description

Technical Field

[0001] This invention relates to the field of marine environmental detection technology, and in particular to a multi-agent cooperative search and target detection simulation system and its usage method. Background Technology

[0002] With the increasing importance of marine environmental exploration and monitoring, and the continuous development of collaborative technologies, multi-agent control models have attracted widespread attention and research in the field of marine environmental exploration. Multi-agent control models are a control theory and method used to study and realize the collaborative work and decision-making among multiple agents, enabling them to jointly solve complex marine exploration tasks. These tasks may occur in complex and variable marine environments, placing higher demands on traditional detection methods. In marine environmental exploration, the detection payloads of traditional platforms are significantly affected by the complexity and variability of the marine environment. Generally, closer-range target observation yields higher target localization accuracy. Therefore, utilizing multi-agent collaboration to complete target exploration in complex sea areas can effectively improve exploration accuracy and reduce safety risks during the exploration process.

[0003] However, there are many problems in building a multi-agent collaborative search and target detection model in the existing technology. On the one hand, the marine environment is complex and changeable, which requires a high level of real-time obstacle avoidance capability from the agents, which is difficult to achieve. On the other hand, the task area is uncertain, and suspicious targets may appear at any time during the search process, resulting in untimely response. Furthermore, the stringent detection conditions, such as the uncertainty of target information, require the agents to approach and detect multiple times at a specified angle, which is time-consuming and labor-intensive. Summary of the Invention

[0004] (I) Purpose of the Invention

[0005] The purpose of this invention is to provide a multi-agent cooperative search and target exploration simulation system and its usage method, so as to solve the above-mentioned technical problems existing in the multi-agent cooperative search and target exploration models in the prior art.

[0006] (II) Technical Solution

[0007] To address the above problems, this invention provides a multi-agent cooperative search and target detection simulation system, comprising:

[0008] The region division module is used to divide the task area into multiple sub-regions of different shapes and sizes according to the agent's capabilities;

[0009] The comb-line search module is used for global path planning within the task area;

[0010] The local path replanning module is used to replan the agent's path for obstacle avoidance.

[0011] The task allocation module is used to allocate targets based on a greedy algorithm; and

[0012] The target detection module is used to approach and detect the assigned target at a specified angle. During the detection phase, obstacle avoidance is achieved by using an improved artificial potential field method to avoid obstacles in dangerous areas.

[0013] Optionally, the target detection module sets the target to be a target that needs to be detected multiple times.

[0014] Optionally, the target detection module is configured to have detection angle requirements and detection interval requirements for the target.

[0015] Optionally, the region division module may set up no-fly zones and areas where the target is threatened within the task region, which the intelligent agent cannot enter.

[0016] In addition, the present invention also provides a method for using a simulation system, wherein the simulation system is the aforementioned multi-agent cooperative search and target detection simulation system, and the method of use includes:

[0017] Step 1: The region segmentation module divides the region into sub-regions based on the agent's capabilities;

[0018] Step 2: Enable the combline search module to perform global path planning for the task area;

[0019] Step 3: Instruct the combline search module to issue a path tracing command;

[0020] Step 4: Expand the local path replanning module to include no-fly zones and target threat areas;

[0021] Step 5: The local path replanning module determines whether the agent has entered the expanded danger zone. If yes, proceed to the next step; otherwise, return to step 3.

[0022] Step 6: Enable the local path replanning module to avoid obstacles by replanning the agent's path;

[0023] Step 7: Instruct the local path replanning module to issue speed and heading control commands;

[0024] Step 8: The task allocation module calculates the time cost of detectable targets using an improved greedy algorithm;

[0025] Step 9: Enable the task allocation module to select the objective with the lowest time cost for each agent;

[0026] Step 10: During the exploration phase, the obstacle avoidance module determines whether the agent enters the expansion threat zone of other targets while moving towards the target.

[0027] Step 11: Enable the obstacle avoidance module in the exploration phase to avoid obstacles in the target's threat area by using an improved artificial potential field method;

[0028] Step 12: If the structure determined in step 10 is yes, then the target detection module will detect the target and the number of times the target needs to be detected will be reduced by one.

[0029] Step 13: The target detection module determines whether the number of times the target needs to be detected is 0. If it is, the target is removed; otherwise, the detection time of the target is recorded and the target enters the cooldown period.

[0030] Step 14: The target detection module determines whether the cooldown period of the agent has ended. If yes, the target cooldown is lifted and the process returns to step 9. Otherwise, the cooling process continues.

[0031] Step 15: In the target exploration phase, determine whether the number of times all targets need to be explored is 0. If yes, continue running; otherwise, end the program.

[0032] Optionally, the steps of the improved greedy algorithm in step 8 include:

[0033] Calculate the travel time t1 from the agent to each detectable target and the target's cooldown time t2;

[0034] Set the larger of the travel time t1 and the target's cooldown time t2 as the time required for the agent to explore the target;

[0035] Find the target with the minimum time cost, target1;

[0036] Calculate the travel time required for the agent to reach target1. If the travel time is greater than the set threshold, stop exploring. If the travel time is less than the threshold, explore the target.

[0037] Optionally, the route replanning step in step 6 includes:

[0038] Calculate the distance between the agent's current position and each endpoint of the restricted area, and set the endpoint with the shortest distance as the first obstacle avoidance waypoint;

[0039] Calculate the deviation of the agent's current position from each endpoint of the restricted area and the agent's current heading, and set the endpoint with the smallest deviation as the second obstacle avoidance waypoint;

[0040] The agent is instructed to navigate through obstacle avoidance waypoints in sequence to avoid obstacles.

[0041] Optionally, the improved artificial potential field algorithm in step 11 includes:

[0042] Determine whether the agent has entered the expanded danger zone;

[0043] If it is determined that the agent has entered the expanded danger zone, then calculate the direction of the gravitational force of the target on the agent and the direction of the repulsive force of the dangerous target on the agent;

[0044] The agent's desired direction is calculated based on the attraction and repulsion forces.

[0045] Issue speed and heading control commands to perform obstacle avoidance.

[0046] (III) Beneficial Effects

[0047] The above-described technical solution of the present invention has the following beneficial technical effects:

[0048] This invention addresses the challenge 1: the difficulty of obstacle avoidance under complex and variable marine environments and multiple agent constraints. Considering the characteristics of different mission stages, it combines two obstacle avoidance strategies. By designing a local path replanning module and an obstacle avoidance module for the exploration phase, it dynamically adjusts to adapt to different operating environments, significantly reducing the obstacle avoidance failure rate.

[0049] This invention addresses challenges 2 and 3: the uncertainty of the task environment, where threats can appear at any time during the search and exploration process. By improving the greedy algorithm in the task allocation module, comprehensively considering travel time and cooldown time as constraints, and constructing a cost function, the efficiency of target allocation is significantly improved. Furthermore, by improving the artificial potential field method in the obstacle avoidance module during the exploration phase, the problem of the agent easily entering other target threat areas while traveling to the target is solved. Attached Figure Description

[0050] Figure 1 This is an overall flowchart of the cooperative search and target exploration in a multi-agent cooperative search and target exploration simulation system according to a specific embodiment of the present invention;

[0051] Figure 2 This is a schematic diagram of region partitioning and combline search in a multi-agent cooperative search and target detection simulation system according to a specific embodiment of the present invention.

[0052] Figure 3 Another schematic diagram of region partitioning and combline search in a multi-agent cooperative search and target exploration simulation system according to a specific embodiment of the present invention;

[0053] Figure 4 This is a schematic diagram of local path replanning in a multi-agent cooperative search and target exploration simulation system according to a specific embodiment of the present invention;

[0054] Figure 5 This is a schematic diagram of an improved artificial potential field in a multi-agent cooperative search and target detection simulation system according to a specific embodiment of the present invention.

[0055] Figure 6This is a schematic diagram of target exploration in a multi-agent cooperative search and target exploration simulation system according to a specific embodiment of the present invention. Detailed Implementation

[0056] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments and the accompanying drawings. It should be understood that these descriptions are merely exemplary and not intended to limit the scope of the invention. Furthermore, descriptions of well-known structures and techniques are omitted in the following description to avoid unnecessarily obscuring the concept of the invention.

[0057] The accompanying drawings illustrate a layer structure according to an embodiment of the present invention. These drawings are not to scale, and some details have been enlarged for clarity, and some details may have been omitted. The shapes of the various regions and layers shown in the drawings, as well as their relative sizes and positional relationships, are merely exemplary and may deviate from reality due to manufacturing tolerances or technical limitations. Furthermore, those skilled in the art can design regions / layers with different shapes, sizes, and relative positions as needed.

[0058] Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0059] Furthermore, the technical features involved in the different embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0060] The invention will now be described in more detail with reference to the accompanying drawings. In the various drawings, the same elements are indicated by similar reference numerals. For clarity, the various parts in the drawings are not drawn to scale.

[0061] For existing components that do not involve the improvements of this invention, they will be briefly described or not described at all, while the focus will be on describing the components that have been improved relative to the prior art.

[0062] See Figures 1 to 6 This embodiment provides a multi-agent cooperative search and target detection simulation system, including:

[0063] The region division module is used to divide the task area into multiple sub-regions of different shapes and sizes according to the agent's capabilities;

[0064] The comb-line search module is used for global path planning within the task area;

[0065] The local path replanning module is used to replan the agent's path for obstacle avoidance.

[0066] The task allocation module is used to allocate targets based on a greedy algorithm; and

[0067] The target detection module is used to approach and detect the assigned target at a specified angle. During the detection phase, obstacle avoidance is achieved by using an improved artificial potential field method to avoid obstacles in dangerous areas.

[0068] See Figures 1 to 6 Furthermore, the target detection module sets the target as one that needs to be detected multiple times.

[0069] See Figures 1 to 6 Furthermore, the target detection module sets the target to have detection angle requirements and detection interval requirements.

[0070] See Figures 1 to 6 Furthermore, the region division module sets up no-fly zones and areas where the target is threatened within the task region, which the intelligent agent cannot enter.

[0071] In other words, the target needs to be explored multiple times, with requirements for exploration angle and exploration interval; the mission area contains no-fly zones, and the target is in a threat area that the agent cannot enter.

[0072] In addition, this embodiment provides a method for using a simulation system, wherein the simulation system is the aforementioned multi-agent cooperative search and target detection simulation system, and the method of use includes:

[0073] Step 1: The region segmentation module divides the region into sub-regions based on the agent's capabilities;

[0074] Step 2: Enable the combline search module to perform global path planning for the task area;

[0075] Step 3: Instruct the combline search module to issue a path tracing command;

[0076] Step 4: Expand the local path replanning module to include no-fly zones and target threat areas;

[0077] Step 5: The local path replanning module determines whether the agent has entered the expanded danger zone. If yes, proceed to the next step; otherwise, return to step 3.

[0078] Step 6: Enable the local path replanning module to avoid obstacles by replanning the agent's path;

[0079] Step 7: Instruct the local path replanning module to issue speed and heading control commands;

[0080] Step 8: The task allocation module calculates the time cost of detectable targets using an improved greedy algorithm;

[0081] Step 9: Enable the task allocation module to select the objective with the lowest time cost for each agent;

[0082] Step 10: During the exploration phase, the obstacle avoidance module determines whether the agent enters the expansion threat zone of other targets while moving towards the target.

[0083] Step 11: Enable the obstacle avoidance module in the exploration phase to avoid obstacles in the target's threat area by using an improved artificial potential field method;

[0084] Step 12: If the structure determined in step 10 is yes, then the target detection module will detect the target and the number of times the target needs to be detected will be reduced by one.

[0085] Step 13: The target detection module determines whether the number of times the target needs to be detected is 0. If it is, the target is removed; otherwise, the detection time of the target is recorded and the target enters the cooldown period.

[0086] Step 14: The target detection module determines whether the cooldown period of the agent has ended. If yes, the target cooldown is lifted and the process returns to step 9. Otherwise, the cooling process continues.

[0087] Step 15: In the target exploration phase, determine whether the number of times all targets need to be explored is 0. If yes, continue running; otherwise, end the program.

[0088] See Figures 1 to 6 Furthermore, the steps of the improved greedy algorithm in step 8 include:

[0089] Calculate the travel time t1 from the agent to each detectable target and the target's cooldown time t2;

[0090] Set the larger of the travel time t1 and the target's cooldown time t2 as the time required for the agent to explore the target;

[0091] Find the target with the minimum time cost, target1;

[0092] Calculate the travel time required for the agent to reach target1. If the travel time is greater than the set threshold, stop exploring. If the travel time is less than the threshold, explore the target.

[0093] The target's cooldown time is the target's detection interval (the interval between the current time and the last detection time).

[0094] Greedy algorithms are prone to getting trapped in local optima. By comprehensively considering travel time, target cooldown time, and number of times to be explored, and by setting a time cost threshold, it is possible to effectively prevent the agent from getting trapped in local optima.

[0095] See Figures 1 to 6 Furthermore, the route replanning step in step 6 includes:

[0096] Calculate the distance between the agent's current position and each endpoint of the restricted area, and set the endpoint with the shortest distance as the first obstacle avoidance waypoint;

[0097] Calculate the deviation of the agent's current position from each endpoint of the restricted area and the agent's current heading, and set the endpoint with the smallest deviation as the second obstacle avoidance waypoint;

[0098] The agent is instructed to navigate through obstacle avoidance waypoints in sequence to avoid obstacles.

[0099] By calculating the distance between the agent and each endpoint of the restricted area, as well as the deviation of the agent's heading from each endpoint of the restricted area, obstacle avoidance by the agent in the restricted area can be achieved simply and efficiently, while meeting real-time requirements.

[0100] See Figures 1 to 6 Furthermore, the improved artificial potential field algorithm in step 11 includes:

[0101] Determine whether the agent has entered the expanded danger zone;

[0102] If it is determined that the agent has entered the expanded danger zone, then calculate the direction of the gravitational force of the target on the agent and the direction of the repulsive force of the dangerous target on the agent;

[0103] The desired orientation of the agent is calculated based on gravity and repulsion.

[0104] Issue speed and heading control commands to perform obstacle avoidance.

[0105] The improved artificial potential field method no longer calculates the direction of the resultant force, thus avoiding the situation where the resultant force is 0 when the magnitudes of the attraction and repulsion forces are equal and their directions are opposite, causing the agent to fall into a local optimum.

[0106] The artificial potential field method calculates the repulsive force of obstacles on the agent and the attractive force of the target on the agent, and calculates the resultant force to guide the agent to avoid obstacles. When the attractive and repulsive forces are equal in magnitude and opposite in direction, the agent gets trapped in a local optimum. This invention solves the problem of the artificial potential field method getting trapped in local optima by changing the attractive and repulsive forces to the desired direction, ensuring that the calculation result points to the side of the agent and the obstacle. It is a conventional method.

[0107] See Figure 1 The multi-agent collaborative search and target exploration simulation system is divided into two phases: search and exploration. The region division module divides the region into sub-regions based on the agent's capabilities. The combline search module performs global path planning. The local path replanning module replans the path for obstacle avoidance during the search phase. The task allocation module assigns targets to each agent based on a greedy algorithm. The obstacle avoidance module in the exploration phase performs obstacle avoidance by using an improved artificial potential field method. The target exploration module explores the target.

[0108] For implementation of region division and combline search, see Figure 2 The capability function of the agent is constructed based on its detection radius r and travel speed v as follows:

[0109]

[0110] Where n is the number of agents, the task area is divided into sub-regions of corresponding ability sizes according to the agent's capabilities.

[0111] See Figure 3 The agent, using the bottom boundary as a reference in its respective sub-region, first doubles the exploration radius in latitude and calculates the intersection point of that latitude with the boundary of the task area. Then, it doubles the exploration radius and calculates the intersection point of the corresponding latitude with the boundary of the task area. All intersection points are arranged in a comb pattern to generate the agent's global planning path, and finally, the path tracking command is issued.

[0112] For implementing local path replanning, see Figure 4 First, the no-fly zone is expanded, with the expansion length exceeding the agent's speed multiplied by the inertial navigation data update interval, as shown in the diagram, expanding to ABCD. After each inertial navigation data update, it is determined whether the agent has entered the expanded no-fly zone. If so, the distances between the agent's current position and each endpoint of the expanded no-fly zone are calculated. The endpoint with the shortest distance is selected as the first obstacle avoidance waypoint. The azimuth angles of the lines connecting the agent to each endpoint of the expanded no-fly zone are calculated and subtracted from the agent's current heading angle. The endpoint with the smallest azimuth deviation is selected as the second obstacle avoidance waypoint. By adding two obstacle avoidance waypoints to the agent, path replanning is performed to achieve obstacle avoidance.

[0113] For implementing improved artificial potential field obstacle avoidance, see Figure 5 The target itself has a threat zone that the agent cannot enter. Obstacle avoidance is determined by constantly checking whether the distance between the agent and the target is less than the target's threat radius plus its expansion radius (the expansion radius must be greater than the agent's speed * inertial navigation data update interval). If the agent enters the expansion threat range of another target while en route to the target being explored, the agent will experience both the repulsive force of that target and the gravitational force of the target being explored. To avoid a situation where the gravitational and repulsive forces are equal in magnitude and opposite in direction, resulting in a net force of 0 and the agent getting trapped in a local optimum, the gravitational force is changed to a gravitational heading, and the repulsive force is changed to a repulsive heading. The desired heading of the agent is obtained by calculating the angle bisector of the smaller angle formed by the two headings, until the agent escapes the expansion threat range of that target.

[0114] For the implementation of target reconnaissance, see Figure 6The target must have an effective detection angle, and the agent must have an effective detection range. Only when both conditions are met can the agent be considered to have performed an effective detection of the target. After assigning the target, the agent first calculates the detection radius r that meets the effective detection conditions (smaller than the agent's detection radius and smaller than the target's effective detection radius). Then, it calculates the intersection point A of the line connecting the agent and the target and the circle containing the detection radius r. Point A is used as the agent's target waypoint, and path tracking control commands are issued. After reaching point A, the agent moves in a circle around the target. When it enters the target's effective detection range, the effective detection is completed.

[0115] Additionally, it should be noted that the appendix Figure 3 The detection radius of batch number 4 is 300m, and the detection radius of batch number 31 is 250m.

[0116] It should be understood that the specific embodiments described above are merely illustrative or explanatory of the principles of the invention and do not constitute a limitation thereof. Therefore, any modifications, equivalent substitutions, improvements, etc., made without departing from the spirit and scope of the invention should be included within the protection scope of the invention. Furthermore, the appended claims are intended to cover all variations and modifications falling within the scope and boundaries of the appended claims, or equivalent forms of such scope and boundaries.

Claims

1. A method for using a multi-agent cooperative search and target detection simulation system, characterized in that, include: The region division module is used to divide the task area into multiple sub-regions of different shapes and sizes according to the agent's capabilities; The comb-line search module is used for global path planning within the task area; The local path replanning module is used to replan the agent's path for obstacle avoidance. The task allocation module is used to allocate targets based on a greedy algorithm; The target detection module is used to approach and detect the assigned target at a specified angle. During the detection phase, obstacle avoidance is achieved by using an improved artificial potential field method to avoid obstacles in dangerous areas. Usage instructions include: Step 1: The region partitioning module divides the region into sub-regions based on the agent's capabilities; specifically, it constructs the agent's capability function based on the agent's detection radius r and travel speed v: Where n is the number of agents, and the task area is divided into sub-regions of corresponding ability sizes according to the abilities of the agents; Step 2: Enable the combline search module to perform global path planning for the task area; specifically: The agent first doubles the exploration radius in the latitude of its respective sub-region, calculates the intersection of the latitude with the boundary of the task region, and then doubles the exploration radius to calculate the intersection of the corresponding latitude with the boundary of the task region. All intersections are arranged in a comb pattern to generate the global planning path of the agent. Step 3: Instruct the combline search module to issue a path tracing command; Step 4: Expand the local path replanning module to include the no-fly zone and target threat area; specifically: Calculate the distance between the agent's current position and each endpoint of the restricted area, and set the endpoint with the shortest distance as the first obstacle avoidance waypoint; calculate the deviation between the agent's current position and each endpoint of the restricted area and the agent's current heading, and set the endpoint with the smallest deviation as the second obstacle avoidance waypoint; make the agent pass through the obstacle avoidance waypoints in sequence to avoid obstacles. Step 5: The local path replanning module determines whether the agent has entered the expanded danger zone. If yes, proceed to the next step; otherwise, return to step 3. Step 6: Enable the local path replanning module to avoid obstacles by replanning the agent's path; Step 7: Instruct the local path replanning module to issue speed and heading control commands; Step 8: The task allocation module calculates the time cost of detectable targets using an improved greedy algorithm. The steps of the improved greedy algorithm include: calculating the travel time t1 from the agent to each detectable target and the target's cooldown time t2; setting the larger value between the travel time t1 and the target's cooldown time t2 as the time required for the agent to detect the target; finding the target with the minimum time cost, target1; calculating the travel time required for the agent to reach target1, stopping detection if it is greater than the set threshold, and detecting the target if it is less than the threshold. Step 9: Enable the task allocation module to select the objective with the lowest time cost for each agent; Step 10: During the exploration phase, the obstacle avoidance module determines whether the agent enters the expansion threat zone of other targets while moving towards the target. Step 11: Enable the obstacle avoidance module in the exploration phase to avoid obstacles in the target's threat area by using an improved artificial potential field method; Step 12: If the judgment result in step 10 is yes, then the target detection module will detect the target and the number of times the target needs to be detected will be reduced by one. Step 13: The target detection module determines whether the number of times the target needs to be detected is 0. If it is, the target is removed; otherwise, the detection time of the target is recorded and the target enters the cooldown period. Step 14: The target detection module determines whether the cooldown period of the agent has ended. If yes, the target cooldown is lifted and the process returns to step 9. Otherwise, the cooling process continues. Step 15: In the target exploration phase, determine whether the number of times all targets need to be explored is 0. If yes, continue running; otherwise, end the program.

2. The method of using the multi-agent cooperative search and target detection simulation system according to claim 1, characterized in that, The target detection module sets the target as one that needs to be detected multiple times.

3. The method of using the multi-agent cooperative search and target detection simulation system according to claim 1, characterized in that, The target detection module sets the target to have detection angle requirements and detection interval requirements.

4. The method of using the multi-agent cooperative search and target detection simulation system according to any one of claims 1 to 3, characterized in that, The region division module sets up no-fly zones and areas where targets pose a threat within the task area, which the agent cannot enter.

5. The method of using the multi-agent cooperative search and target detection simulation system according to claim 1, characterized in that, The steps in step 8 of the improved greedy algorithm include: Calculate the travel time t1 from the agent to each detectable target and the target's cooldown time t2; Set the larger of the travel time t1 and the target's cooldown time t2 as the time required for the agent to explore the target; Find the target with the minimum time cost, target1; Calculate the travel time required for the agent to reach target1. If the travel time is greater than the set threshold, stop exploring. If the travel time is less than the threshold, explore the target.

6. The method of using the multi-agent cooperative search and target detection simulation system according to claim 1, characterized in that, The steps in step 6 for replanning the route include: Calculate the distance between the agent's current position and each endpoint of the restricted area, and set the endpoint with the shortest distance as the first obstacle avoidance waypoint; Calculate the deviation of the agent's current position from each endpoint of the restricted area and the agent's current heading, and set the endpoint with the smallest deviation as the second obstacle avoidance waypoint; The agent is instructed to navigate through obstacle avoidance waypoints in sequence to avoid obstacles.

7. The method of using the multi-agent cooperative search and target detection simulation system according to claim 1, characterized in that, The improved artificial potential field algorithm in step 11 includes: Determine whether the agent has entered the expanded danger zone; If it is determined that the agent has entered the expanded danger zone, then calculate the direction of the gravitational force of the target on the agent and the direction of the repulsive force of the dangerous target on the agent; The agent's desired direction is calculated based on the attraction and repulsion forces. Issue speed and heading control commands to perform obstacle avoidance.

Images