Method and device for autonomously avoiding obstacles and bypassing obstacles based on artificial intelligence, and medium

Abstract

The invention relates to the technical field of artificial intelligence, and particularly discloses an artificial intelligence-based autonomous obstacle avoidance and obstacle bypassing method and device and a medium, and the method comprises the steps: firstly collecting the surrounding environment information of a target mobile device in real time, and constructing a virtual environment model; and then integrating a device starting point, a target position, a three-dimensional environment map in the virtual environment model and obstacle information, and building a path planning information database. And performing initial path planning by using an ant colony algorithm in combination with the database to obtain an initial path from the starting point to the target position. And then evaluating the initial path from multiple aspects based on the virtual environment model to obtain an evaluation result. If the evaluation result does not meet the preset standard, optimizing the initial path by adopting a particle swarm algorithm so as to obtain a better path; through comprehensive data acquisition and processing, effective algorithm application and strict evaluation optimization, a path which is higher in quality and better meets requirements can be planned for target mobile equipment.

Description

Technical Field

[0001] The present invention relates to the technical field of artificial intelligence, and particularly to a method, device and medium for autonomous obstacle avoidance and detouring around obstacles based on artificial intelligence. Background Art

[0002] In the era of rapid technological development today, the applications of autonomous mobile devices are constantly expanding in various fields, from material handling robots in industrial production, to domestic cleaning robots in daily life, to autonomous driving vehicles in the field of transportation, etc. During the operation of these autonomous mobile devices, being able to accurately and efficiently achieve autonomous obstacle avoidance and detouring around obstacles is the key to ensuring their safe and stable operation.

[0003] Currently, there are various technical means for autonomous obstacle avoidance and detouring around obstacles. When constructing a virtual environment model, existing methods often have difficulty comprehensively and finely restoring the complex situation of the real environment. For example, they cannot accurately present the shape, size, material and other characteristics of obstacles, as well as the dynamic change factors in the environment. For path planning, traditional methods usually calculate based on simple geometric rules, fixed mathematical models or empirical formulas. These methods may be able to play a certain role when dealing with relatively simple and static environments, but when facing actual scenarios with complex changes, a large number of obstacles and dynamic change factors, they often seem powerless. They are difficult to quickly and effectively process a large amount of environmental data, with low calculation efficiency. The planned paths may not be optimized, having problems such as too long paths, too many turns, and too close distances to obstacles. This will not only reduce the operating efficiency of the mobile device, but also increase the risk of collision, and cannot fully meet the obstacle avoidance requirements of autonomous mobile devices in complex environments.

[0004] Therefore, the present invention proposes a method, device and medium for autonomous obstacle avoidance and detouring around obstacles based on artificial intelligence. Summary of the Invention

[0005] The present invention provides a method, device and medium for autonomous obstacle avoidance and detouring around obstacles based on artificial intelligence, including: S1 Real-time collecting environmental information to construct a virtual environment model, providing an accurate environmental basis for subsequent path planning, and improving the authenticity and feasibility of path planning. S2 Integrating a variety of key information to construct a path planning information database, ensuring the comprehensiveness and accuracy of the data required for planning. S3 Using the ant colony algorithm for initial path planning, which can quickly find a possible path and improve the planning efficiency. S4 Conducting multi-faceted evaluation on the initial path, comprehensively considering the advantages and disadvantages of the path, and ensuring the quality of path planning. S5 When the evaluation does not meet the preset criteria, using the particle swarm algorithm to optimize the path, further enhancing the rationality and adaptability of the path. Through comprehensive data collection and processing, effective algorithm application, and strict evaluation and optimization, this solution can plan a better and more compliant path for the target mobile device.

[0006] The present invention provides a method for autonomous obstacle avoidance and detouring around obstacles based on artificial intelligence, including:

[0007] S1: Collect the surrounding environment information of the target mobile device in real time and construct a virtual environment model;

[0008] S2: Integrate the starting point, target position information, three-dimensional environment map information and obstacle information in the virtual environment model of the target mobile device to construct a path planning information database;

[0009] S3: Use the ant colony algorithm and the path planning information database to perform initial path planning to obtain an initial path from the starting point of the target mobile device to the target position;

[0010] S4: Perform multi-directional evaluation on the initial path based on the virtual environment model to obtain the multi-directional evaluation result of the initial path;

[0011] S5: When the multi-directional evaluation result of the initial path does not meet the preset standard, the particle swarm algorithm is used to optimize the initial path to obtain a better path.

[0012] Preferably, for the method for autonomous obstacle avoidance and detouring around obstacles based on artificial intelligence, S1: Collect the surrounding environment information of the target mobile device in real time and construct a virtual environment model, including:

[0013] Use the multi-sensor system mounted on the target mobile device to collect the surrounding environment information of the target mobile device, where the multi-sensor system includes a lidar, a camera, and an ultrasonic sensor;

[0014] Construct a virtual environment model based on the surrounding environment information of the target mobile device.

[0015] Preferably, for the method for autonomous obstacle avoidance and detouring around obstacles based on artificial intelligence, constructing a virtual environment model based on the surrounding environment information of the target mobile device includes:

[0016] Construct a three-dimensional environment map based on the surrounding environment information of the target mobile device, where the three-dimensional environment map includes the position, shape and size of the obstacle and the relative distance information of the target mobile device;

[0017] Create a fixed object model corresponding to the real environment based on the fixed object information in the surrounding environment information of the target mobile device;

[0018] Synchronously map the dynamic quantization environment information and the three-dimensional environment map in the surrounding environment information of the target mobile device to the fixed object model to obtain a virtual environment model.

[0019] Preferably, for the method of autonomous obstacle avoidance and detouring around obstacles based on artificial intelligence, S3: Use the ant colony algorithm and the path planning information database to perform initial path planning to obtain an initial path from the starting point of the target mobile device to the target location, including:

[0020] Set the number of ants m, determine the pheromone evaporation coefficient ρ where ρ ∈ [0, 1], and determine the pheromone intensity Q;

[0021] Initialize the pheromone matrix, and initialize the heuristic information matrix based on the obstacle information and target location information in the path planning information database;

[0022] Calculate the probability that ant k at node i chooses to move to node j based on the initialized pheromone matrix and heuristic information matrix

[0023] Based on the preset probability rule and the probability that each ant chooses to move from the current node to each node in the next node set, determine the next node that the ant chooses to move to;

[0024] Based on the pheromone evaporation coefficient ρ, the pheromone intensity Q, and the number of ants m, continue to update the pheromone concentration and continuously determine the next node that the ant chooses to move to until the termination condition is met, and obtain the selection paths of all ants;

[0025] Among the selection paths of all ants, select the selection path with the shortest path length as the initial path.

[0026] Preferably, for the method of autonomous obstacle avoidance and detouring around obstacles based on artificial intelligence, calculate the probability that ant k at node i chooses to move to node j based on the initialized pheromone matrix and heuristic information matrix Including:

[0027]

[0028] Where τ ij (t) is the pheromone concentration on the path from node i to node j at time t, α is the pheromone relative importance factor, η ij is the heuristic information from node i to node j, β is the heuristic information relative importance factor, s is an index variable for summation and represents any node in the next node set allowd that ant k can choose k in it.

[0029] Preferably, for the method of autonomous obstacle avoidance and detouring around obstacles based on artificial intelligence, continue to update the pheromone concentration based on the pheromone evaporation coefficient ρ, the pheromone intensity Q, and the number of ants m, and continuously determine the next node that the ant chooses to move to, including:

[0030] Volatile part: For each path (i, j), the pheromone concentration volatilizes according to the following formula:

[0031] τ ij (t + 1) = (1 - ρ)τ ij (t)

[0032] Enhancement part: For each path (i, j), the pheromone concentration enhances according to the following formula:

[0033]

[0034] Among them, τ ij (t + 1) is the pheromone concentration on the path from node i to node j at time t + 1. The amount of pheromone increased for the path (i, j) passed by ant k is:

[0035]

[0036] Among them, L k is the path length around ant k.

[0037] Preferably, for the method of autonomous obstacle avoidance and detouring around obstacles based on artificial intelligence, S4: Multidirectional evaluation of the initial path is performed based on the virtual environment model to obtain the multidirectional evaluation result of the initial path, including:

[0038] Determine the minimum distance between each path segment in the initial path and the obstacle based on the virtual environment model and the initial path, and calculate the safety evaluation index of the initial path:

[0039]

[0040] In the formula, S is the safety evaluation index of the initial path, q is the total number of path segments included in the initial path, e is the natural constant and its value is 2.71828, d min,p is the minimum distance between the p-th path segment in the initial path and the obstacle, d safe is the preset safety distance;

[0041] Calculate the energy consumption when the target mobile device moves along the initial path:

[0042]

[0043] In the formula, E is the energy consumption when the target mobile device moves along the initial path, E 0 is the energy consumed per unit distance when the target mobile device moves in the horizontal direction, b is the number of nodes included in the initial path, x a+1 is the abscissa value of the (a + 1)-th node in the initial path on the two-dimensional plane, x aThe abscissa value of the a-th node in the initial path on the two-dimensional plane, y a+1 The ordinate value of the (a + 1)-th node in the initial path on the two-dimensional plane, y a The ordinate value of the a-th node in the initial path on the two-dimensional plane, cosθ a The cosine value of the slope angle of the initial path, E 1 The energy consumed per unit distance when the target mobile device moves in the vertical direction, sinθ a The sine value of the slope angle of the initial path, E 2 The energy consumption coefficient related to the path tortuosity, γ a The included angle between two path segments connected to the a-th node in the initial path;

[0044] The path length of the initial path, the safety evaluation index, and the energy consumption when the target mobile device moves along the initial path are regarded as the multi-faceted evaluation results of the initial path.

[0045] Preferably, the method for autonomous obstacle avoidance and bypassing obstacles based on artificial intelligence further includes:

[0046] During the process of the target mobile device moving along the initial path or a better path, continuously and real-time collect the surrounding environment of the target mobile device to obtain the real-time surrounding environment information of the target mobile device;

[0047] When it is determined based on the real-time surrounding environment information of the target mobile device that the obstacle information in the current virtual environment model has changed, then feedback the new obstacle information to the virtual environment model to obtain the latest virtual environment model;

[0048] Based on the latest virtual environment model, perform real-time adjustment on the initial path or better path executed by the target mobile device to obtain a real-time adjusted path;

[0049] Send the real-time adjusted path to the motion control system of the target mobile device, control the target mobile device to move along the real-time adjusted path, and at the same time, feedback the real-time motion state of the target mobile device and the latest virtual environment model to the path planning information database.

[0050] The present invention provides a device for autonomous obstacle avoidance and bypassing obstacles based on artificial intelligence, which is used to execute any one of the above methods for autonomous obstacle avoidance and bypassing obstacles based on artificial intelligence, including:

[0051] An environment acquisition module, which is used to collect the surrounding environment information of the target mobile device in real time and construct a virtual environment model;

[0052] An information integration module for integrating the starting point of the target mobile device, the target location information, the three-dimensional environmental map information in the virtual environment model, and the obstacle information to construct a path planning information database;

[0053] An initial planning module for performing initial path planning using the ant colony algorithm and the path planning information database to obtain an initial path from the starting point of the target mobile device to the target location;

[0054] A path evaluation module for multi-directionally evaluating the initial path based on the virtual environment model to obtain a multi-directional evaluation result of the initial path;

[0055] A path optimization module for, when the multi-directional evaluation result of the initial path does not meet the preset standard, using the particle swarm algorithm to optimize the initial path to obtain a better path.

[0056] The present invention provides a readable storage medium with a program or instructions stored thereon. When the program or instructions are executed by a processor, the steps of any one of the above methods for autonomous obstacle avoidance and bypassing obstacles based on artificial intelligence are implemented.

[0057] The beneficial effects of the present invention compared with the prior art are as follows: S1 Real-time collection of environmental information to construct a virtual environment model provides an accurate environmental basis for subsequent path planning, improving the authenticity and feasibility of path planning. S2 Integrating multiple key information to construct a path planning information database ensures the comprehensiveness and accuracy of the data required for planning. S3 Using the ant colony algorithm for initial path planning can quickly find a possible path, improving the planning efficiency. S4 Conducting multi-directional evaluation of the initial path comprehensively considers the advantages and disadvantages of the path, ensuring the quality of path planning. S5 Using the particle swarm algorithm to optimize the path when the evaluation does not meet the preset standard further improves the rationality and adaptability of the path. This solution can plan a better and more compliant path for the target mobile device through comprehensive data collection and processing, effective algorithm application, and strict evaluation and optimization.

[0058] Other features and advantages of the present invention will be described in the following specification, and, in part, will be obvious from the specification, or will be understood by implementing the present invention. The objectives and other advantages of the present invention can be achieved and obtained through the structures specifically pointed out in this application document.

[0059] The technical solutions of the present invention will be further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] The accompanying drawings are used to provide a further understanding of the present invention, and constitute a part of the specification, and are used together with the embodiments of the present invention to explain the present invention, and do not constitute a limitation to the present invention. In the accompanying drawings:

[0061] Figure 1 Flowchart of the method for autonomous obstacle avoidance and detouring around obstacles based on artificial intelligence in the embodiments of the present invention;

[0062] Figure 2 Schematic diagram of the device for autonomous obstacle avoidance and detouring around obstacles based on artificial intelligence in the embodiments of the present invention. Specific embodiments

[0063] The following describes the preferred embodiments of the present invention with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are only for the purpose of illustrating and explaining the present invention, and are not intended to limit the present invention.

[0064] Embodiment 1:

[0065] Referring to Figure 1 , the present invention provides a method for autonomous obstacle avoidance and detouring around obstacles based on artificial intelligence, including:

[0066] S1: Collect the surrounding environment information of the target mobile device in real time and construct a virtual environment model;

[0067] S2: Integrate the starting point, target position information, three-dimensional environment map information, and obstacle information in the virtual environment model of the target mobile device to construct a path planning information database;

[0068] S3: Use the ant colony algorithm and the path planning information database to perform initial path planning to obtain an initial path from the starting point of the target mobile device to the target position;

[0069] S4: Perform multi-directional evaluation on the initial path based on the virtual environment model to obtain the multi-directional evaluation result of the initial path;

[0070] S5: When the multi-directional evaluation result of the initial path does not meet the preset standard, the particle swarm optimization algorithm is used to optimize the initial path to obtain a better path.

[0071] In this embodiment, the target mobile device refers to a specific mobile device that needs to perform autonomous obstacle avoidance and detour around obstacles, such as an autonomous driving vehicle, a logistics handling robot, etc.

[0072] In this embodiment, the surrounding environment information of the target mobile device includes data such as the distribution of obstacles around the device, terrain conditions, the position and movement state of other moving objects, etc. For example, information such as other vehicles, pedestrians, and road conditions around an autonomous driving vehicle.

[0073] In this embodiment, the virtual environment model is a digital simulation of the real environment where the target mobile device is located, including various elements and features in the environment. Just like a virtual scene constructed for an autonomous driving vehicle that includes elements such as roads, buildings, and traffic lights.

[0074] In this embodiment, the starting point of the target mobile device and the starting point of the target location information. The starting point is the position where the device starts to move, and the target location is the final destination that the device needs to reach. For example, a logistics robot starts from point A in a warehouse and needs to reach point B for goods delivery.

[0075] In this embodiment, the three-dimensional environmental map information and obstacle information in the virtual environment model. The three-dimensional environmental map information refers to the three-dimensional spatial layout of the environment, and the obstacle information includes the position, shape, size, etc. of the obstacles. For example, in a virtual environment, an obstacle may be a cuboid box, and its position coordinates, size, etc.

[0076] In this embodiment, the path planning information database integrates various data sets related to the target mobile device for path planning, such as starting point, target location, environmental map, obstacle, etc.

[0077] In this embodiment, the preset criteria are pre-set criteria for determining whether the initial path needs to be optimized, such as the path length exceeding a certain value, the distance from an obstacle being less than the safety threshold, etc.

[0078] In this embodiment, the particle swarm optimization algorithm is used to optimize the initial path to obtain a better path: that is, the particle swarm optimization algorithm is used to improve and adjust the initially planned path to obtain a path that better meets the requirements (such as shorter, safer, lower energy consumption).

[0079] Path planning provides an accurate environmental basis, improving the authenticity and feasibility of path planning. S2 integrates various key information to build a path planning information database to ensure the comprehensiveness and accuracy of the data required for planning. S3 uses the ant colony algorithm for initial path planning, which can quickly find a possible path and improve the planning efficiency. S4 conducts a multi-faceted evaluation of the initial path, comprehensively considering the advantages and disadvantages of the path, and ensuring the quality of path planning. S5 uses the particle swarm optimization algorithm to optimize the path when the evaluation does not meet the preset criteria, further improving the rationality and adaptability of the path. Through comprehensive data collection and processing, effective algorithm application, and strict evaluation and optimization, this solution can plan a better-quality and more compliant path for the target mobile device.

[0080] Embodiment 2:

[0081] Based on Embodiment 1, a method for autonomous obstacle avoidance and detouring around obstacles based on artificial intelligence. S1: Real-time collect the surrounding environment information of the target mobile device and build a virtual environment model, including:

[0082] Use the multi-sensor system mounted on the target mobile device to collect the surrounding environment information of the target mobile device, where the multi-sensor system includes lidar, camera, and ultrasonic sensor;

[0083] Construct a virtual environment model based on the surrounding environment information of the target mobile device.

[0084] The beneficial effects of the above technical solutions are as follows: The multi-sensor system is used to collect the surrounding environment information. Among them, the lidar can provide accurate distance information, the camera can obtain rich image information, and the ultrasonic sensor can play a role in short-distance detection. The combination of multiple sensors ensures the comprehensiveness and accuracy of the environmental information collection. Based on the comprehensive and accurate environmental information, a virtual environment model is constructed, providing a reliable basis for subsequent path planning and obstacle avoidance operations, and improving the authenticity and effectiveness of the model. This part of the content can provide high-quality environmental data support for the autonomous obstacle avoidance and path planning of the target mobile device, improving the performance and reliability of the system.

[0085] Embodiment 3:

[0086] Based on the method of autonomous obstacle avoidance and detouring around obstacles based on artificial intelligence in Embodiment 2, a virtual environment model is constructed based on the surrounding environment information of the target mobile device, including:

[0087] Construct a three-dimensional environmental map based on the surrounding environment information of the target mobile device, where the three-dimensional environmental map contains the position, shape and size of obstacles and the relative distance information of the target mobile device;

[0088] Create a fixed object model corresponding to the real environment based on the fixed object information in the surrounding environment information of the target mobile device;

[0089] Synchronously map the dynamic quantization environmental information and the three-dimensional environmental map in the surrounding environment information of the target mobile device into the fixed object model to obtain a virtual environment model.

[0090] In this embodiment, the three-dimensional environmental map refers to a map containing three-dimensional information such as the spatial position, shape, and size of the surrounding environment of the target mobile device. For example, a map constructed for an autonomous driving vehicle that shows the undulations of the road and the three-dimensional shape of buildings.

[0091] In this embodiment, the relative distance information of the target mobile device refers to the distance data between the target mobile device and other objects or reference points in the surrounding environment. For example, the distance value between a robot and surrounding obstacles.

[0092] In this embodiment, the fixed object information in the surrounding environment information refers to the relevant information of objects with relatively fixed positions in the surrounding environment, such as buildings, utility poles, etc.

[0093] In this embodiment, based on the information of stationary objects in the surrounding environment information of the target mobile device, a stationary object model corresponding to the real environment is created. According to the information such as the characteristics and positions of the obtained stationary objects, a model that conforms to them is established in the virtual environment. For example, a virtual building model is created according to the shape and position of the building.

[0094] In this embodiment, the dynamic quantization environment information and the three-dimensional environment map in the surrounding environment information of the target mobile device are synchronously mapped into the stationary object model to obtain a virtual environment model. The dynamically changing data (such as the positions and speeds of other moving objects) and the three-dimensional environment map Figure 1 are matched to the stationary object model, thereby forming a complete virtual environment model. For example, the real-time changing vehicle positions and the three-dimensional map of the road are added to the building model to construct a comprehensive virtual environment.

[0095] The beneficial effects of the above technical solutions are as follows: constructing a three-dimensional environment map containing detailed information about obstacles provides an accurate spatial reference for path planning, which helps to more accurately plan a path to avoid obstacles. Creating a stationary object model corresponding to the real environment enhances the realism and stability of the virtual environment. Synchronously mapping the dynamic quantization environment information and the three-dimensional environment map into the stationary object model enables the virtual environment model to reflect the changes in the environment in real time, improving the timeliness and adaptability of the model. The constructed virtual environment model is more comprehensive, realistic, and dynamic, which can provide more powerful support for autonomous obstacle avoidance and path planning, and enhances the coping ability of the target mobile device in complex environments.

[0096] Embodiment 4:

[0097] Based on the method for autonomous obstacle avoidance and bypassing obstacles using artificial intelligence in Embodiment 1, S3: Use the ant colony algorithm and the path planning information database to perform initial path planning to obtain an initial path from the starting point of the target mobile device to the target position, including:

[0098] Set the number of ants m, which determines the search scale and efficiency of the algorithm. For example, in a relatively complex environment, m = 50 can be set to increase the comprehensiveness of the search

[0099] Determine the pheromone evaporation coefficient ρ and ρ ∈ [0, 1] (such as ρ = 0.5), which controls the attenuation degree of pheromone over time. The more complex the environment, the smaller this value can be appropriately reduced to retain more historical information.

[0100] And determine the pheromone intensity Q, which affects the amount of pheromone left by ants and has an important impact on the convergence speed of the algorithm. Usually, it is set according to specific problems and experience, such as Q = 100.

[0101] Initialize the pheromone matrix τ ij, representing the pheromone concentration on the path (i, j), which can be set to a small constant initially, such as τ ij (0) = 0.1, indicating that the pheromone concentrations on all paths are the same at the beginning.

[0102] And initialize the heuristic information matrix based on the obstacle information and target location information in the path planning information database, which is calculated according to the obstacle information and target location information. For example, it can be represented by calculating the reciprocal of the Euclidean distance between two points. The closer the distance, the greater the heuristic information;

[0103] Place m ants at the current position of the mobile device, which is the starting point of the ant colony;

[0104] Each ant selects the next moving direction according to a certain probability rule, and calculates the probability that ant k at node i moves to node j based on the initialized pheromone matrix and heuristic information matrix

[0106] Based on the preset probability rule and the probability that each ant moves from the current node to each node in the next node set, determine the next node that the ant chooses to move to;

[0107] When all ants have completed one move, update the pheromone concentration on the path. The update of pheromone is divided into two parts: evaporation and enhancement;

[0108] Based on the pheromone evaporation coefficient ρ, pheromone intensity Q, and the number of ants m, continue to update the pheromone concentration and continuously determine the next node that the ant chooses to move to until the termination condition is met, and obtain the selected paths of all ants;

[0109] Select the selected path with the shortest path length among the selected paths of all ants as the initial path.

[0110] In this embodiment, the number of ants is set to refer to the number of ants participating in the path search when using the ant colony algorithm for path planning. For example, it is set to 50 ants.

[0111] In this embodiment, determining the pheromone evaporation coefficient refers to clarifying the speed ratio of pheromone reduction in the ant colony algorithm. For example, it is set to 0.5.

[0112] In this embodiment, determining the pheromone intensity is a parameter that determines the influence of pheromone on the path selection of ants.

[0113] In this embodiment, the pheromone matrix is initialized, and the heuristic information matrix is initialized based on the obstacle information and target position information in the path planning information database. First, an initial value is assigned to the pheromone matrix, and at the same time, starting values are set for the heuristic information matrix according to data such as obstacles and target positions required for path planning.

[0114] In this embodiment, the preset probability rule is a pre-set probability calculation method or condition for determining the probability that an ant selects the next node.

[0115] In this embodiment, based on the preset probability rule and the probability that each ant moves from the current node to each node in the next node set, the next node that the ant selects to move to is determined. According to the pre-set rule and the calculated probability, the node that the ant will move to next is clarified.

[0116] In this embodiment, the termination condition is the condition for the ant colony algorithm to stop searching for a path, such as reaching the maximum number of iterations (e.g., 100) or finding a path that meets specific requirements.

[0117] In this embodiment, the selected path refers to the path determined by the ant from the starting point to the target point during the search process.

[0118] The beneficial effects of the above technical solutions are as follows: Setting the number of ants, the pheromone evaporation coefficient, and the intensity provides basic parameters for the algorithm operation, ensuring the stability and effectiveness of the algorithm. Initializing the pheromone matrix and the heuristic information matrix provides initial conditions for calculating the probability of ant movement, which helps to quickly start the algorithm. By calculating the probability that an ant moves to the next node and determining the next node based on the preset probability rule, the ant search path can be effectively guided. Continuously updating the pheromone concentration enables the algorithm to continuously optimize the path selection and increase the possibility of finding the optimal path. Finally, the path with the shortest length is selected as the initial path, ensuring the rationality and efficiency of the initial path. Through reasonable parameter settings and calculation steps, this part of the content can effectively plan a relatively reasonable initial path using the ant colony algorithm, providing a good foundation for subsequent optimization.

[0119] Embodiment 5:

[0120] Based on the method for autonomous obstacle avoidance and bypassing obstacles based on artificial intelligence in Embodiment 4, the probability that ant k at node i moves to node j is calculated based on the initialized pheromone matrix and heuristic information matrix including:

[0121]

[0122] where τ ijτ(i, j, t) is the pheromone concentration on the path from node i to node j at time t. The pheromone concentration changes as ants move and over time. It reflects the amount of pheromone left by past ants on this path. The higher the pheromone concentration, the greater the likelihood that this path will be chosen by ants. α is the relative importance factor of pheromone. The larger α is, the greater the influence of pheromone concentration on the path selection of ants. Ants tend to choose paths that have been traversed by more ants before. The convergence speed of the algorithm may increase, but it may also lead to premature convergence to a local optimal solution. For example, if α = 1, η ij η(i, j) is the heuristic information from node i to node j. It is calculated based on the characteristics of the problem and target information, etc. It is usually related to the target location and obstacle information and is used to guide ants to move in the direction of the target. For example, it can be represented by calculating the reciprocal of the Euclidean distance between two points. The closer the distance, the greater the heuristic information, and the relatively greater the likelihood that ants will choose this path. β is the relative importance factor of heuristic information and is used to measure the importance of heuristic information when ants select paths. The larger the value of β, the greater the influence of heuristic information on the path selection of ants. Ants tend to choose paths that are closer to the target or are better according to heuristic rules, which helps the algorithm find the target faster, but may weaken the algorithm's ability to explore new paths. Generally, it can be adjusted through experiments. For example, if β = 2, s is an index variable used for summation and represents the set of next nodes allowed k for any node in, that is, it does not include the nodes that have been visited and the nodes where obstacles are located. That is, ants can only choose the next moving direction from the nodes in this set.

[0123] In this embodiment, the pheromone concentration refers to the amount of pheromone on the path in the ant colony algorithm, which reflects the likelihood of this path being chosen by ants. For example, a high pheromone concentration on a certain path means that ants are more likely to choose this path.

[0124] In this embodiment, the relative importance factor of pheromone is a parameter used to measure the importance of pheromone when ants select paths. For example, a larger relative importance factor of pheromone indicates that pheromone has a greater influence on the path selection of ants.

[0125] In this embodiment, the relative importance factor of heuristic information is a parameter used to represent the importance of heuristic information in determining the path selection of ants. Heuristic information is usually related to distance, direction, etc. The relative importance factor of heuristic information determines the magnitude of its influence on ant decision-making.

[0126] The beneficial effects of the above technical solutions are as follows: By comprehensively considering the pheromone concentration and heuristic information through a clear calculation formula, the probability calculation for ants to move to the next node is made more scientific and accurate. The introduction of the relative importance factor of pheromone and the relative importance factor of heuristic information can flexibly adjust the weights of pheromone and heuristic information in probability calculation to adapt to different environments and requirements. This calculation formula provides a quantitative basis for the path selection of ants in the ant colony algorithm, helping to improve the search efficiency of the algorithm and the quality of path planning. It makes the ant colony algorithm more reasonable and effective when calculating the movement probability of ants, so as to better plan the initial path for the target mobile device.

[0127] Embodiment 6:

[0128] Based on Embodiment 4, for the method of autonomous obstacle avoidance and detouring around obstacles based on artificial intelligence, continue to update the pheromone concentration based on the pheromone evaporation coefficient ρ, pheromone intensity Q, and the number of ants m, and continuously determine the next node that the ant chooses to move to, including:

[0129] Volatile part: For each path (i,j), the pheromone concentration evaporates according to the following formula:

[0130] τ ij (t + 1) = (1 - ρ)τ ij (t)

[0131] Enhancement part: For each path (i,j), the pheromone concentration is enhanced according to the following formula:

[0132]

[0133] Among them, τ ij (t + 1) is the pheromone concentration on the path from node i to node j at time t + 1, and the amount of pheromone increased for the path (i,j) passed by ant k is:

[0134]

[0135] Among them, L k is the path length around ant k.

[0136] The beneficial effects of the above technical solutions are as follows: The formula for the evaporation part can reasonably reduce the pheromone concentration on the path, avoid the algorithm converging prematurely to the local optimal solution, and maintain the diversity of the search. The formula for the enhancement part increases the pheromone amount according to the path length traveled by the ants, so that better paths are more strengthened, guiding subsequent ants to be more inclined to choose these paths. By integrating the evaporation and enhancement processes, the pheromone concentration can be dynamically adjusted during the search process, balancing the relationship between exploring new paths and exploiting existing better paths, and increasing the possibility of finding the global optimal solution. Through a scientific pheromone concentration update method, the ant colony algorithm is made more efficient and accurate in path planning, and a better path is planned for the target mobile device.

[0137] Example 7:

[0138] Based on the method for autonomous obstacle avoidance and detouring around obstacles based on artificial intelligence in Example 1, S4: Perform multi-directional evaluation on the initial path based on the virtual environment model to obtain the multi-directional evaluation results of the initial path, including:

[0139] Determine the minimum distance between each path segment in the initial path and the obstacle based on the virtual environment model and the initial path, and calculate the safety evaluation index of the initial path:

[0140]

[0141] In the formula, S is the safety evaluation index of the initial path, q is the total number of path segments included in the initial path, e is the natural constant with a value of 2.71828, and d min,p is the minimum distance between the p-th path segment in the initial path and the obstacle (assuming the obstacle is a complex geometric shape, and d is determined by calculating the distance between the path line segment and the surface points of the obstacle min,p ), and d safe is the preset safety distance;

[0142] When d min,p is much greater than d safe , it indicates that the path is very safe; when d min,p is less than d safe , S will decrease significantly, indicating that there is a safety risk on the path. This formula utilizes the characteristics of the exponential function and can more delicately reflect the relationship between the path and the safety distance, and can better reflect the complexity of the safety evaluation than simple distance comparison;

[0143] Calculate the energy consumption when the target mobile device moves along the initial path (energy consumption evaluation needs to consider the motion characteristics of the mobile device and factors such as the slope and tortuosity of the path. E is a comprehensive index for evaluating path energy consumption and is used to measure the total energy consumed by the mobile device when operating on this path):

[0144]

[0145] Where E is the energy consumption of the target mobile device when it moves along the initial path, E 0 It is the energy consumed per unit distance when the target mobile device moves horizontally. It is a basic energy consumption parameter, representing the energy consumption per unit distance when moving on an ideal horizontal path without slope and twists and turns. It is related to factors such as the characteristics of the mobile device itself and the resistance during horizontal movement. b is the number of nodes included in the initial path, which determines the discretization degree of the path and the scope of calculation. The more nodes on the path, the larger the value of, the more detailed the description of the path, and the more accurate the corresponding energy consumption calculation. a+1 is the horizontal coordinate value of the a+1th node in the initial path on the two-dimensional plane, x a is the horizontal coordinate value of the ath node in the initial path on the two-dimensional plane, y a+1 is the ordinate value of the a+1th node in the initial path on the two-dimensional plane, y a is the ordinate value of the ath node in the initial path on the two-dimensional plane. In the energy consumption calculation formula, only the horizontal direction is considered. The coordinate is used to calculate the horizontal moving distance, cosθ a is the cosine value of the slope angle of the initial path, E 1 The energy consumed per unit distance when the target mobile device moves in the vertical direction reflects the energy consumption per unit distance of ascent or descent when the mobile device overcomes gravity or moves on a path with height changes. It mainly depends on factors such as the weight of the mobile device and the efficiency of vertical movement. a is the slope angle of the initial path (i.e. the angle between the line segment and the horizontal direction, which reflects the inclination of the path at this position, θ a The larger the value, the steeper the slope of the path in this section, and the higher the energy consumption of the mobile device in the vertical direction in this section. 2 is the energy consumption coefficient related to the path tortuosity, which is used to measure the impact of the path tortuosity on energy consumption. It comprehensively considers the additional energy consumed by the mobile device when turning and changing direction, and is related to the steering flexibility of the mobile device, the energy loss mechanism during turning, etc. a is the angle between the two path segments connected to the ath node in the initial path, which is used to measure the tortuosity of the path. a The larger it is, the more drastic the path turns at that location, the higher the tortuosity of the path, and the more additional energy the mobile device consumes due to turning at that location;

[0146] This formula comprehensively considers the energy consumption of the path in the horizontal and vertical directions as well as the additional energy consumption caused by the tortuosity, and fully reflects the energy consumption of the mobile device moving along the path;

[0147] Take the path length of the initial path, the safety evaluation index, and the energy consumption when the target mobile device moves along the initial path as the multi-faceted evaluation results of the initial path.

[0148] In this embodiment, the minimum distance between each path segment in the initial path and the obstacle refers to the closest distance between each segment on the initially planned path and the surrounding obstacles. For example, the closest distance between a certain path segment and the obstacle is 2 meters.

[0149] In this embodiment, the preset safety distance is the minimum distance value set in advance that is considered to keep the path safe from obstacles. For example, the preset safety distance is 1.5 meters.

[0150] In this embodiment, the energy consumption per unit distance when the target mobile device moves in the horizontal (or vertical) direction refers to the amount of energy required for the device to move one unit length in the horizontal (or vertical) direction. For example, the device consumes 10 joules of energy when moving 1 meter in the horizontal direction.

[0151] In this embodiment, the slope angle of the initial path is the angle between the straight line where the initially planned path is located and the horizontal direction. For example, the slope angle of a certain initial path segment is 30 degrees.

[0152] In this embodiment, the energy consumption coefficient related to the path tortuosity is a parameter that reflects the magnitude of the impact of the path bending degree on the energy consumption. The more tortuous the path, the greater the energy consumption coefficient.

[0153] In this embodiment, the angle between two path segments connected to a single node in the initial path refers to the angle formed by two adjacent path segments with a certain node as the vertex in the initial path. For example, the angle between two path segments at a certain node is 45 degrees.

[0154] The beneficial effects of the above technical solutions are as follows: By calculating the minimum distance between each path segment in the initial path and the obstacle to obtain the safety evaluation index, it is possible to quantitatively evaluate the safety of the path and ensure the safety of the target mobile device during operation. Calculating the energy consumption when the target mobile device moves along the initial path takes into account factors such as horizontal and vertical movement and path tortuosity, which helps to optimize energy consumption and improve the battery life of the device. Taking the path length, safety evaluation index, and energy consumption as multi-faceted evaluation results comprehensively and comprehensively considers the performance of the initial path, providing a basis from multiple angles for subsequent optimization. It is possible to conduct a comprehensive and accurate evaluation of the initial path, thus providing strong support for selecting a better path and enhancing the overall effect of the target mobile device's autonomous obstacle avoidance and path planning.

[0155] Embodiment 8:

[0156] Based on the method of the autonomous obstacle avoidance and detouring around obstacles in Embodiment 1, it further includes:

[0157] During the movement of the target mobile device along the initial path or the optimal path, continuously and real-time collect the surrounding environment of the target mobile device to obtain the real-time surrounding environment information of the target mobile device;

[0158] When it is determined based on the real-time surrounding environment information of the target mobile device that the obstacle information in the current virtual environment model has changed, the new obstacle information is fed back into the virtual environment model to obtain the latest virtual environment model;

[0159] Based on the latest virtual environment model, perform real-time adjustment on the initial path or the optimal path executed by the target mobile device to obtain the real-time adjusted path;

[0160] Send the real-time adjusted path to the motion control system of the target mobile device to control the target mobile device to move along the real-time adjusted path. At the same time, feed back the real-time motion state of the target mobile device and the latest virtual environment model to the path planning information database.

[0161] In this embodiment, the real-time surrounding environment information of the target mobile device refers to various data of its surrounding environment collected at the current moment during the operation of the target mobile device, including but not limited to the position change of obstacles, the dynamics of other moving objects, etc. For example, when an autonomous vehicle is driving, it detects a newly emerged pedestrian in front in real time.

[0162] In this embodiment, determining that the obstacle information in the current virtual environment model has changed based on the real-time surrounding environment information of the target mobile device depends on the environment information obtained by the target mobile device in real time to judge whether the description of obstacles in the originally constructed virtual environment model is different. For example, there was no obstacle at a certain position in the original model, but the real-time information shows that there is an obstacle at this position.

[0163] In this embodiment, performing real-time adjustment on the initial path or the optimal path executed by the target mobile device based on the latest virtual environment model to obtain the real-time adjusted path means modifying the previously planned initial path or the optimized optimal path according to the latest virtual environment model that can accurately reflect the current real environment, so as to obtain a path suitable for the current environment. For example, due to the emergence of a new obstacle, the original path needs to be re-planned to bypass the new obstacle.

[0164] In this embodiment, the real-time motion state of the target mobile device refers to the specific operating conditions of the target mobile device at the current moment, such as speed, position, direction, etc. For example, the current speed of the robot is 2 meters per second, the position is at coordinates (10, 20), and the orientation is eastward.

[0165] The beneficial effects of the above technical solutions are as follows: continuously and real-time collect environmental information during movement, which can timely detect environmental changes, improve the response timeliness and adaptability of the system. Feed the new obstacle information back into the virtual environment model to obtain the latest model, ensuring the real-time and accuracy of the environmental model. Based on the latest model, adjust the path in real time to ensure that the target mobile device can timely avoid newly emerged obstacles, improving the safety and efficiency of movement. Send the real-time adjusted path to the motion control system, and feedback the real-time motion state and the latest model to the database, forming a closed-loop control and optimization system, continuously improving the path planning and obstacle avoidance effects. It can enable the target mobile device to move more flexibly, safely and efficiently in a complex and changeable environment, improving the practicability and reliability of autonomous obstacle avoidance and path planning.

[0166] Embodiment 9:

[0167] Reference Figure 2 , the present invention provides a device for autonomous obstacle avoidance and bypassing obstacles based on artificial intelligence, which is used to execute the method for autonomous obstacle avoidance and bypassing obstacles based on artificial intelligence according to any one of Embodiments 1 to 8, including:

[0168] An environment acquisition module, which is used to collect the surrounding environmental information of the target mobile device in real time and construct a virtual environment model;

[0169] An information integration module, which is used to integrate the starting point, target position information, three-dimensional environmental map information and obstacle information in the virtual environment model of the target mobile device to construct a path planning information database;

[0170] An initial planning module, which is used to perform initial path planning by using the ant colony algorithm and the path planning information database to obtain an initial path from the starting point of the target mobile device to the target position;

[0171] A path evaluation module, which is used to perform multi-faceted evaluation on the initial path based on the virtual environment model to obtain a multi-faceted evaluation result of the initial path;

[0172] A path optimization module, which is used to optimize the initial path by using the particle swarm algorithm when the multi-faceted evaluation result of the initial path does not meet the preset standard to obtain a better path.

[0173] The beneficial effects of the above technical solutions are as follows: The environmental acquisition module can accurately obtain the surrounding environmental information, providing a basis for subsequent path planning and improving the accuracy and reliability of path planning. The information integration module constructs a database by integrating key information to ensure the comprehensiveness and effectiveness of the data required for path planning. The initial planning module uses the ant colony algorithm for initial path planning, which can quickly generate a preliminary feasible path and improve the planning efficiency. The path evaluation module evaluates the initial path from multiple aspects, helping to discover potential problems and deficiencies in the path and ensuring the path quality. The path optimization module uses the particle swarm algorithm to optimize the path when necessary, further enhancing the performance and adaptability of the path. Through the collaborative work of each module, the device can provide an efficient and high-quality autonomous obstacle avoidance and detour path planning service for the target mobile device.

[0174] Embodiment 10:

[0175] The present invention provides a readable storage medium, on which a program or instruction is stored. When the program or instruction is executed by a processor, the steps of the method for autonomous obstacle avoidance and detour based on artificial intelligence according to any one of Embodiments 1 to 8 are implemented.

[0176] The beneficial effects of the above technical solutions are as follows: It provides a convenient storage and dissemination method, enabling the method for autonomous obstacle avoidance and detour based on artificial intelligence to be more widely applied and shared. By storing relevant programs or instructions on the readable storage medium, it is convenient for different devices and systems to read and execute, increasing the applicability and generality of the method. It helps to ensure the consistency and stability of method execution, reducing errors caused by device differences or human operations. It provides convenience for subsequent updates and optimizations, and the improvement of the method can be achieved simply by modifying the programs or instructions in the storage medium. It promotes the popularization and development of autonomous obstacle avoidance and detour technologies, bringing convenience and progress to related application fields.

[0177] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention also intends to include these changes and modifications.

Claims

1. A method for autonomous obstacle avoidance and detouring based on artificial intelligence, characterized in that: include: S1: Collect the surrounding environment information of the target mobile device in real time and build a virtual environment model; S2: Integrate the starting point of the target mobile device, the target location information, the three-dimensional environment map information in the virtual environment model, and the obstacle information to construct a path planning information database; S3: Using the ant colony algorithm and the path planning information database to perform initial path planning, and obtain the initial path from the starting point of the target mobile device to the target location; S4: Perform a multi-faceted evaluation on the initial path based on the virtual environment model to obtain a multi-faceted evaluation result of the initial path; S5: When the multi-dimensional evaluation results of the initial path do not meet the preset standards, the particle swarm algorithm is used to optimize the initial path to obtain a better path.

2. The method for autonomous obstacle avoidance and detouring based on artificial intelligence according to claim 1, characterized in that: S1: Real-time collection of the surrounding environment information of the target mobile device and construction of a virtual environment model, including: Using a multi-sensor system mounted on the target mobile device to collect the surrounding environment information of the target mobile device, wherein the multi-sensor system includes a laser radar, a camera, and an ultrasonic sensor; A virtual environment model is constructed based on the surrounding environment information of the target mobile device.

3. The method for autonomous obstacle avoidance and detouring based on artificial intelligence according to claim 2, characterized in that: A virtual environment model is constructed based on the surrounding environment information of the target mobile device, including: Constructing a three-dimensional environment map based on the surrounding environment information of the target mobile device, wherein the three-dimensional environment map includes the location, shape and size of obstacles and the relative distance information of the target mobile device; Creating a fixed object model corresponding to the real environment based on the fixed object information in the surrounding environment information of the target mobile device; The dynamic quantitative environment information and the three-dimensional environment map in the surrounding environment information of the target mobile device are synchronously mapped to the fixed object model to obtain a virtual environment model.

4. The method for autonomous obstacle avoidance and detouring based on artificial intelligence according to claim 1, characterized in that: S3: Use the ant colony algorithm and the path planning information database to perform initial path planning to obtain an initial path from the starting point of the target mobile device to the target location, including: Set the number of ants m, determine the pheromone volatility coefficient ρ and ρ∈[0,1], and determine the pheromone intensity Q; Initialize the pheromone matrix, and initialize the heuristic information matrix based on the obstacle information and target location information in the path planning information database; Based on the initialized pheromone matrix and the heuristic information matrix, the probability that ant k at node i chooses to move to node j is calculated. Determine the next node that the ant chooses to move to based on a preset probability rule and the probability of each ant choosing to move from the current node to each node in the next node set; Based on the pheromone volatility coefficient ρ, pheromone intensity Q and the number of ants m, the pheromone concentration is continuously updated and the next node to which the ants choose to move is continuously determined until the termination condition is met, and the selection path of all ants is obtained; The path with the shortest path length is selected from all the paths selected by the ants as the initial path.

5. The method for autonomous obstacle avoidance and detouring based on artificial intelligence according to claim 4, characterized in that: Based on the initialized pheromone matrix and the heuristic information matrix, the probability that ant k at node i chooses to move to node j is calculated. include: Among them, τ ij (t) is the pheromone concentration on the path from node i to node j at time t, α is the relative importance factor of pheromone, η ij is the heuristic information from node i to node j, β is the relative importance factor of the heuristic information, s is an index variable used for summation and represents the next node set allowd that ant k can choose k Any node in .

6. The method for autonomous obstacle avoidance and detouring based on artificial intelligence according to claim 4, characterized in that: Based on the pheromone volatility coefficient ρ, the pheromone intensity Q, and the number of ants m, the pheromone concentration is continuously updated and the next node to which the ants choose to move is continuously determined, including: Volatile part: For each path (i, j), the pheromone concentration evaporates according to the following formula: t ij (t+1)=(1-ρ)τ ij (t) Enhancement part: For each path (i, j), the pheromone concentration is enhanced according to the following formula: Among them, τ ij (t+1) is the pheromone concentration on the path from node i to node j at time t+1. The amount of pheromone added to the path (i, j) taken by ant k is for: Among them, L k is the length of the path of the ant k.

7. The method for autonomous obstacle avoidance and detouring based on artificial intelligence according to claim 1, characterized in that: S4: Perform a multi-faceted evaluation of the initial path based on the virtual environment model to obtain a multi-faceted evaluation result of the initial path, including: Based on the virtual environment model and the initial path, the minimum distance between each path segment in the initial path and the obstacle is determined, and the safety assessment index of the initial path is calculated: Where S is the safety assessment index of the initial path, q is the total number of path segments contained in the initial path, e is a natural constant with a value of 2.71828, and d min,p is the minimum distance between the pth path segment in the initial path and the obstacle, d safe To preset a safe distance; Calculate the energy consumption of the target mobile device when it moves along the initial path: Where E is the energy consumption of the target mobile device when it moves along the initial path, E0 is the energy consumed per unit distance when the target mobile device moves horizontally, b is the number of nodes included in the initial path, and x a+1 is the horizontal coordinate value of the a+1th node in the initial path on the two-dimensional plane, x a is the horizontal coordinate value of the ath node in the initial path on the two-dimensional plane, y a+1 is the ordinate value of the a+1th node in the initial path on the two-dimensional plane, y a is the ordinate value of the ath node in the initial path on the two-dimensional plane, cosθ a is the cosine value of the slope angle of the initial path, E1 is the energy consumed per unit distance when the target mobile device moves in the vertical direction, sinθ a is the sine value of the slope angle of the initial path, E2 is the energy consumption coefficient related to the path tortuosity, γ a is the angle between the two path segments connected to the a-th node in the initial path; The path length of the initial path, the safety evaluation index, and the energy consumption of the target mobile device when moving along the initial path are taken as the multi-faceted evaluation results of the initial path.

8. The method for autonomous obstacle avoidance and detouring based on artificial intelligence according to claim 1, characterized in that: Also includes: When the target mobile device moves along the initial path or the better path, the surrounding environment of the target mobile device is continuously collected in real time to obtain real-time surrounding environment information of the target mobile device; When it is determined that the obstacle information in the current virtual environment model has changed based on the real-time surrounding environment information of the target mobile device, the new obstacle information is fed back to the virtual environment model to obtain the latest virtual environment model; Based on the latest virtual environment model, an initial path or a better path executed by the target mobile device is adjusted in real time to obtain a real-time adjusted path; The real-time adjusted path is sent to the motion control system of the target mobile device to control the target mobile device to move according to the real-time adjusted path. At the same time, the real-time motion state of the target mobile device and the latest virtual environment model are fed back to the path planning information database.

9. An autonomous obstacle avoidance device based on artificial intelligence, characterized in that: A method for executing the autonomous obstacle avoidance and detouring method based on artificial intelligence as claimed in any one of claims 1 to 8, comprising: The environment acquisition module is used to collect the surrounding environment information of the target mobile device in real time and build a virtual environment model; An information integration module is used to integrate the starting point of the target mobile device, the target location information, the three-dimensional environment map information in the virtual environment model, and the obstacle information to construct a path planning information database; The initial planning module is used to perform initial path planning using an ant colony algorithm and a path planning information database to obtain an initial path from a starting point of the target mobile device to a target location; A path evaluation module is used to perform a multi-faceted evaluation on the initial path based on the virtual environment model to obtain a multi-faceted evaluation result of the initial path; The path optimization module is used to optimize the initial path using a particle swarm algorithm to obtain a better path when the multi-dimensional evaluation results of the initial path do not meet the preset standards.

10. A readable storage medium, characterized in that: The readable storage medium stores programs or instructions, and when the programs or instructions are executed by the processor, the steps of the method for autonomous obstacle avoidance and circumventing obstacles based on artificial intelligence as claimed in any one of claims 1 to 8 are implemented.

Images