Unmanned aerial vehicle auxiliary decision-making method, device, equipment, storage medium and product

Abstract

The invention discloses an unmanned aerial vehicle auxiliary decision-making method and device, equipment, a storage medium and a product, and relates to the technical field of unmanned aerial vehicles, and the unmanned aerial vehicle auxiliary decision-making method comprises the steps: responding to a data processing instruction, and detecting the data type of an unmanned aerial vehicle sensor; according to the data type of the unmanned aerial vehicle sensor, matching a corresponding capability component, and receiving and preprocessing unmanned aerial vehicle sensor data; the unmanned aerial vehicle sensor data is sent to a preset auxiliary decision-making model, an unmanned aerial vehicle auxiliary decision-making result is obtained, the auxiliary decision-making model comprises a target recognition model, a path planning model and a meteorological activity model, and parameters of the auxiliary decision-making model are adjusted in real time through a closed-loop feedback module based on the current unmanned aerial vehicle sensor data; and sending the unmanned aerial vehicle auxiliary decision result to the main control module. Sensor data processing and decision-making models are flexibly matched according to different task requirements, and the decision-making model parameters are fed back and adjusted by monitoring the state in real time, so that the decision-making generation efficiency of the unmanned aerial vehicle is improved.

Description

Technical Field

[0001] The present application relates to the field of drone technology, and in particular to drone-assisted decision-making methods, devices, equipment, storage media, and products. Background Art

[0002] With the expansion of drone application scenarios, such as search and rescue, patrol monitoring, and disaster monitoring, the demand for drones is growing. Existing drones usually rely on preset flight programs or manual control to complete tasks and lack autonomous decision-making capabilities.

[0003] In related technologies, traditional drone decision-making methods mainly rely on the computing resources of the main control module, and have problems such as protocol incompatibility, heavy computing burden, and limited number of interfaces, resulting in low efficiency in drone decision-making. Summary of the Invention

[0004] The main purpose of this application is to provide a drone-assisted decision-making method, device, equipment, storage medium and product, aiming to solve the technical problem of low efficiency in drone decision-making.

[0005] To achieve the above objectives, this application proposes a drone-assisted decision-making method, which includes:

[0006] In response to the data processing instruction, detecting a data type of a sensor of the drone;

[0007] According to the data type of the drone sensor, matching the corresponding capability component, receiving and preprocessing the drone sensor data;

[0008] The drone sensor data is sent to a preset auxiliary decision model to obtain a drone auxiliary decision result, wherein the auxiliary decision model includes a target recognition model based on reinforcement learning, a path planning model based on deep reinforcement learning, and a meteorological activity model based on data fusion. Based on the current drone sensor data, the parameters of the auxiliary decision model are adjusted in real time through a closed-loop feedback module;

[0009] The UAV auxiliary decision result is sent to the main control module, wherein the sending process is carried out through an optimized SPI communication protocol, and the optimized SPI communication protocol includes a dynamic adjustment of the data transmission rate and an automatic retransmission request mechanism.

[0010] In one embodiment, the step of matching corresponding capability components according to the data type of the drone sensor, and receiving and pre-processing the drone sensor data includes:

[0011] According to the data type of the drone sensor, the corresponding component version is searched in the preset component library to determine the corresponding capability component;

[0012] Based on the capability component, receiving and adaptively decoding drone sensor data;

[0013] Preprocessing the decoded drone sensor data, wherein the preprocessing includes normalizing and filtering the data.

[0014] In one embodiment, the step of sending the drone sensor data to a preset auxiliary decision model to obtain a drone auxiliary decision result includes:

[0015] Sending the drone sensor data to a preset target recognition auxiliary decision model;

[0016] Based on the target recognition auxiliary decision model, a convolutional neural network structure is used to extract feature information from the drone sensor data to form a target feature vector;

[0017] The extracted target feature vector is input into the fully connected network, and classification and recognition training is performed based on the preset adversarial sample generation algorithm;

[0018] Based on the training results, human targets, vehicle targets and building targets in the drone sensor data are identified, and the location information, size information and category information of the corresponding targets are generated to determine the target recognition auxiliary decision results.

[0019] In one embodiment, the step of sending the drone sensor data to a preset auxiliary decision model to obtain a drone auxiliary decision result further includes:

[0020] Sending the drone sensor data to a preset path planning auxiliary decision model;

[0021] Based on the path planning auxiliary decision model, a topological graph of the current environment is constructed through a multi-scale analysis method, where nodes represent reachable locations and edges represent path connection relationships;

[0022] Integrate obstacle information into the topological map and construct the path planning objective function based on the preset obstacle avoidance flight requirements of the UAV;

[0023] According to the path planning objective function, an expansion tree search algorithm is used to search for an optimal path on the topological map, and the optimal path is determined as a path planning auxiliary decision result.

[0024] In one embodiment, the step of sending the drone sensor data to a preset auxiliary decision model to obtain a drone auxiliary decision result further includes:

[0025] Sending the drone sensor data to a preset meteorological activity auxiliary decision model;

[0026] parsing meteorological information from drone sensor data based on the meteorological activity decision-making support model;

[0027] The meteorological information is integrated with preset meteorological forecast data, and an air quality early warning algorithm is used to identify potential risk items in the meteorological information;

[0028] Based on the potential risk items identified, corresponding flight recommendations and warning information are generated to determine the auxiliary decision-making results of meteorological activities.

[0029] In one embodiment, the step of adjusting the parameters of the auxiliary decision model in real time through a closed-loop feedback module based on current drone sensor data includes:

[0030] Monitor the current drone sensor data in real time to determine the current drone status;

[0031] Optimize the decision by pre-set parameters and adaptively adjust the parameters of the auxiliary decision model according to the current state of the drone;

[0032] Evaluate the performance of the adaptively adjusted auxiliary decision-making model, and determine whether to update the model parameters based on the evaluation results.

[0033] In addition, to achieve the above objectives, the present application also proposes a drone decision-making assistance device, which includes:

[0034] a detection module, configured to detect a data type of a drone sensor in response to a data processing instruction;

[0035] A matching module, configured to match corresponding capability components according to the data type of the drone sensor, and receive and pre-process the drone sensor data;

[0036] A decision-making module is configured to send the drone sensor data to a preset auxiliary decision-making model to obtain a drone auxiliary decision result, wherein the auxiliary decision-making model includes a target recognition model based on reinforcement learning, a path planning model based on deep reinforcement learning, and a meteorological activity model based on data fusion. Based on the current drone sensor data, the parameters of the auxiliary decision-making model are adjusted in real time through a closed-loop feedback module;

[0037] The sending module is used to send the UAV auxiliary decision-making result to the main control module, wherein the sending process is carried out through the optimized SPI communication protocol, and the optimized SPI communication protocol includes dynamically adjusting the data transmission rate and the automatic retransmission request mechanism.

[0038] In addition, to achieve the above-mentioned purpose, the present application also proposes a drone-assisted decision-making device, which includes: a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein the computer program is configured to implement the steps of the drone-assisted decision-making method as described above.

[0039] In addition, to achieve the above-mentioned purpose, the present application also proposes a storage medium, which is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by the processor, the steps of the drone assisted decision-making method as described above are implemented.

[0040] In addition, to achieve the above-mentioned purpose, the present application also provides a computer program product, which includes a computer program, and when the computer program is executed by a processor, it implements the steps of the drone assisted decision-making method as described above.

[0041] One or more technical solutions proposed in this application have at least the following technical effects:

[0042] Compared to the conventional drone decision-making methods in related technologies, which primarily rely on the computing resources of the main control module and suffer from protocol incompatibility, heavy computing burden, and limited number of interfaces, resulting in low drone decision-making efficiency, the present application detects the data type of the drone sensor in response to data processing instructions; matches the corresponding capability component based on the data type of the drone sensor, receives and pre-processes the drone sensor data; sends the drone sensor data to a preset auxiliary decision model to obtain a drone auxiliary decision result, wherein the auxiliary decision model includes a target recognition model based on reinforcement learning, a path planning model based on deep reinforcement learning, and a meteorological activity model based on data fusion. Based on the current drone sensor data, the parameters of the auxiliary decision model are adjusted in real time through a closed-loop feedback module; and the drone auxiliary decision result is sent to the main control module. The sending process is carried out through an optimized SPI communication protocol, which includes a dynamic adjustment of the data transmission rate and an automatic retransmission request mechanism. It can be understood that the present application uses the capability component mechanism to flexibly match sensor data processing and decision models according to different task requirements, thereby improving the efficiency of drone decision-making. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0044] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0045] Figure 1 A flowchart of the first embodiment of the drone-assisted decision-making method of this application is provided;

[0046] Figure 2 A flowchart of the second embodiment of the drone-assisted decision-making method of this application is provided;

[0047] Figure 3 This is a schematic diagram of the module structure of the drone decision-making support device according to an embodiment of the present application;

[0048] Figure 4 This is a schematic diagram of the device structure of the hardware operating environment involved in the drone assisted decision-making method in the embodiment of the present application.

[0049] The purpose, features and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0050] It should be understood that the specific embodiments described herein are merely used to explain the technical solutions of the present application and are not intended to limit the present application.

[0051] In order to better understand the technical solution of the present application, a detailed description will be given below in conjunction with the accompanying drawings and specific implementation methods.

[0052] The main solutions of the embodiments of this application are:

[0053] In response to the data processing instruction, detecting a data type of a sensor of the drone;

[0054] According to the data type of the drone sensor, matching the corresponding capability component, receiving and preprocessing the drone sensor data;

[0055] The drone sensor data is sent to a preset auxiliary decision model to obtain a drone auxiliary decision result, wherein the auxiliary decision model includes a target recognition model based on reinforcement learning, a path planning model based on deep reinforcement learning, and a meteorological activity model based on data fusion. Based on the current drone sensor data, the parameters of the auxiliary decision model are adjusted in real time through a closed-loop feedback module;

[0056] The UAV auxiliary decision result is sent to the main control module, wherein the sending process is carried out through an optimized SPI communication protocol, and the optimized SPI communication protocol includes a dynamic adjustment of the data transmission rate and an automatic retransmission request mechanism.

[0057] In this embodiment, the application uses a drone decision-making assistance device as the execution subject. For the sake of convenience, it is specifically described below as "device".

[0058] Since the traditional drone decision-making method in existing technology mainly relies on the computing resources of the main control module, there are problems such as protocol incompatibility, heavy computing burden, and limited number of interfaces, resulting in low efficiency of drone decision-making.

[0059] The present application provides a solution that detects the data type of a drone sensor in response to a data processing instruction; matches the corresponding capability component based on the data type of the drone sensor, receives and pre-processes the drone sensor data; sends the drone sensor data to a preset auxiliary decision model to obtain a drone auxiliary decision result, wherein the auxiliary decision model includes a target recognition model based on reinforcement learning, a path planning model based on deep reinforcement learning, and a meteorological activity model based on data fusion. Based on the current drone sensor data, the parameters of the auxiliary decision model are adjusted in real time through a closed-loop feedback module; and the drone auxiliary decision result is sent to a main control module. The sending process is carried out through an optimized SPI communication protocol, and the optimized SPI communication protocol includes a dynamic adjustment of the data transmission rate and an automatic retransmission request mechanism. The present application uses a capability component mechanism to flexibly match sensor data processing and decision models according to different task requirements, thereby improving the efficiency of drone decision generation.

[0060] Based on this, the embodiment of the present application provides a drone-assisted decision-making method, referring to Figure 1 , Figure 1 This is a flow chart of the first embodiment of the drone-assisted decision-making method of this application.

[0061] In this embodiment, the drone-assisted decision-making method includes steps S10 to S40:

[0062] Step S10, detecting the data type of the drone sensor in response to the data processing instruction;

[0063] It should be noted that data processing instructions are a series of electronic signals or commands issued by the drone system to instruct the drone's data processing unit to perform specific data collection, processing, and analysis tasks. These instructions are usually generated by the main control module in response to specific flight missions or environmental changes.

[0064] Drone sensors refer to a series of hardware devices installed on drones that can detect and measure the drone's flight status, surrounding environment, and other relevant parameters. These sensors include but are not limited to GPS, IMU (Inertial Measurement Unit), altimeter, visual camera, infrared sensor, etc.

[0065] Data type refers to the category or format of the data collected by drone sensors. For example, a vision sensor might produce image data, while an IMU might produce acceleration and angular velocity data. Each data type requires specific processing methods and algorithms to interpret and utilize it.

[0066] Understandably, this step is the starting point of the drone data processing flow, and its purpose is to ensure that the drone can correctly understand and process the data from the various sensors it carries.

[0067] Specifically, when a drone receives a data processing instruction, it initiates a detection process to identify and classify the data from each sensor. This process is crucial for subsequent data preprocessing, analysis, and decision-making, as different types of data require different processing methods. For example, image data may require analysis using an image recognition algorithm, while IMU data requires a posture estimation algorithm.

[0068] Step S20, matching corresponding capability components according to the data type of the drone sensor, receiving and preprocessing the drone sensor data;

[0069] It should be noted that drone sensor data types refer to the categories of data collected by the sensors installed on the drone. These categories vary based on the sensor's functionality and measurement parameters. For example, a vision sensor may generate image data, while a barometer may generate air pressure data. Capability components refer to software modules or hardware units within a drone's data processing system that are specifically designed to process specific types of sensor data. These components possess the ability to parse, decode, and pre-process specific data types. Matching refers to the process by which the system automatically identifies and selects the appropriate capability component based on the sensor data type. This process may involve searching a pre-defined component library to determine the component best suited for processing a specific data type. Receiving refers to the act of a capability component acquiring data from a drone sensor. This typically involves hardware interfaces and communication protocols to ensure data can be transmitted from the sensor to the processing unit. Preprocessing refers to the preliminary processing steps performed on raw sensor data before further analysis. Preprocessing may include filtering, normalization, format conversion, data augmentation, and other operations designed to improve data quality and make it more suitable for subsequent processing and analysis.

[0070] For example, this step is a key part of the drone's data processing and decision-making process. This step ensures that the drone can efficiently and accurately process large amounts of data from different sensors.

[0071] Specifically, drones generate and receive large amounts of sensor data while performing missions. This data must be correctly identified and classified so that the device can match it with the appropriate capability component for processing. For example, if a drone is equipped with a visual sensor, the device will be matched with a capability component capable of processing image data, such as an image recognition module.

[0072] Once the corresponding capability component is matched, the device begins receiving data streams from sensors. This data may be in the form of analog or digital signals and must be received through the appropriate interface and communication protocol. Upon receipt, the capability component preprocesses the data to ensure quality and consistency. Preprocessing steps may include noise removal, data range adjustment, and data format conversion to facilitate subsequent analysis and decision-making.

[0073] In summary, this step involves not only technical data matching and processing, but also the integrity and coordination of system design to ensure the drone can operate stably in various environments and conditions. In this way, the drone can better adapt to complex and changing environments, improving its performance and reliability in various application scenarios.

[0074] In a feasible implementation, step S20 may include the following steps:

[0075] According to the data type of the drone sensor, the corresponding component version is searched in the preset component library to determine the corresponding capability component;

[0076] Based on the capability component, receiving and adaptively decoding drone sensor data;

[0077] Preprocessing the decoded drone sensor data, wherein the preprocessing includes normalizing and filtering the data.

[0078] It should be noted that drone sensor data types refer to the formats and categories of data collected by drone-mounted sensors. These data types may include, but are not limited to, image data, temperature data, pressure data, and position data, each corresponding to different physical quantities and information. A pre-defined component library refers to a collection of software components pre-configured and stored within a drone system. These components are designed to process specific types of sensor data. The component library allows the system to quickly find and use the appropriate processing component based on the sensor data type. A component version refers to the specific version number of each capability component in the library, identifying different iterations and update states of the component. Version control helps ensure consistent and reliable data processing and facilitates traceability and maintenance. Capability components are software modules within the drone system that process specific sensor data. These components have the ability to receive, decode, and pre-process sensor data to facilitate subsequent data analysis and decision-making. Receiving refers to the process by which capability components acquire raw data from drone sensors. This typically involves hardware interfaces and communication protocols to ensure secure and accurate data transmission from the sensor to the processing unit. Decoding refers to the process of converting raw sensor data into a format suitable for further processing. This may include converting analog signals to digital signals, parsing data packet structures, and converting encoding formats. Preprocessing refers to a series of preliminary processing steps performed on decoded data before further analysis. This may include filtering, normalization, data augmentation, feature extraction, and other operations aimed at improving the quality of the data and making it more suitable for subsequent processing and analysis.

[0079] For example, when a drone sensor collects data, the device first identifies the data type. It then searches a pre-defined component library for a component capable of processing that data type. The library contains a variety of capable components, each designed for a specific data type to ensure specialized data processing. The device selects the most appropriate component based on the data type and component version to ensure accurate and efficient data processing. Once a capable component has been identified, the device uses it to receive raw sensor data. This reception process may involve direct communication with the sensor or through an intermediary such as a data bus. The received data is typically in raw format and requires decoding to convert it into a more easily processable format. The decoding process may include steps such as parsing data packets, converting signal formats, and identifying data protocols. The decoded data requires preprocessing to improve data quality and prepare it for subsequent analysis. Preprocessing steps may include noise removal, imputing missing values, normalizing the data range, and enhancing signals. These steps help improve data usability, reduce errors in subsequent processing, and enhance the accuracy of data analysis.

[0080] In summary, through this coherent process, drones can transform raw sensor data into useful information, supporting various functions such as flight control, environmental perception, and mission execution. This data processing flow design not only improves the drone's autonomy but also enhances its adaptability and mission execution efficiency in complex environments.

[0081] Step S30: Sending the drone sensor data to a preset decision-making support model to obtain a drone decision-making support result, wherein the decision-making support model includes a target recognition model based on reinforcement learning, a path planning model based on deep reinforcement learning, and a meteorological activity model based on data fusion. Based on the current drone sensor data, the parameters of the decision-making support model are adjusted in real time through a closed-loop feedback module;

[0082] It should be noted that drone sensor data refers to the raw data collected by various sensors onboard drones. This data reflects the drone's flight status, surrounding environment information, and other relevant parameters. Data types may include location information, speed, attitude, image, temperature, air pressure, etc.

[0083] Pre-set decision-making support models refer to a series of algorithms and models pre-set in the drone system. These models are used to process and analyze sensor data to assist the drone in making decisions. These models may include machine learning models, deep learning networks, rule engines, etc.

[0084] Decision-support results refer to the output generated by the decision-support model after processing sensor data. These results typically include an assessment of the drone's current state, an understanding of the environment, and recommendations for future actions. These results can help the drone perform tasks such as path planning, obstacle avoidance, and target identification.

[0085] For example, data collected by drone sensors needs to be transmitted to a processing unit, typically via an internal communication network. Data transmission must ensure data integrity and timeliness so that the decision-making support model receives the required information promptly and accurately. The pre-set decision-making support model processes and analyzes the received sensor data. This may include steps such as data cleaning, feature extraction, and pattern recognition. For example, image data may be processed using a convolutional neural network for object recognition, while location data may be used for path planning using a machine learning algorithm. The decision-making support model not only analyzes the data but also makes decisions based on the analysis results. These decisions may include selecting a flight path, implementing obstacle avoidance strategies, or tracking a target. The decision results must be sufficiently clear for the drone's main control system to implement. The decision-making support results must be output in a usable form for the drone's main control system. These outputs may include control signals, warning messages, and recommended courses of action. These results directly impact the drone's behavior and performance. The drone system may provide feedback based on the effectiveness of the decision-making support results to optimize and adjust the decision-making support model. This iterative process helps improve the system's accuracy and adaptability, enabling the drone to operate stably in various environments and conditions.

[0086] Through this process, drones can achieve a higher level of autonomy and intelligence, improving their performance and reliability in various application scenarios. This data processing and decision-making process design not only improves the safety and efficiency of drones, but also opens up new possibilities for the development of drone technology.

[0087] In a feasible implementation, step S30 may include the following steps:

[0088] Sending the drone sensor data to a preset target recognition auxiliary decision model;

[0089] Based on the target recognition auxiliary decision model, a convolutional neural network structure is used to extract feature information from the drone sensor data to form a target feature vector;

[0090] The extracted target feature vector is input into the fully connected network, and classification and recognition training is performed based on the preset adversarial sample generation algorithm;

[0091] Based on the training results, human targets, vehicle targets and building targets in the drone sensor data are identified, and the location information, size information and category information of the corresponding targets are generated to determine the target recognition auxiliary decision results.

[0092] It should be noted that drone sensor data refers to the raw data collected by various sensors onboard drones. This data includes, but is not limited to, images, temperature, air pressure, position, and speed, and is used to reflect the drone's flight status and surrounding environment. A preset target recognition decision-making support model is an algorithmic model pre-set within the drone system specifically for target recognition tasks. This model can process sensor data and identify and classify different targets, such as people, vehicles, and buildings. A convolutional neural network is a deep learning model that automatically extracts feature information from image data through a structure consisting of convolutional, pooling, and fully connected layers.

[0093] In target recognition, CNNs can identify local features in images and abstract them layer by layer to form high-level feature representations. Feature information refers to useful information extracted from drone sensor data. This information can represent important attributes of the data and is used for subsequent classification and recognition tasks. The target feature vector is a vector representation of the feature information extracted by a convolutional neural network. It contains the key attributes of the target and is used for target classification and recognition. A fully connected network is a structure in a neural network in which each neuron is connected to all neurons in the previous layer. In target recognition, the fully connected network receives the feature vectors extracted by the CNN and performs further classification and recognition training. Classification and recognition training refers to training the fully connected network using labeled training data to enable it to recognize different target categories. The target recognition decision-making support result is the output generated by the target recognition decision-making support model after processing sensor data. It includes the target's location, size, and category information, which is used to assist the drone in making decisions.

[0094] Exemplarily, the sensor data collected by the drone is sent to a preset target recognition auxiliary decision model. This step ensures that the data can be correctly transmitted to the processing unit for subsequent target recognition tasks. Using the convolutional neural network structure, feature information is automatically extracted from the drone sensor data. Through its multi-layer structure, CNN can identify features such as edges, textures, and shapes in the image and form a target feature vector; the extracted feature information is organized into a target feature vector, which is a point in a high-dimensional space that contains the key attributes of the target and provides a basis for subsequent classification and recognition. The target feature vector is input into the fully connected network for classification and recognition training;

[0095] This process involves using large amounts of labeled data to train the network so that it can learn to distinguish the features of different target categories. Based on the training results, the fully connected network can identify human targets, vehicle targets, and building targets in the drone sensor data, and generate the location information, size information, and category information of the corresponding targets. The recognition results are organized into target recognition auxiliary decision results. These results include not only the category information of the target, but also the spatial information of the target, providing support for the drone's flight decisions. The target recognition auxiliary decision results can be used for drone navigation, obstacle avoidance, mission planning, etc. At the same time, the accuracy and effectiveness of these results can also be used as feedback to optimize and adjust the target recognition model, thereby improving the drone's recognition capabilities and decision-making efficiency.

[0096] Through this process, drones can achieve intelligent perception and understanding of their surroundings, improving their adaptability and mission execution capabilities in complex environments. This design of a target recognition and decision-making support process not only enhances drone autonomy but also opens up new possibilities for the development of drone technology.

[0097] In a feasible implementation, step S30 may include the following steps:

[0098] Sending the drone sensor data to a preset path planning auxiliary decision model;

[0099] Based on the path planning auxiliary decision model, a topological graph of the current environment is constructed through a multi-scale analysis method, where nodes represent reachable locations and edges represent path connection relationships;

[0100] Integrate obstacle information into the topological map and construct the path planning objective function based on the preset obstacle avoidance flight requirements of the UAV;

[0101] According to the path planning objective function, an expansion tree search algorithm is used to search for an optimal path on the topological map, and the optimal path is determined as a path planning auxiliary decision result.

[0102] It should be noted that drone sensor data refers to the raw data collected by various sensors onboard drones. This data includes, but is not limited to, position, speed, attitude, image, temperature, air pressure, and other information, reflecting the drone's flight status and surrounding environment. The pre-set path planning decision-making support model is a pre-configured algorithmic model within the drone system specifically designed for path planning tasks. This model processes sensor data, constructs a topological map of the environment, and plans the optimal flight path.

[0103] A topological graph is a mathematical model used to represent the connectivity between points in a space. In path planning, a topological graph consists of nodes and edges. Nodes represent key locations in space, and edges represent the paths connecting these locations. In a topological graph, nodes represent reachable locations and are the basic units of path planning. A node can be any location in space, such as the starting point, the end point, or any turning point in a path. In a topological graph, edges represent the connectivity of a path, connecting two nodes and indicating a feasible path from one node to another.

[0104] Obstacle information refers to data about obstacles that a drone needs to avoid during flight, including information such as their location, size, and shape. The path planning objective function is a mathematical function used to evaluate the pros and cons of different paths and find the optimal one. This function typically considers factors such as path length, safety, and energy consumption. The expansion tree search algorithm is an algorithm used to find the optimal path from a starting point to an end point in a graph. It explores possible paths by gradually expanding nodes and selects the optimal one. The path planning decision-making support result is the output generated by the path planning decision-making support model after processing sensor data. It includes detailed information about the optimal path and is used to guide the drone's flight.

[0105] For example, the sensor data collected by the drone is sent to a pre-set path planning auxiliary decision model. This step ensures that the data can be correctly transmitted to the processing unit for subsequent path planning tasks.

[0106] Using drone sensor data, the path planning and decision-making model constructs a topological map of the current environment. The topological map is constructed based on the drone's perception of the environment, including information such as terrain, buildings, and obstacles.

[0107] Integrating obstacle information into the topology map to ensure that these obstacles can be avoided during path planning is crucial to ensuring the flight safety of the drone.

[0108] Based on the preset obstacle avoidance flight requirements of the UAV, a path planning objective function is constructed. The objective function defines the optimization goal of path planning, such as the shortest path, the safest path, etc.

[0109] Use the expansion tree search algorithm to search for the optimal path in a topological graph. The algorithm starts from the starting point, gradually expands the nodes, evaluates the cost of each path, and selects the optimal path.

[0110] Based on the results of the expansion tree search algorithm, the optimal path is determined and a path planning auxiliary decision result is generated. These results include detailed information about the path, such as path length, turning points, and altitude changes.

[0111] The results of path planning aided decision-making can be used for navigation and flight control of the UAV. Furthermore, the accuracy and effectiveness of these results can be used as feedback to optimize and adjust the path planning model, improving the UAV’s path planning capabilities and flight efficiency.

[0112] Through this process, drones can achieve intelligent path planning in complex environments, improving their adaptability and mission execution capabilities in complex environments. This path planning and decision-making process design not only improves drone autonomy, but also provides new possibilities for the development of drone technology.

[0113] In a feasible implementation, step S30 may include the following steps:

[0114] Sending the drone sensor data to a preset meteorological activity auxiliary decision model;

[0115] parsing meteorological information from drone sensor data based on the meteorological activity decision-making support model;

[0116] The meteorological information is integrated with preset meteorological forecast data, and an air quality early warning algorithm is used to identify potential risk items in the meteorological information;

[0117] Based on the potential risk items identified, corresponding flight recommendations and warning information are generated to determine the auxiliary decision-making results of meteorological activities.

[0118] It should be noted that drone sensor data refers to the raw data collected by various sensors onboard drones. This data includes, but is not limited to, meteorological parameters such as temperature, humidity, air pressure, wind speed, and wind direction, as well as possible image and sound data, reflecting the drone's flight status and surrounding environment. A pre-set meteorological activity decision-making support model refers to an algorithmic model pre-configured within the drone system specifically for meteorological information processing and decision support. This model processes sensor data, interprets meteorological information, and provides flight decision support. Meteorological information refers to data collected by drone sensors regarding atmospheric conditions, including temperature, humidity, air pressure, wind speed, and wind direction. This data is crucial for assessing flight conditions and safety.

[0119] Meteorological forecast data refers to predicted information about future weather conditions provided by weather stations or weather forecast systems, including forecast data such as temperature, precipitation, and wind speed. An air quality warning algorithm refers to a computational method used to assess air quality conditions and predict potential risks. Based on meteorological information and forecast data, this algorithm can identify air quality issues that may affect drone flight. Potential risk items refer to factors identified based on meteorological information and the air quality warning algorithm that may pose a threat to drone flight safety, such as thunderstorms, strong winds, and haze. Flight recommendations and warning information refer to guidance information provided based on the results of meteorological activity decision support. These include recommendations for flight path adjustments, altitude changes, flight speed control, and possible risk warnings. The results of meteorological activity decision support are the output generated by the meteorological activity decision support model after processing sensor data. They include flight recommendations, risk warnings, and other information used to guide drone flight decisions.

[0120] For example, the sensor data collected by the drone is sent to a pre-set meteorological activity decision-making support model. This step ensures that the data is correctly transmitted to the processing unit for subsequent meteorological information analysis and decision support. Using the meteorological activity decision-making support model, meteorological information is extracted from the drone's sensor data. This process may involve data preprocessing, feature extraction, and parameter conversion to ensure the accuracy and usability of the meteorological information. The extracted meteorological information is then integrated with pre-set weather forecast data. This step helps improve the accuracy and predictive power of meteorological information, providing drones with more comprehensive meteorological environmental information.

[0121] Using air quality warning algorithms, potential risk items are identified from the integrated meteorological information. This step is crucial for preventing flight accidents and ensuring flight safety.

[0122] Based on the potential risks identified, corresponding flight recommendations and warning information are generated. These information will guide the drone to make necessary flight adjustments to avoid risks.

[0123] Based on the above steps, the meteorological activity decision-making support results are determined. These results include not only specific flight recommendations but also flight safety assessments and warnings.

[0124] The results of meteorological activity decision-making support can be used for drone navigation and flight control. Furthermore, the accuracy and effectiveness of these results can be used as feedback to optimize and adjust meteorological activity decision-making support models, improving the drone's meteorological risk assessment capabilities and flight safety.

[0125] Through this process, drones can achieve intelligent perception and response to meteorological conditions, improving their adaptability and mission execution capabilities in complex weather environments. This design of a meteorological activity-assisted decision-making process not only enhances drone autonomy but also opens up new possibilities for the development of drone technology.

[0126] Step S40: Send the UAV auxiliary decision result to the main control module, wherein the sending process is performed through the optimized SPI communication protocol, and the optimized SPI communication protocol includes dynamic adjustment of data transmission rate and automatic retransmission request mechanism.

[0127] It should be noted that the drone's assisted decision-making results refer to the outputs generated by the assisted decision-making model after processing sensor data. These outputs include but are not limited to flight path suggestions, obstacle avoidance strategies, target identification information, environmental assessments, etc., and are intended to assist drones in making more intelligent and safer flight decisions.

[0128] The main control module (MCM) is the core processing unit within a drone system responsible for coordinating and managing all of the drone's functions. It typically processes the output from the decision-making support system and translates it into specific flight control commands, such as adjusting altitude, changing flight direction, and controlling speed.

[0129] Dynamic data rate adjustment involves changing the data transmission speed in real time during a drone's flight, based on actual communication needs and environmental conditions. For example, when a drone is flying at high speed or performing urgent missions, it needs to transmit large amounts of data quickly to ensure real-time transmission. In these situations, the data transmission rate can be increased. Meanwhile, when the drone is flying at low speed or in standby mode, the transmission rate can be appropriately reduced to conserve energy and minimize interference. This dynamic adjustment mechanism optimizes communication efficiency and enables the communication system to better adapt to the data transmission needs of the drone under different operating conditions.

[0130] The automatic retransmission request mechanism allows for the transmission of data from a transmitter to a receiver that fails to receive the data correctly (e.g., due to interference, errors, or other reasons resulting in data loss or corruption). The receiver automatically sends a retransmission request to the transmitter, which then retransmits the data. This mechanism ensures the integrity and accuracy of data transmission, effectively guaranteeing the reliable transmission of important information, such as drone-assisted decision-making results, even in the presence of poor communication link quality or interference.

[0131] For example, the decision support system generates a series of decision results based on drone sensor data and a pre-set decision model. These results may be recommendations based on data analysis, such as path planning results, obstacle avoidance instructions, target tracking information, etc.

[0132] The generated decision-making assistance results need to be transmitted to the main control module via the drone's internal communication system. This transmission process must ensure data integrity, security, and real-time performance, ensuring that the main control module accurately receives the decision results. To ensure effective communication between the decision-making assistance results and the main control module, the drone system may employ specific data interfaces and communication protocols. These interfaces and protocols define the data transmission format, rate, and control mechanisms, ensuring smooth data transfer between the different modules of the drone. After receiving the decision-making assistance results, the main control module performs further processing based on them. This may include verifying the results, integrating them with data from other systems, and converting the results into specific flight control commands. The main control module converts the processed decision-making results into flight control commands, such as adjusting the drone's flight attitude, speed, and altitude to execute the intended flight mission. These operations ensure that the drone can safely and effectively complete the mission based on the decision-making assistance results. After executing the flight maneuver, the main control module collects feedback to evaluate the accuracy and effectiveness of the decision-making results. This feedback can be used to optimize the decision-making assistance model, improving the drone's decision-making capabilities and flight performance.

[0133] Through this coherent process, drones can achieve a higher level of autonomous flight and intelligent decision-making, improving their adaptability and reliability in various environments and conditions. This process design, which transmits auxiliary decision-making results to the main control module, not only improves the safety and efficiency of drones, but also opens up new possibilities for the development of drone technology.

[0134] This embodiment provides a drone-assisted decision-making method that, through a capability component mechanism, flexibly matches sensor data processing and decision-making models according to different mission requirements, thereby improving the efficiency of drone decision-making.

[0135] In a feasible implementation, the step of adjusting the parameters of the auxiliary decision model in real time through a closed-loop feedback module based on current drone sensor data includes:

[0136] Monitor the current drone sensor data in real time to determine the current drone status;

[0137] Optimize the decision by pre-set parameters and adaptively adjust the parameters of the auxiliary decision model according to the current state of the drone;

[0138] Evaluate the performance of the adaptively adjusted auxiliary decision-making model, and determine whether to update the model parameters based on the evaluation results.

[0139] It should be noted that real-time monitoring means that the UAV system continuously observes and detects the data collected by sensors in order to obtain the latest status information of the UAV and its surrounding environment in a timely manner, ensure the timeliness and accuracy of the data, and provide a basis for subsequent decision-making model parameter adjustments.

[0140] The current drone sensor data is the raw information collected by the drone through various sensors at the current moment, including but not limited to location information (such as GPS coordinates), speed, attitude (such as heading angle, pitch angle, etc.), image data (from the camera), and meteorological data (such as air pressure, temperature, humidity, etc.). These data reflect the drone's own status and the surrounding environment.

[0141] The current drone status is a comprehensive assessment of the drone's flight conditions based on sensor data, such as flight altitude, flight speed, whether the flight attitude is stable, whether it is on the planned flight route, and whether there are obstacles or weather risks in the surrounding area. It describes the drone's overall current operating conditions.

[0142] Preset parameter optimization decisions are a set of pre-defined rules and strategies within the drone system that adjust the parameters of the decision-making support model based on the drone's state. These rules are based on extensive experimental data, flight experience, and a deep understanding of the drone's mission requirements. Their goal is to ensure that the decision-making support model operates optimally under different conditions.

[0143] Decision-support model parameters are adjustable variables within decision-support models (such as target recognition models, path planning models, and weather activity models). These variables influence the model's processing and analysis of input data. For example, in a path planning model, parameters might include obstacle distance weights and flight path smoothness coefficients. Adjusting these parameters can alter the flight path planning results generated by the model.

[0144] Adaptive adjustment is to automatically and flexibly change the parameters of the auxiliary decision-making model according to the changes in the current UAV state, so that the model can better adapt to the current flight environment and mission requirements and optimize the decision results.

[0145] When updating model parameters, if the evaluation results show that the adjusted parameters can significantly improve the performance of the decision-making support model, these new parameter values ​​will be applied to the model to replace the original parameters to ensure that the model can continue to provide better results in subsequent decision-making processes.

[0146] For example, consider a drone conducting a search and rescue mission in a mountainous area. Equipped with multiple sensors, including a camera, GPS, and weather sensors, the system monitors the drone's weather sensor data for a sudden increase in wind speed, while GPS data indicates the drone has deviated from its intended route. The system then determines that the drone is experiencing strong winds and a course deviation.

[0147] Based on pre-set parameters, the system automatically adjusts the parameters of the decision-support model. For the path planning model, the system increases the weighting of wind speed and adjusts the safety factor in the path planning objective function, enabling the model to generate safer flight paths that better meet the current complex weather conditions. The adjusted path planning model will be more inclined to plan flight routes along the leeward slopes of mountains and away from complex obstacles such as tall trees, thereby reducing the impact of strong winds on the drone.

[0148] To evaluate the performance of the adaptively adjusted decision-making support model, the system simulates flight data or historical data from similar scenarios, calculating safety metrics (such as minimum distance to obstacles and path stability) and mission efficiency metrics (such as the estimated time to reach the target area) for the new path planning results and comparing them with the results of the pre-adjustment model. If the evaluation results show that the new model outperforms the original model in both safety and efficiency, the model parameters are updated and applied to actual flight decisions to guide the drone to continue its search and rescue mission safely and efficiently.

[0149] For example, in order to help understand the implementation process of the drone auxiliary decision-making method obtained by combining this embodiment with the above embodiment 1, please refer to Figure 2 , Figure 2 A brief flowchart of a drone-assisted decision-making method is provided. Specifically:

[0150] The UAV system responds to the data processing instruction and begins to detect the sensor data type.

[0151] According to the sensor data type, the corresponding capability component is matched in the component library, the sensor data is received and decoded, and then preprocessed.

[0152] The preprocessed data is sent to the preset auxiliary decision model for processing.

[0153] In the target recognition auxiliary decision model, a convolutional neural network is used to extract feature information to form a feature vector, which is then input into a fully connected network for classification and recognition training;

[0154] In the path planning auxiliary decision model, a topological map of the current environment is constructed, obstacle information is incorporated, and the optimal path is planned.

[0155] In the meteorological activity decision support model, meteorological information is analyzed, meteorological forecast data is integrated, potential risk items are identified, and flight recommendations and warning information are generated.

[0156] The auxiliary decision results are sent to the main control module, and the main control module executes corresponding flight decisions based on these results.

[0157] It should be noted that the above examples are only used to understand the present application and do not constitute a limitation on the drone-assisted decision-making method of the present application. More simple transformations based on this technical concept are all within the scope of protection of the present application.

[0158] This application also provides a drone decision-making assistance device, please refer to Figure 3 , the UAV decision-making assistance device includes:

[0159] a detection module 10 for detecting a data type of a drone sensor in response to a data processing instruction;

[0160] A matching module 20 is used to match corresponding capability components according to the data type of the drone sensor, and receive and pre-process the drone sensor data;

[0161] A decision module 30 is configured to send the drone sensor data to a preset auxiliary decision model to obtain a drone auxiliary decision result, wherein the auxiliary decision model includes a target recognition model based on reinforcement learning, a path planning model based on deep reinforcement learning, and a meteorological activity model based on data fusion. The parameters of the auxiliary decision model are adjusted in real time based on the current drone sensor data through a closed-loop feedback module;

[0162] The sending module 40 is used to send the UAV auxiliary decision-making result to the main control module, wherein the sending process is carried out through the optimized SPI communication protocol, and the optimized SPI communication protocol includes dynamic adjustment of data transmission rate and automatic retransmission request mechanism.

[0163] And / or, the matching module 20 includes:

[0164] A first determination module is configured to search a preset component library for a corresponding component version according to the data type of the drone sensor and determine a corresponding capability component;

[0165] a first receiving module, configured to receive and adaptively decode drone sensor data based on the capability component;

[0166] The first decoding module is used to preprocess the decoded drone sensor data, wherein the preprocessing includes normalizing and filtering the data.

[0167] And / or, the decision module 30 includes:

[0168] A first sending module is used to send the drone sensor data to a preset target recognition auxiliary decision model;

[0169] A first extraction module is used to extract feature information from the drone sensor data using a convolutional neural network structure based on the target recognition auxiliary decision model to form a target feature vector;

[0170] The first classification module is used to input the extracted target feature vector into the fully connected network and perform classification and recognition training based on the preset adversarial sample generation algorithm;

[0171] The first recognition module is used to identify human targets, vehicle targets and building targets in the drone sensor data based on the training results, and generate the location information, size information and category information of the corresponding targets to determine the target recognition auxiliary decision results.

[0172] And / or, the decision module 30 includes:

[0173] A second sending module is used to send the drone sensor data to a preset path planning auxiliary decision model;

[0174] A first construction module is configured to construct a topological graph of the current environment based on the path planning auxiliary decision model using a multi-scale analysis method, wherein nodes represent reachable locations and edges represent connection relationships of paths;

[0175] The second building block is used to integrate obstacle information into the topological map and construct the path planning objective function according to the preset obstacle avoidance flight requirements of the UAV;

[0176] The first search module is used to search for the optimal path on the topological map using an expansion tree search algorithm according to the path planning objective function, and determine the optimal path as the path planning auxiliary decision result.

[0177] And / or, the decision module 30 includes:

[0178] A third sending module is used to send the drone sensor data to a preset meteorological activity auxiliary decision model;

[0179] A first parsing module is configured to parse meteorological information in the drone sensor data based on the meteorological activity decision support model;

[0180] a second identification module, configured to integrate the meteorological information with preset meteorological forecast data and identify potential risk items in the meteorological information using an air quality early warning algorithm;

[0181] The second determination module is used to generate corresponding flight recommendations and warning information based on the identified potential risk items and determine the auxiliary decision-making results of meteorological activities.

[0182] And / or, the decision module 30 includes:

[0183] The first real-time monitoring module is used to monitor the current drone sensor data in real time and determine the current drone status;

[0184] A first adaptive adjustment module is used to optimize the decision by using preset parameters and adaptively adjust the parameters of the auxiliary decision model according to the current state of the drone;

[0185] The first performance evaluation module is used to evaluate the performance of the adaptively adjusted auxiliary decision model and determine whether to update the model parameters according to the evaluation results.

[0186] The drone decision-making assistance device provided in this application utilizes the drone decision-making assistance method described in the aforementioned embodiment to address the low efficiency of drone decision-making. Compared to the prior art, the drone decision-making assistance device provided in this application achieves the same beneficial effects as the drone decision-making assistance method described in the aforementioned embodiment. Other technical features of the drone decision-making assistance device are the same as those disclosed in the aforementioned embodiment and are not further elaborated here.

[0187] The present application provides a drone-assisted decision-making device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the drone-assisted decision-making method in the above-mentioned embodiment one.

[0188] Reference below Figure 4 , which shows a schematic structural diagram of a drone decision-making assistance device suitable for implementing the embodiments of the present application. The drone decision-making assistance device in the embodiments of the present application may include, but is not limited to, mobile terminals such as mobile phones, tablet computers, laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PMPs (Portable Media Players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), and fixed terminals such as digital TVs and desktop computers. Figure 4 The drone decision-making assistance device shown is merely an example and should not limit the functions and scope of use of the embodiments of the present application.

[0189] like Figure 4As shown, the drone decision-making support device may include a processing device 1001 (e.g., a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes based on programs stored in a read-only memory (ROM) 1002 or programs loaded from a storage device 1003 into a random access memory (RAM) 1004. RAM 1004 also stores various programs and data required for the operation of the drone decision-making support device. Processing device 1001, ROM 1002, and RAM 1004 are interconnected via a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to I / O interface 1006: input devices 1007, such as a touchscreen, touchpad, keyboard, mouse, image sensor, microphone, accelerometer, gyroscope, etc.; output devices 1008, such as a liquid crystal display (LCD), speaker, vibrator, etc.; storage device 1003, such as a magnetic tape or hard disk; and communication devices 1009. Communication device 1009 can allow the drone-assisted decision-making device to communicate wirelessly or wired with other devices to exchange data. Although the figure shows a drone-assisted decision-making device with various systems, it should be understood that it is not required to implement or include all of the illustrated systems. More or fewer systems may alternatively be implemented or included.

[0190] In particular, according to the embodiments disclosed in the present application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, the embodiments disclosed in the present application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program comprising program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network via a communication device, or installed from a storage device 1003, or installed from a ROM 1002. When the computer program is executed by the processing device 1001, the above-mentioned functions defined in the method of the embodiment disclosed in the present application are executed.

[0191] The drone decision-making assistance device provided in this application utilizes the drone decision-making assistance method described in the aforementioned embodiment to address the low efficiency of drone decision-making. Compared to the prior art, the drone decision-making assistance device provided in this application achieves the same beneficial effects as the drone decision-making assistance method described in the aforementioned embodiment. Other technical features of the drone decision-making assistance device are the same as those disclosed in the aforementioned embodiment and are not further elaborated here.

[0192] It should be understood that the various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any one or more embodiments or examples in a suitable manner.

[0193] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

[0194] The present application provides a computer-readable storage medium having computer-readable program instructions (i.e., a computer program) stored thereon, and the computer-readable program instructions are used to execute the drone-assisted decision-making method in the above-mentioned embodiment.

[0195] The computer-readable storage medium provided in this application may be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, systems or devices, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In this embodiment, the computer-readable storage medium may be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, system or device. The program code contained on the computer-readable storage medium may be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination thereof.

[0196] The above-mentioned computer-readable storage medium may be included in the drone decision-making support device; or it may exist independently without being assembled into the drone decision-making support device.

[0197] The computer-readable storage medium carries one or more programs. When the one or more programs are executed by the drone decision-making support device, the drone decision-making support device: detects the data type of the drone sensor in response to the data processing instruction;

[0198] According to the data type of the drone sensor, matching the corresponding capability component, receiving and preprocessing the drone sensor data;

[0199] The drone sensor data is sent to a preset auxiliary decision model to obtain a drone auxiliary decision result, wherein the auxiliary decision model includes a target recognition model based on reinforcement learning, a path planning model based on deep reinforcement learning, and a meteorological activity model based on data fusion. Based on the current drone sensor data, the parameters of the auxiliary decision model are adjusted in real time through a closed-loop feedback module;

[0200] The UAV auxiliary decision result is sent to the main control module, wherein the sending process is carried out through an optimized SPI communication protocol, and the optimized SPI communication protocol includes a dynamic adjustment of the data transmission rate and an automatic retransmission request mechanism.

[0201] Computer program code for performing the operations of the present application may be written in one or more programming languages, or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0202] The flow charts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present application. In this regard, each box in the flow chart or block diagram can represent a module, program segment or a part of code, and the module, program segment or a part of code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order than that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flow chart, and the combination of the boxes in the block diagram and / or flow chart can be implemented by a dedicated hardware-based system that performs the specified function or operation, or can be implemented by a combination of dedicated hardware and computer instructions.

[0203] The modules described in the embodiments of the present application may be implemented in software or hardware, wherein the name of a module does not necessarily limit the unit itself.

[0204] The computer-readable storage medium provided in this application is a computer-readable storage medium that stores computer-readable program instructions (i.e., a computer program) for executing the aforementioned drone-assisted decision-making method, thereby addressing the low efficiency of drone decision-making. Compared to the prior art, the beneficial effects of the computer-readable storage medium provided in this application are the same as those of the drone-assisted decision-making method provided in the aforementioned embodiment, and are not further elaborated here.

[0205] The present application also provides a computer program product, including a computer program, which, when executed by a processor, implements the steps of the above-mentioned drone assisted decision-making method.

[0206] The computer program product provided in this application can solve the problem of low efficiency in drone decision-making. Compared with the prior art, the beneficial effects of the computer program product provided in this application are the same as those of the drone-assisted decision-making method provided in the above embodiment, and will not be repeated here.

[0207] The above description is only part of the embodiments of the present application and does not limit the patent scope of the present application. All equivalent structural transformations made by using the contents of the present application specification and drawings under the technical concept of the present application, or direct / indirect application in other related technical fields are included in the patent protection scope of the present application.

Claims

1. A drone-assisted decision-making method, characterized in that: The method includes: In response to the data processing instruction, detecting a data type of a sensor of the drone; According to the data type of the drone sensor, matching the corresponding capability component, receiving and preprocessing the drone sensor data; The drone sensor data is sent to a preset auxiliary decision model to obtain a drone auxiliary decision result, wherein the auxiliary decision model includes a target recognition model based on reinforcement learning, a path planning model based on deep reinforcement learning, and a meteorological activity model based on data fusion. Based on the current drone sensor data, the parameters of the auxiliary decision model are adjusted in real time through a closed-loop feedback module; The UAV auxiliary decision result is sent to the main control module, wherein the sending process is carried out through an optimized SPI communication protocol, and the optimized SPI communication protocol includes a dynamic adjustment of the data transmission rate and an automatic retransmission request mechanism.

2. The method according to claim 1, wherein The step of matching corresponding capability components according to the data type of the drone sensor, and receiving and pre-processing the drone sensor data includes: According to the data type of the drone sensor, the corresponding component version is searched in the preset component library to determine the corresponding capability component; Based on the capability component, receiving and adaptively decoding drone sensor data; Preprocessing the decoded drone sensor data, wherein the preprocessing includes normalizing and filtering the data.

3. The method according to claim 1, wherein The step of sending the drone sensor data to a preset auxiliary decision model to obtain a drone auxiliary decision result includes: Sending the drone sensor data to a preset target recognition auxiliary decision model; Based on the target recognition auxiliary decision model, a convolutional neural network structure is used to extract feature information from the drone sensor data to form a target feature vector; The extracted target feature vector is input into the fully connected network, and classification and recognition training is performed based on the preset adversarial sample generation algorithm; Based on the training results, human targets, vehicle targets and building targets in the drone sensor data are identified, and the location information, size information and category information of the corresponding targets are generated to determine the target recognition auxiliary decision results.

4. The method according to claim 1, wherein The step of sending the drone sensor data to a preset auxiliary decision model to obtain a drone auxiliary decision result further includes: Sending the drone sensor data to a preset path planning auxiliary decision model; Based on the path planning auxiliary decision model, a topological graph of the current environment is constructed through a multi-scale analysis method, where nodes represent reachable locations and edges represent path connection relationships; Integrate obstacle information into the topological map and construct the path planning objective function based on the preset obstacle avoidance flight requirements of the UAV; According to the path planning objective function, an expansion tree search algorithm is used to search for an optimal path on the topological map, and the optimal path is determined as a path planning auxiliary decision result.

5. The method according to claim 1, wherein The step of sending the drone sensor data to a preset auxiliary decision model to obtain a drone auxiliary decision result further includes: Sending the drone sensor data to a preset meteorological activity auxiliary decision model; parsing meteorological information from drone sensor data based on the meteorological activity decision-making support model; The meteorological information is integrated with preset meteorological forecast data, and an air quality early warning algorithm is used to identify potential risk items in the meteorological information; Based on the potential risk items identified, corresponding flight recommendations and warning information are generated to determine the auxiliary decision-making results of meteorological activities.

6. The method according to claim 1, wherein The step of adjusting the parameters of the auxiliary decision model in real time through a closed-loop feedback module based on the current drone sensor data includes: Monitor the current drone sensor data in real time to determine the current drone status; Optimize the decision by pre-set parameters and adaptively adjust the parameters of the auxiliary decision model according to the current state of the drone; Evaluate the performance of the adaptively adjusted auxiliary decision-making model, and determine whether to update the model parameters based on the evaluation results.

7. A drone decision-making assistance device, characterized in that: The device comprises: a detection module, configured to detect a data type of a sensor of the drone in response to a data processing instruction; A matching module, configured to match corresponding capability components according to the data type of the drone sensor, and receive and pre-process the drone sensor data; A decision-making module is configured to send the drone sensor data to a preset auxiliary decision-making model to obtain a drone auxiliary decision-making result, wherein the auxiliary decision-making model includes a target recognition model based on reinforcement learning, a path planning model based on deep reinforcement learning, and a meteorological activity model based on data fusion. Based on the current drone sensor data, the parameters of the auxiliary decision-making model are adjusted in real time through a closed-loop feedback module; The sending module is used to send the UAV auxiliary decision-making result to the main control module, wherein the sending process is carried out through the optimized SPI communication protocol, and the optimized SPI communication protocol includes dynamically adjusting the data transmission rate and the automatic retransmission request mechanism.

8. A drone decision-making assistance device, characterized in that: The device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program is configured to implement the steps of the drone assisted decision-making method according to any one of claims 1 to 6.

9. A storage medium, characterized in that: The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, the steps of the drone assisted decision-making method according to any one of claims 1 to 6 are implemented.

10. A computer program product, characterized in that The computer program product includes a computer program, and when the computer program is executed by a processor, the steps of the drone assisted decision-making method according to any one of claims 1 to 6 are implemented.

Images