Multi-unmanned aerial vehicle navigation method, device and equipment based on artificial intelligence and medium

Abstract

The method is mainly applied to the technical field of intelligent robots. The invention discloses a multi-unmanned aerial vehicle navigation method, device and equipment based on artificial intelligence and a medium, and the method comprises the steps: obtaining the current environment information of the position of each unmanned aerial vehicle, the current environment information comprising the position information of a target object, the depth image information and the current position information of each unmanned aerial vehicle; on the basis of the position information of the target object and the current position information of each unmanned aerial vehicle, the moving route of each unmanned aerial vehicle is updated; and when the unmanned aerial vehicle moves to the target object according to the updated moving route, generating an obstacle avoidance strategy corresponding to the unmanned aerial vehicle by using a preset obstacle avoidance strategy model based on the depth image information, so that the unmanned aerial vehicle executes the obstacle avoidance strategy and avoids the obstacle. The navigation capability of the unmanned aerial vehicle can be improved, and the obstacle can be efficiently avoided in the navigation process.

Description

Technical Field

[0001] The present invention relates to the field of intelligent robot technology, and in particular to an artificial intelligence-based multi-unmanned aerial vehicle navigation method, device, equipment and medium. Background Art

[0002] With the rapid development of unmanned aerial vehicle (UAV) technology, UAVs have been applied to many fields, including agriculture, environmental monitoring, search and rescue, mining, and patrol inspection, due to their high flexibility, wide applicability, and economic controllability. In application scenarios, effective collaboration between multiple UAVs, especially when large-scale UAV groups perform tasks, is crucial to find an optimal path to avoid obstacles and reach the target location. However, traditional UAV obstacle avoidance methods rely on simultaneous positioning and mapping technology, which usually requires a lot of computing resources and is limited by prior map information, and is not adaptable to unknown environments. In order to adapt to unknown environments and reduce dependence on computing resources, new UAV navigation and obstacle avoidance strategies need to be developed. Summary of the invention

[0003] The present invention provides a multi-UAV navigation method, device, equipment and medium based on artificial intelligence, which can improve the navigation capability of UAVs and efficiently avoid obstacles during navigation.

[0004] The present invention provides a multi-UAV navigation method based on artificial intelligence, the method comprising: Acquire current environment information of each drone's location, the current environment information including location information of the target object, depth image information, and current location information of each drone; Based on the location information of the target object and the current location information of each of the drones, respectively updating the moving route of each of the drones; When the drone moves toward the target object according to the updated moving route, an obstacle avoidance strategy corresponding to the drone is generated based on the depth image information using a preset obstacle avoidance strategy model, so that the drone executes the obstacle avoidance strategy and avoids obstacles.

[0005] Furthermore, the multi-UAV navigation method based on artificial intelligence also includes: Based on the position information of the target object and the current position information of the drone, an artificial potential field is generated between the drone and the target object; Through the gravitational effect of the artificial potential field, the drone is controlled to move in the direction of the location of the target object.

[0006] Further, the obstacle avoidance strategy includes a second speed used by the drone to avoid obstacles; The multi-UAV navigation method based on artificial intelligence also includes: When the drone moves, determining a first speed of the drone based on the gravitational effect of the artificial potential field; Based on the first speed and the second speed, the current speed of the drone is updated.

[0007] Further, the obstacle avoidance strategy model includes a first model network and a second model network; When the drone moves toward the target object according to the updated moving route, based on the depth image information, a preset obstacle avoidance strategy model is used to generate an obstacle avoidance strategy corresponding to the drone, so that the drone executes the obstacle avoidance strategy and avoids obstacles, including: Based on the current environment information, generating an observation vector; After the observation vector is input into the first model network, an action vector is generated through the first model network, and the obstacle avoidance strategy is generated according to the action vector; After inputting the action vector into the second model network, determining a score value corresponding to the action vector through the second model network; Based on the score value, an optimized action vector is generated through the first model network, and the obstacle avoidance strategy is updated according to the optimized action vector.

[0008] Further, generating an observation vector based on the current environment information includes: Acquire a depth image through the drone; Generate a spatial vector corresponding to the depth image using a preset encoder; Obtaining a direction vector pointing to the target object and a current speed of the drone; The observation vector is generated based on the space vector, the direction vector and the current speed.

[0009] Furthermore, the artificial intelligence-based multi-UAV navigation method further includes training the obstacle avoidance strategy model; The training of the obstacle avoidance strategy model includes: Generate multiple reward values; Constructing a reward function based on each of the reward values ​​and a preset weight coefficient corresponding to each of the reward values; When the obstacle avoidance strategy model is trained according to a preset training data set, feedback data of the obstacle avoidance strategy model training process is generated according to the reward function, and the obstacle avoidance strategy model is optimized according to the feedback data.

[0010] Further, the generating of multiple reward values ​​includes: Acquire a first distance value between the drone and the target object; If it is detected that the first distance value decreases, a positive first reward value is generated; if it is detected that the first distance value increases, a negative first reward value is generated; A second distance value between the drone and the obstacle is obtained, and when the second distance is within a preset value range, a negative second reward value is generated.

[0011] The present invention also provides a multi-UAV navigation device based on artificial intelligence, the device comprising: An acquisition module is used to acquire current environment information of each drone's location, wherein the current environment information includes location information of the target object, depth image information, and current location information of each drone; An updating module, configured to update the moving route of each of the drones based on the location information of the target object and the current location information of each of the drones; A generation module is used to generate an obstacle avoidance strategy corresponding to the drone based on the depth image information and using a preset obstacle avoidance strategy model when the drone moves toward the target object according to the updated moving route, so that the drone executes the obstacle avoidance strategy and avoids obstacles.

[0012] The present invention also provides an electronic device, comprising a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the multi-UAV navigation method based on artificial intelligence as described in any one of the above items is implemented.

[0013] The present invention also provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, it implements the artificial intelligence-based multi-UAV navigation method as described in any of the above items.

[0014] The present invention has at least the following beneficial effects: In the technical solution of the present application, first, the drone obtains the current environmental information through sensors, including the location of the target object, the depth image information and the current location of the drone itself. This real-time perception capability enables the drone to navigate and avoid obstacles directly according to the current environment without pre-building a map; then, based on the acquired location of the target object and the current location information of the drone, the movement route of each drone is updated in real time. This dynamic path planning ensures that the drone can adjust its flight trajectory according to the latest environmental information, thereby improving the flexibility and accuracy of navigation; finally, when the drone moves toward the target object according to the updated movement route, the preset obstacle avoidance strategy model is used to generate the corresponding obstacle avoidance strategy according to the depth image information. These strategies guide the drone to effectively avoid obstacles during flight, avoid the limitations of relying on prior map information, and enable the drone to flexibly adapt to environmental changes. This technical solution enables drones to achieve efficient navigation and obstacle avoidance in complex environments, and improve their adaptability and task completion rate in unknown or dynamically changing environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The accompanying drawings are used to provide a further understanding of the technical solution of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the technical solution of the present invention and do not constitute a limitation on the technical solution of the present invention.

[0016] Figure 1 is a flowchart of the steps of a multi-UAV navigation method based on artificial intelligence in this embodiment; Figure 2 is a flowchart of step S103 in a multi-UAV navigation method based on artificial intelligence in this embodiment; Figure 3 is a structural schematic diagram of an obstacle avoidance strategy generation model in an artificial intelligence-based multi-UAV navigation method of this embodiment; Figure 4 is a flowchart of step S201 in a multi-UAV navigation method based on artificial intelligence in this embodiment; Figure 5 is a flowchart of a depth image processing process in a multi-UAV navigation method based on artificial intelligence in this embodiment; Figure 6 This is a flowchart of the steps of training an obstacle avoidance strategy model in an artificial intelligence-based multi-UAV navigation method of this embodiment.

[0017] Figure 7 is a structural schematic diagram of a multi-UAV navigation device based on artificial intelligence in this embodiment; Figure 8 It is a structural diagram of an electronic device. DETAILED DESCRIPTION

[0018] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0019] In the related field, traditional multi-UAV obstacle avoidance methods mainly rely on real-time simultaneous localization and mapping (SLAM) technology. These methods usually use sensors such as laser radar (LiDAR) to perceive the surrounding environment and generate flight trajectories through path planning algorithms. In order to improve the performance of SLAM systems, prior map information is usually introduced to assist the navigation of UAVs in known environments. However, these traditional methods have some limitations. They usually require a lot of computing resources, which may be impractical in resource-constrained environments. In addition, these methods are limited by the existing prior map information and have poor adaptability to unknown or dynamically changing environments. Among them, "A UAV Navigation Approach Basedon Deep Reinforcement Learning in Large Cluttered 3D Environments" proposes a UAV navigation method based on deep reinforcement learning, which can achieve end-to-end obstacle avoidance navigation without the need for environmental maps. This method uses the sensor data of the UAV, including seven horizontal and five vertical ranging sensor data, as well as the attitude and target information of the UAV, to train the reinforcement learning strategy. In addition, a fast cyclic random value gradient algorithm is designed to solve the reinforcement learning model online to improve the obstacle avoidance ability of the UAV in complex environments. This method directly maps the raw perception information into navigation signals, avoiding the memory usage and computational consumption required for real-time map construction, thereby improving the navigation efficiency of the UAV in unknown environments.

[0020] However, the researchers of this application found that the above-mentioned prior art solutions mainly rely on the ranging sensor data in seven horizontal directions and five vertical directions carried by the drone to identify obstacles. This method performs well when dealing with obstacles of specific shapes, but its recognition ability is insufficient when faced with diverse obstacle shapes, and it is difficult to be widely applied to all types of obstacles. Since the obstacle perception of drones mainly relies on the ranging sensor data in seven horizontal directions and five vertical directions, this configuration works better when identifying obstacles of regular shapes such as rectangular blocks, but the recognition ability for irregular shapes such as triangular pyramids or slender obstacles is insufficient, which limits the obstacle avoidance ability of drones in complex environments.

[0021] When the navigation target changes, the navigation performance of the drone, including the ability to find the shortest path and the mission success rate, may drop significantly. This shows that the flexibility and robustness of the above-mentioned existing technologies in adapting to target changes need to be improved. During the training process of reinforcement learning, the navigation target of the drone is pre-set and fixed. The drone will train the optimal deep reinforcement learning strategy for this fixed target. However, once the target changes, the original strategy may not be able to adapt to the new target, resulting in a decrease in navigation performance and an increase in failure rate. This shows that the navigation strategy of the drone is not highly flexible and adaptable.

[0022] Under the above-mentioned prior art framework, drones cannot independently obtain the location information of other drones, which limits their ability to achieve distributed control. Instead, they rely on a centralized control center to obtain and coordinate the positions of all drones in real time, which may lead to inefficiency and response delays in large-scale operations. Drones mainly obtain information about obstacles in front through ranging sensors, but this method cannot obtain the location information of other drones. Therefore, a centralized controller is required to obtain the positions of all drones in real time to achieve effective collaborative control. In addition, the above-mentioned prior art does not take into account mutual obstacle avoidance between drones, and additional control instructions need to be assigned to each drone to avoid mutual interference.

[0023] When training a drone’s deep reinforcement learning policy, it takes a long time to achieve policy convergence. The time-consuming process may delay the deployment of drones in practical applications, especially when they need to quickly adapt to new environments or tasks. The reward value of the reinforcement learning model is calculated based on the path connecting the starting point and the end point and the ranging sensor data, rather than the optimal path from the drone’s current position to the end point. This results in the drone being unable to effectively find the best path based on its current position after completing obstacle avoidance, thereby prolonging the learning process. This limits the drone’s ability to quickly adapt and make decisions in dynamic environments.

[0024] In order to solve the above technical problems, the present application proposes the following embodiments, which can improve the navigation capability of the drone and avoid obstacles efficiently during navigation.

[0025] Please refer to Figure 1 , Figure 1 This is a flowchart of the steps of a multi-UAV navigation method based on artificial intelligence in this embodiment.

[0026] This embodiment provides a multi-UAV navigation method based on artificial intelligence, including: S101, obtaining current environment information of each drone's location, where the current environment information includes location information of a target object, depth image information, and current location information of each drone.

[0027] S102: Based on the location information of the target object and the current location information of each drone, the moving route of each drone is updated respectively.

[0028] S103: When the UAV moves toward the target object according to the updated moving route, an obstacle avoidance strategy corresponding to the UAV is generated based on the depth image information using a preset obstacle avoidance strategy model, so that the UAV executes the obstacle avoidance strategy and avoids obstacles.

[0029] In step S101 of some embodiments, the drone may be equipped with a variety of sensors, such as GPS, laser radar, visual sensor, ultrasonic sensor, etc. These sensors provide information such as location, obstacle detection, height, etc. Through the data fusion algorithm, the data obtained by different sensors are integrated to obtain the current environment information of the drone's location, which includes the extraction of detailed information of obstacles, such as location, size, etc.

[0030] In step S102 of some embodiments, a dynamic window method (DWA) is used for local path planning. Within a given current state and dynamic window, all possible movements are evaluated and an optimal movement strategy is selected to enable the drone to avoid obstacles and move toward the target. In addition, an improved A* algorithm or RRT algorithm can also be used for path planning. These algorithms can dynamically adjust the flight path according to real-time updated environmental information and drone status to ensure that the drone reaches the target point safely and efficiently.

[0031] In step S102 of some embodiments, in a multi-UAV collaborative task, each UAV may utilize a distributed path planning algorithm. For example, the Hybrid A* algorithm is used for distributed path planning, and each UAV independently plans a path based on its own target point and environmental information, and continuously detects obstacles and collision risks during flight. If there is a risk, the safe flight speed is calculated based on the three-dimensional speed obstacle model, and the path is adjusted.

[0032] In some embodiments, a multi-UAV navigation method based on artificial intelligence further includes: Based on the location information of the target object and the current location information of the UAV, an artificial potential field is generated between the UAV and the target object; through the gravitational effect of the artificial potential field, the UAV is controlled to move in the direction of the target object.

[0033] Furthermore, the navigation speed of the drone is calculated by combining the current position, the target position and the attraction of the artificial potential field to calculate the speed of the drone to the target velocity1. The specific calculation method is as follows:

[0034]

[0035] in, Indicates attraction, represents the attraction coefficient, represents the target point coordinates, represents the coordinates of the drone, V represents the speed, Represents the distance between two coordinate points.

[0036] It is understandable that through the gravitational effect of the artificial potential field, the UAV can quickly and accurately move toward the location of the target object, reducing the time to find the target in a complex environment and improving the overall navigation efficiency. Moreover, the basic principle of the artificial potential field method is intuitive, and the UAV movement is guided by simulating the interaction between gravity and repulsion. The algorithm structure is simple and easy to implement in the UAV system. In addition, the artificial potential field method does not need to generate a complete path in advance, but calculates the movement direction of the UAV in real time according to the current potential field state. The amount of calculation is relatively small, which can meet the needs of real-time control of the UAV and better handle dynamic obstacle environments. When the position or state of the obstacle changes, the UAV can quickly adjust the flight path according to the new potential field information, and has strong dynamic adaptability. Moreover, the path it generates is relatively smooth, avoiding frequent sharp turns or mutations, which helps the UAV maintain a stable flight attitude, reduce energy consumption, and reduce flight time.

[0037] In some embodiments, a multi-UAV navigation method based on artificial intelligence further includes: When the UAV moves, the first speed of the UAV is determined based on the gravitational effect of the artificial potential field; the current speed of the UAV is updated based on the first speed and the second speed; and the obstacle avoidance strategy includes the second speed used by the UAV to avoid obstacles.

[0038] It can be understood that this embodiment combines the speed calculated by the artificial potential field method with the speed and parameters generated by the reinforcement learning model to enable the drone to complete obstacle avoidance and navigation tasks.

[0039] A specific implementation of this embodiment is to use the speed calculated by the artificial potential field as the first speed, and the speed generated by the reinforcement learning model as the second speed, and update the current speed of the drone based on the first speed and the second speed:

[0040] in, is the speed of the drone at the next moment, are the parameters generated by the reinforcement learning model, is the velocity calculated by the artificial potential field, is the speed at which reinforcement learning models are generated.

[0041] It can be understood that this embodiment effectively combines the control navigation speed with the obstacle avoidance speed through the parameter fusion strategy to improve the obstacle avoidance and navigation capabilities of the drone in unknown environments, and has a high degree of generalization ability, can adapt to various unknown environments, and provide a flexible and effective obstacle avoidance and navigation solution for the drone. By effectively combining the control navigation speed with the obstacle avoidance speed, the drone can autonomously learn and optimize its obstacle avoidance strategy in a complex environment while maintaining navigation efficiency to the target.

[0042] In some embodiments, the obstacle avoidance strategy model includes a first model network and a second model network.

[0043] Please refer to Figure 2 , Figure 2 This is a flowchart of step S103 in the multi-UAV navigation method based on artificial intelligence in this embodiment.

[0044] In some embodiments, step S103 includes: S201. Generate an observation vector based on current environment information.

[0045] S202: After the observation vector is input into the first model network, an action vector is generated through the first model network, and an obstacle avoidance strategy is generated according to the action vector.

[0046] S203: After inputting the action vector into the second model network, the score value corresponding to the action vector is determined through the second model network.

[0047] S204: Based on the score value, generate an optimized action vector through the first model network, and update the obstacle avoidance strategy according to the optimized action vector.

[0048] In some embodiments, the obstacle avoidance strategy model is a strategy model based on the DQN algorithm, including an evaluation network (Evaluate Network) and a target network (Target Network), the first model network is a strategy network, and the second model network is a value function network. The first model network is used to generate an action vector. It outputs the Q value of each possible action by inputting an observation vector. During the training process, the parameters of the evaluation network are continuously updated through the gradient descent method to better approximate the true Q value. The second model network is used to determine the score value corresponding to the action vector. The parameters of the target network are not updated in real time, but are copied from the evaluation network regularly. The purpose of this is to stabilize the training process and reduce the problem of network training instability caused by frequent changes in the target Q value.

[0049] Specifically, an observation vector is generated based on the current environmental information, including the current position and speed of the drone, the position of the target object, and the position of obstacles.

[0050] The observation vector is input to the evaluation network (first model network). The evaluation network outputs the Q value of each possible action and selects the action with the highest Q value as the current action vector.

[0051] The action vector is input to the target network (the second model network). The target network is used to calculate the score value of the action vector, that is, the expected return value in the current state.

[0052] The parameters of the evaluation network are updated according to the score value, and a loss function (such as mean square error) is used to measure the difference between the predicted Q value and the target Q value, and then optimized by gradient descent.

[0053] When the score value meets the preset conditions, the optimized action vector is output.

[0054] The obstacle avoidance strategy is determined based on the optimized action vector. The action vector contains the optimal obstacle avoidance action of the drone in the current environment, such as adjusting the flight direction and speed. The drone executes the obstacle avoidance strategy to avoid obstacles and move towards the target object.

[0055] Please refer to Figure 3 , Figure 3 It is a structural schematic diagram of an obstacle avoidance strategy generation model in an artificial intelligence-based multi-UAV navigation method of this embodiment.

[0056] In some embodiments, the obstacle avoidance strategy model is a strategy model based on the Soft Actor-Critic algorithm, including a strategy network (Actor Network) and a value function network (Critic Networks), the first model network is a strategy network, and the second model network is a value function network. In a specific embodiment, the first model network and the second model network both adopt a three-layer structure, each layer includes a 1024-dimensional hidden layer, forming a multi-layer perceptron (MLP). The output of the strategy network is the parameters of the Gaussian action distribution, for example, a 4-dimensional vector is output, corresponding to the three control speeds of the drone: and a speed control parameter These parameters include mean and variance, which define the probability distribution of action selection when the drone performs obstacle avoidance and navigation tasks. The value function network outputs the evaluation of the action to achieve policy training.

[0057] It can be understood that this embodiment generates a corresponding obstacle avoidance strategy for each drone through an obstacle avoidance strategy model, does not rely on a centralized controller to individually assign control instructions, does not rely on the global coordinates of other drones, and is processed like other obstacles in the environment without the need for additional obstacle avoidance algorithms.

[0058] Please refer to Figure 4, Figure 4 This is a flowchart of step S201 in the multi-UAV navigation method based on artificial intelligence in this embodiment.

[0059] In some embodiments, step S201 includes: S301. Acquire a depth image through a drone.

[0060] S302: Generate a spatial vector corresponding to the depth image using a preset encoder.

[0061] S303: Obtain a direction vector pointing to the target object and a current speed of the drone.

[0062] S304: Generate an observation vector based on the space vector, the direction vector and the current speed.

[0063] In this embodiment, the drone is equipped with a depth camera that can capture depth images in front of it, which provide distance information to surrounding objects. Using a regularized autoencoder (RAE), the drone can process depth images captured at three consecutive moments. The regularized autoencoder encodes these images into a 50-dimensional spatial vector , then, the 50-dimensional space vector , the direction vector pointing to the target object and current speed Integrate to form a comprehensive observation vector .

[0064] In some embodiments, this observation vector can be used as input to a Soft Actor-Critic (SAC) algorithm to train the obstacle avoidance navigation strategy of the drone.

[0065] It is understandable that when the preset encoder is used to generate the spatial vector corresponding to the depth image, the encoder will encode the image into a spatial vector and then reconstruct these representations back to the original image through the decoder. This process effectively extracts and highlights the key visual information in the environment related to the drone's mission, thereby enhancing the drone's ability to understand the surrounding environment.

[0066] Please refer to Figure 5 , Figure 5 It is a flowchart of a depth image processing process in an artificial intelligence-based multi-UAV navigation method of this embodiment.

[0067] like Figure 5 As shown in Figure 1, in an encoder-decoder structure, after the depth image is input, the encoder converts the original image into a feature representation that is easier to process, and the decoder may convert these features back into image form. In this process, the spatial vector is extracted.

[0068] It can be understood that this embodiment uses a regularized autoencoder to encode the depth image in front of the drone and extract a low-dimensional representation vector of the image. This process effectively converts complex environmental information into a concise potential representation vector, allowing the drone to automatically identify and filter out visual features that are closely related to the obstacle avoidance task. Since the environmental information is obtained in the form of images, the drone can learn the ability to avoid obstacles of any shape. By combining the autoencoder and the depth image, the present invention significantly improves the generalization ability of the drone's obstacle avoidance.

[0069] Please refer to Figure 6 , Figure 6 This is a flowchart of the steps of training an obstacle avoidance strategy model in an artificial intelligence-based multi-UAV navigation method of this embodiment.

[0070] In some embodiments, a multi-UAV navigation method based on artificial intelligence further includes the step of training an obstacle avoidance strategy model, and specific implementation methods include: S401. Generate multiple reward values.

[0071] S402: Construct a reward function based on each reward value and a preset weight coefficient corresponding to each reward value.

[0072] S403: When the obstacle avoidance strategy model is trained according to the preset training data set, feedback data of the obstacle avoidance strategy model training process is generated according to the reward function, and the obstacle avoidance strategy model is optimized according to the feedback data.

[0073] In some embodiments, step S401 includes: A first distance value between the drone and the target object is obtained; if it is detected that the first distance value decreases, a positive first reward value is generated; if it is detected that the first distance value increases, a negative first reward value is generated.

[0074] In some embodiments, step S401 includes: A second distance value between the drone and the obstacle is obtained, and when the second distance value is within a preset value range, a negative second reward value is generated.

[0075] In some embodiments, step S401 includes: When it is detected that the current state of the drone is a stopped state, a negative third reward value is generated.

[0076] In a specific embodiment, the obstacle avoidance strategy model is a strategy model based on the Soft Actor-Critic algorithm. In the process of training the obstacle avoidance strategy model, the reward calculation of the drone involves three key aspects: the distance to the target, the distance to the obstacle, and its own motion state. When the drone approaches the target point, it will receive a positive reward, otherwise it will receive a negative reward. When the drone is within a certain range of the obstacle, it will receive a negative reward. When the drone stops moving, it will receive a negative reward.

[0077]

[0078]

[0079]

[0080]

[0081]

[0082] in, is the reward for moving toward the goal, is the distance the UAV moves relative to the target in one time step, is the change in distance moved relative to the target, is the attraction coefficient. is the repulsive force exerted by the obstacle point on the drone, is the repulsion coefficient, is the minimum distance at which the repulsive force formula takes effect, It's a safe distance. is the distance between two coordinate points, is the direction between two coordinate points. is the reward value for approaching an obstacle. The drone stops moving and gets a reward. and is the reward value coefficient, and r is the reward value obtained by the drone after performing the action.

[0083] It is understandable that this embodiment uses a reward function to train the obstacle avoidance strategy model. This function combines the principle of artificial potential field and effectively reduces the learning time of deep reinforcement learning through linear target movement reward value and nonlinear obstacle avoidance reward value. This design not only improves the learning efficiency, but also enhances the adaptability and flexibility of drones in complex environments.

[0084] In any of the above method embodiments, first, the drone obtains current environmental information through sensors, including the location of the target object, depth image information, and the current location of the drone itself. This real-time perception capability enables the drone to navigate and avoid obstacles directly according to the current environment without pre-building a map; then, based on the acquired location of the target object and the current location information of the drone, the movement route of each drone is updated in real time. This dynamic path planning ensures that the drone can adjust its flight trajectory according to the latest environmental information, thereby improving the flexibility and accuracy of navigation; finally, when the drone moves toward the target object according to the updated movement route, the preset obstacle avoidance strategy model is used to generate a corresponding obstacle avoidance strategy according to the depth image information. These strategies guide the drone to effectively avoid obstacles during flight, avoid the limitations of relying on prior map information, and enable the drone to flexibly adapt to environmental changes. This technical solution enables drones to achieve efficient navigation and obstacle avoidance in complex environments, and improve their adaptability and task completion rate in unknown or dynamically changing environments.

[0085] Please refer to Figure 7 , Figure 7 Schematic diagram of the structure of a multi-UAV navigation device based on artificial intelligence in this embodiment.

[0086] This embodiment also provides a multi-UAV navigation device based on artificial intelligence, including: The acquisition module 501 is used to acquire the current environment information of each drone's location, where the current environment information includes the location information of the target object, the depth image information, and the current location information of each drone.

[0087] The updating module 502 is used to update the moving route of each drone based on the location information of the target object and the current location information of each drone.

[0088] The generation module 503 is used to generate an obstacle avoidance strategy corresponding to the drone based on the depth image information and using a preset obstacle avoidance strategy model when the drone moves toward the target object according to the updated moving route, so that the drone executes the obstacle avoidance strategy and avoids obstacles.

[0089] It will be appreciated by those skilled in the art that all or some of the steps and devices in the methods disclosed above may be implemented as software, firmware, hardware, and appropriate combinations thereof. Some or all physical components may be implemented as software executed by a processor, such as a central processing unit, a digital signal processor, or a microprocessor, or as hardware, or as an integrated circuit, such as an application-specific integrated circuit. Such software may be distributed on a computer-readable medium, which may include a computer storage medium (or non-transitory medium) and a communication medium (or transient medium). As known to those skilled in the art, the term computer storage medium includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules, or other data). Computer storage media include, but are not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile disk (DVD) or other optical disk storage, magnetic cassettes, magnetic tapes, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and can be accessed by a computer. As is well known to those skilled in the art, communication media typically embodies computer readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transport mechanism, and may include any information delivery media.

[0090] It can be understood that the contents of the above method embodiments are all applicable to the present device embodiments, the functions specifically implemented by the present device embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0091] An embodiment of the present application also provides an electronic device, which includes a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, any of the above artificial intelligence-based multi-UAV navigation methods is implemented.

[0092] refer to Figure 8 , Figure 8 The hardware structure of an electronic device of another embodiment is illustrated, and the electronic device includes: The processor 601 may be implemented by a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of the present application; The memory 602 can be implemented in the form of a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 602 can store operating devices and other applications. When the technical solution provided in the embodiment of this specification is implemented by software or firmware, the relevant program code is stored in the memory 602, and the processor 601 is called to execute the multi-UAV navigation method based on artificial intelligence in the embodiment of the present application; Input / output interface 603, used to implement information input and output; Communication interface 604, used to realize communication interaction between the device and other devices, which can be realized through wired mode (such as USB, network cable, etc.) or wireless mode (such as mobile network, WIFI, Bluetooth, etc.); A bus 605 that transmits information between the various components of the device (e.g., the processor 601, the memory 602, the input / output interface 603, and the communication interface 604); The processor 601 , the memory 602 , the input / output interface 603 and the communication interface 604 are connected to each other in communication within the device via a bus 605 .

[0093] It can be understood that the contents of the above method embodiments are all applicable to the electronic device embodiment, the functions specifically implemented by the electronic device embodiment are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0094] An embodiment of the present application also provides a computer-readable storage medium, which stores a program executable by a processor. When the program executable by the processor is executed by the processor, it is used to implement the artificial intelligence-based multi-UAV navigation method as described in any one of the above-mentioned specific embodiments.

[0095] An embodiment of the present application also discloses a computer program product, including a computer program or computer instructions, wherein the computer program or computer instructions are stored in a computer-readable storage medium, a processor of a computer device reads the computer program or computer instructions from the computer-readable storage medium, and the processor executes the computer program or computer instructions, so that the computer device executes the artificial intelligence-based multi-UAV navigation method as described in any of the previous embodiments.

[0096] It can be understood that the contents of the above method embodiments are all applicable to the present storage medium embodiments, the functions specifically implemented by the present storage medium embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0097] The terms "first", "second", "third", "fourth", etc. (if any) in the specification of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein, for example. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, device, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices. It should be understood that in the present application, "at least one (item)" refers to one or more, and "a plurality" refers to two or more.

[0098] In the several embodiments provided in the present application, it should be understood that the disclosed devices, apparatuses and methods can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0099] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0100] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0101] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, a server or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (Read-Only Memory, referred to as ROM), random access memory (Random Access Memory, referred to as RAM), disk or optical disk and other media that can store program codes.

[0102] Although the description of the present application has been quite detailed and specifically describes several described embodiments, it is not intended to be limited to any of these details or embodiments or any particular embodiment, but should be regarded as providing a broad possible interpretation of these claims by reference to the attached claims, taking into account the prior art, so as to effectively cover the intended scope of the present application. In addition, the above description of the present application is based on the embodiments foreseeable by the inventor, and its purpose is to provide a useful description, and those non-substantial changes to the present application that have not yet been foreseen may still represent equivalent changes to the present application.

Claims

1. A multi-UAV navigation method based on artificial intelligence, characterized in that: The method comprises: Acquire current environment information of each drone's location, the current environment information including location information of the target object, depth image information, and current location information of each drone; Based on the location information of the target object and the current location information of each of the drones, respectively updating the moving route of each of the drones; When the drone moves toward the target object according to the updated moving route, an obstacle avoidance strategy corresponding to the drone is generated based on the depth image information using a preset obstacle avoidance strategy model, so that the drone executes the obstacle avoidance strategy and avoids obstacles.

2. The multi-UAV navigation method based on artificial intelligence according to claim 1 is characterized in that: The method further comprises: Based on the position information of the target object and the current position information of the drone, an artificial potential field is generated between the drone and the target object; Through the gravitational effect of the artificial potential field, the drone is controlled to move in the direction of the location of the target object.

3. The multi-UAV navigation method based on artificial intelligence according to claim 2 is characterized in that: The obstacle avoidance strategy includes a second speed used by the drone to avoid obstacles; The method further comprises: When the drone moves, determining a first speed of the drone based on the gravitational effect of the artificial potential field; Based on the first speed and the second speed, the current speed of the drone is updated.

4. The multi-UAV navigation method based on artificial intelligence according to claim 1 is characterized in that: The obstacle avoidance strategy model includes a first model network and a second model network; When the drone moves toward the target object according to the updated moving route, based on the depth image information, a preset obstacle avoidance strategy model is used to generate an obstacle avoidance strategy corresponding to the drone, so that the drone executes the obstacle avoidance strategy and avoids obstacles, including: Based on the current environment information, generating an observation vector; After the observation vector is input into the first model network, an action vector is generated through the first model network, and the obstacle avoidance strategy is generated according to the action vector; After inputting the action vector into the second model network, determining a score value corresponding to the action vector through the second model network; Based on the score value, an optimized action vector is generated through the first model network, and the obstacle avoidance strategy is updated according to the optimized action vector.

5. The multi-UAV navigation method based on artificial intelligence according to claim 4 is characterized in that: The generating an observation vector based on the current environment information includes: Acquire a depth image through the drone; Generate a spatial vector corresponding to the depth image using a preset encoder; Obtaining a direction vector pointing to the target object and a current speed of the drone; The observation vector is generated based on the space vector, the direction vector and the current speed.

6. The multi-UAV navigation method based on artificial intelligence according to claim 1 is characterized in that: The method further comprises training the obstacle avoidance strategy model; The training of the obstacle avoidance strategy model includes: Generate multiple reward values; Constructing a reward function based on each of the reward values ​​and a preset weight coefficient corresponding to each of the reward values; When the obstacle avoidance strategy model is trained according to a preset training data set, feedback data of the obstacle avoidance strategy model training process is generated according to the reward function, and the obstacle avoidance strategy model is optimized according to the feedback data.

7. The multi-UAV navigation method based on artificial intelligence according to claim 6 is characterized in that: The generating of multiple reward values ​​comprises: Acquire a first distance value between the drone and the target object; If it is detected that the first distance value decreases, a positive first reward value is generated; if it is detected that the first distance value increases, a negative first reward value is generated; A second distance value between the drone and the obstacle is obtained, and when the second distance is within a preset value range, a negative second reward value is generated.

8. A multi-UAV navigation device based on artificial intelligence, characterized in that: The device comprises: An acquisition module is used to acquire current environment information of each drone's location, wherein the current environment information includes location information of the target object, depth image information, and current location information of each drone; An updating module, configured to update the moving route of each of the drones based on the location information of the target object and the current location information of each of the drones; A generation module is used to generate an obstacle avoidance strategy corresponding to the drone based on the depth image information and using a preset obstacle avoidance strategy model when the drone moves toward the target object according to the updated moving route, so that the drone executes the obstacle avoidance strategy and avoids obstacles.

9. An electronic device, characterized in that: The electronic device includes a memory and a processor, the memory stores a computer program, and the processor implements the artificial intelligence-based multi-UAV navigation method described in any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the multi-UAV navigation method based on artificial intelligence described in any one of claims 1 to 7 is implemented.

Images