A method and system for identifying unmanned aerial vehicle flight intentions

Abstract

This application provides a method and system for identifying drone flight intentions. The system receives drone flight data and constructs a multidimensional behavior expectation model based on a preset behavioral pattern library and airspace usage rules. A deep belief network is used to map the flight data to the model, calculate the probability distribution of different flight intentions, and generate simulated behavioral performance and probability distribution results. A Markov decision process is used to compare deviation metrics between simulations and actual flight activities, evaluate, and predict optimal flight intentions. Based on the prediction results and the immediate airspace safety status, operational recommendations are generated and sent to the drone, and flight guidance recommendations are generated based on execution feedback information. This application improves the accuracy of drone flight intention identification and the reliability of real-time operational recommendations.

Description

Technical Field

[0001] The embodiments of the present application relate to the technical field of behavior prediction and intelligent management of drones, and in particular to a method and system for identifying drone flight intentions. Background Art

[0002] The widespread application of drones in fields such as logistics and distribution, agricultural monitoring, environmental monitoring, and military reconnaissance has placed higher demands on accurate identification and real-time response to drone flight intentions. To ensure the safe and efficient operation of drones, it is imperative to develop a technology that can accurately analyze drone flight intentions. This technology not only requires processing a large amount of flight parameter and environmental perception data from drones, but also requires spatiotemporal consistency calibration to ensure data accuracy and consistency. Furthermore, the system must combine a pre-defined behavioral pattern library with airspace usage rules to construct a multidimensional behavioral prediction model. Using advanced deep learning methods such as deep belief networks, flight data is mapped to this model, probabilistic distributions of different flight intentions are calculated, and simulated behavioral performance and probability distributions are generated. Based on these predictions, the system also uses a Markov decision process to evaluate and predict the optimal flight intention of the drone. It then generates operational recommendations based on the current airspace safety status, sends operational instructions to the drone, and further optimizes the flight guidance recommendations based on post-execution feedback, thereby achieving closed-loop control and improving the success rate and stability of drone mission execution.

[0003] Existing technical solutions usually use spatiotemporal consistency calibration to process flight parameters and environmental perception data to ensure data consistency and accuracy. Then, a behavioral pattern library is established based on historical flight data or theoretical models, and a multi-dimensional behavioral expectation model is constructed in combination with airspace usage rules. Next, a deep belief network is used to map the flight data into the model, calculate the probability distribution of different flight intentions, and generate simulated behavioral performance and probability distribution results. Through the Markov decision process, the deviation measures between the simulated behavioral performance and the actual flight activities are compared, such as position deviation, speed change, and time difference in response to environmental interactions, to evaluate and predict the flight intention of the drone. Finally, based on the prediction results and the real-time airspace safety status, operation recommendations are generated, instructions are sent to the drone, and flight guidance recommendations are generated based on the feedback information after the drone executes to optimize subsequent operations.

[0004] While existing solutions address the issue of drone flight intention identification to a certain extent, they still suffer from several key flaws. First, there is a lack of detailed description of the specific methods, algorithms, or standards for data calibration and preprocessing, resulting in limited operability in practical applications. Second, the mechanism for establishing and updating the behavioral pattern library is unclear, making it unable to dynamically adapt to new data and changing environmental conditions, affecting the system's adaptability and long-term accuracy. Furthermore, the application details of deep belief networks are missing, including key information such as the specific architecture, loss function selection during training, and optimizer configuration, making it difficult to replicate and optimize the method. Furthermore, the selection of deviation metrics and the quantification of their relative importance are not detailed. The importance of these factors may vary in different mission scenarios, affecting prediction accuracy. Finally, the lack of detailed description of core elements of the Markov decision process, such as the state space definition, transition probability calculation, and reward function design, makes it difficult to fully understand and effectively implement the method. Summary of the Invention

[0005] The embodiments of the present application provide a method and system for identifying drone flight intentions, which are used to solve the problems of low accuracy in drone flight intention identification and poor reliability of real-time operation suggestions in the prior art.

[0006] In a first aspect, an embodiment of the present application provides a method for identifying drone flight intentions, comprising:

[0007] Receiving flight data from the UAV, wherein the flight data includes flight parameters and environmental perception data that have been calibrated for spatiotemporal consistency;

[0008] Using the flight data, a multi-dimensional behavior prediction model of the UAV is constructed based on a preset behavior pattern library and airspace usage rules;

[0009] Using a deep belief network, the flight data is mapped to the multi-dimensional behavior expectation model to calculate the probability distribution of the drone under different flight intentions, and generate simulated behavior performance and corresponding probability distribution results;

[0010] Based on the probability distribution results, a Markov decision process is used to compare the deviation metrics between the simulated behavior performance and the actual flight activity to evaluate and predict the UAV's flight intention and obtain an optimal flight intention prediction, wherein the deviation metrics include position deviation, speed change, and time difference in response to environmental interactions;

[0011] Based on the flight intention and the immediate airspace safety status, an operation suggestion for the drone is generated, and the operation suggestion is sent to the drone. Feedback information after the drone executes the operation is obtained, and flight guidance suggestions are generated based on the feedback information.

[0012] Optionally, the deep belief network is used to map the flight data to the multi-dimensional behavior expectation model to calculate the probability distribution of the UAV under different flight intentions, and generate simulated behavior performance and corresponding probability distribution results, including:

[0013] Using a deep belief network, performing nonlinear transformation processing on the flight data, and dynamically adjusting the weight of the flight data using an attention mechanism to obtain a flight data representation;

[0014] mapping the flight data representation to different portions of the multi-dimensional behavior expectation model to identify behavior patterns that match the flight data representation;

[0015] Based on the behavior pattern, a flight intention model of the UAV in the multi-dimensional behavior expectation model is constructed, and a similarity metric learning method is applied to evaluate the similarity between the flight intention models to determine the optimal parameter configuration of the flight intention model and generate a behavior pattern configuration;

[0016] Based on the behavior pattern configuration and the Bayesian reasoning framework, the probability distribution of the drone executing each flight intention is calculated, and the variational inference technique is used to obtain the probability distribution result;

[0017] Based on the probability distribution results and the behavior pattern configuration, a Monte Carlo method is used to generate a series of simulated behavior performances and corresponding probability distribution results.

[0018] Optionally, based on the behavior pattern, constructing a flight intention model of the UAV in the multi-dimensional behavior expectation model, and applying a similarity metric learning method to evaluate the similarity between the flight intention models to determine the optimal parameter configuration of the flight intention model and generate a behavior pattern configuration, including:

[0019] Mapping the flight intention of the behavior pattern to different parts of the multi-dimensional behavior expectation model to construct multiple flight intention models;

[0020] Using the Bayesian optimization method, the optimal initial parameter combination is selected from the flight behavior patterns in the pre-training dataset, and the Gaussian process is used to initialize the parameter configuration of the flight intention model in combination with the optimal initial parameter combination;

[0021] Utilizing multiple distance measurement methods and kernel methods in combination with twin network learning, the similarity between the flight data representation and the flight intention model is calculated, and the degree of matching between the flight data representation and the flight intention model is quantified to generate a similarity measurement result;

[0022] Determining an optimal parameter configuration from the parameters of the flight intention model by applying a gradient descent algorithm and a reinforcement learning mechanism based on the similarity measurement result and the initialization parameter configuration;

[0023] Based on the optimal parameter configuration, a behavior pattern configuration is generated using a context-aware reasoning mechanism.

[0024] Optionally, the determining the optimal parameter configuration from the parameters of the flight intention model by applying a gradient descent algorithm and a reinforcement learning mechanism based on the similarity measurement result in combination with the initialization parameter configuration includes:

[0025] According to the similarity measurement result and the initialization parameter configuration, an initial loss function is set, and a gradient descent algorithm is applied to obtain initial parameters of the flight intention model by calculating the gradient of the initial loss function with respect to the initialization parameter configuration and adopting an adaptive learning rate strategy;

[0026] Using a reinforcement learning mechanism, a reward function is defined, and based on the reward function and actual flight data feedback, initial parameters of the flight intention model are dynamically adjusted to generate an intermediate parameter configuration;

[0027] Based on the flight data, the intermediate parameter configuration is evaluated to obtain a verification result, and the weights in the loss function and the reward function are adjusted according to the verification result to generate a closed-loop feedback mechanism. Based on the closed-loop feedback mechanism, the optimal parameter configuration is determined.

[0028] Optionally, based on the probability distribution result, comparing the deviation metric between the simulated behavior performance and the actual flight activity through a Markov decision process to evaluate and predict the flight intention of the UAV and obtain the optimal flight intention prediction includes:

[0029] Based on the probability distribution result, a state transition model is constructed; according to the state transition model, in combination with a preset behavior pattern library and airspace usage rules, a Markov decision process model is initialized, a reward function is defined, and an initial Markov decision process configuration is generated based on the reward function;

[0030] By comparing the deviation measures between the simulated behavioral performance and the actual flight activities, a series of deviation measure results are obtained;

[0031] Based on the deviation measurement result and the initial Markov decision process configuration, applying a dynamic time warping algorithm to evaluate the match between the simulated behavior performance and the actual flight activity, and generating a match evaluation result;

[0032] According to the matching evaluation result, the flight intention corresponding to the simulated behavior performance of the actual flight activity is selected as the prediction result to obtain the optimal flight intention prediction.

[0033] Optionally, applying a dynamic time warping algorithm based on the deviation measurement result and the initial Markov decision process configuration to evaluate the match between the simulated behavior performance and the actual flight activity and generating a match evaluation result includes:

[0034] Preprocessing the simulated behavior performance according to the deviation measurement result and in combination with the initial Markov decision process configuration to obtain a preprocessed simulated behavior performance;

[0035] Determining parameter configuration of a dynamic time warping algorithm based on the characteristics of the simulated behavior performance after the preprocessing and the deviation measurement result;

[0036] Calculating the time series matching between the simulated behavior performance and the actual flight activity using the parameter configuration of the dynamic time warping algorithm to obtain a time series matching result;

[0037] quantifying the similarity between the simulated behavior performance and the actual flight activity based on the time series matching result to generate a matching score;

[0038] Based on the matching score combined with the Markov decision process configuration, the matching of the simulated behavior performance is evaluated to generate a matching evaluation result.

[0039] Optionally, the use of the flight data to construct a multi-dimensional behavior prediction model for the UAV according to a preset behavior pattern library and airspace usage rules includes:

[0040] Performing feature extraction processing on the flight data to obtain a set of key feature sets;

[0041] According to a preset behavior pattern library, the key feature set is compared with the flight behavior patterns in the preset behavior pattern library, the pattern related to the UAV flight behavior is identified, and a behavior pattern result is generated;

[0042] Using the behavior pattern results, applying airspace use rules, evaluating whether the UAV flight behavior complies with the predicted aviation management regulations and safety requirements, and obtaining a flight behavior pattern;

[0043] Based on the flight behavior pattern, a multi-dimensional behavior expectation model is constructed.

[0044] In a second aspect, an embodiment of the present application provides a drone flight intention recognition system, comprising:

[0045] A receiving module, configured to receive flight data from the UAV, wherein the flight data includes flight parameters and environmental perception data that have been calibrated for spatiotemporal consistency;

[0046] A construction module is used to use the flight data to build a multi-dimensional behavior expectation model of the UAV based on a preset behavior pattern library and airspace usage rules;

[0047] a computing module, configured to map the flight data to the multi-dimensional behavior expectation model using a deep belief network to calculate the probability distribution of the UAV under different flight intentions and generate simulated behavior performance and corresponding probability distribution results;

[0048] a comparison module, configured to compare, based on the probability distribution result, a deviation metric between the simulated behavior performance and the actual flight activity through a Markov decision process to evaluate and predict the flight intention of the UAV and obtain an optimal flight intention prediction, wherein the deviation metric includes position deviation, speed change, and time difference in response to environmental interaction;

[0049] The generation module is used to generate operation suggestions for the drone based on the flight intention and the immediate airspace safety status, send the operation suggestions to the drone, obtain feedback information after the drone executes the operation, and generate flight guidance suggestions based on the feedback information.

[0050] In a third aspect, an embodiment of the present application provides a computing device comprising a processor and a memory, wherein the memory stores a computer program, and the processor is configured to run the computer program to execute a method for identifying drone flight intentions as described in any one of the first aspects.

[0051] In a fourth aspect, an embodiment of the present application provides a computer storage medium having computer program instructions stored thereon, wherein the computer program instructions, when executed by a processor, implement a method for identifying the flight intention of a drone as described in any one of the first aspects.

[0052] In an embodiment of the present application, flight data from a drone is received, wherein the flight data includes flight parameters and environmental perception data that have been calibrated for spatiotemporal consistency; using the flight data, a multidimensional behavior expectation model for the drone is constructed based on a preset behavior pattern library and airspace usage rules; using a deep belief network, the flight data is mapped to the multidimensional behavior expectation model to calculate the probability distribution of the drone under different flight intentions, and simulated behavior performance and corresponding probability distribution results are generated; based on the probability distribution results, a deviation metric between the simulated behavior performance and the actual flight activity is compared through a Markov decision process to evaluate and predict the flight intention of the drone, and obtain an optimal flight intention prediction, wherein the deviation metric includes position deviation, speed change, and time difference in response to environmental interaction; based on the flight intention and in combination with the immediate airspace safety status, an operation suggestion for the drone is generated, and the operation suggestion is sent to the drone to obtain feedback information after the drone executes the suggestion, and based on the feedback information, a flight guidance suggestion is generated.

[0053] The technical solution of this application has the following beneficial effects:

[0054] By integrating advanced technologies such as deep learning, behavioral modeling, and Markov decision processes, this application not only improves the accuracy of drone flight intention recognition but also provides timely and effective operational recommendations in dynamically changing airspace environments, significantly enhancing drone safety and operational efficiency, and providing a solid technical foundation for large-scale drone deployment. Furthermore, a feedback mechanism continuously optimizes flight guidance strategies, further ensuring the success rate and stability of drone mission execution.

[0055] Furthermore, a deep belief network is used to perform nonlinear transformations on the flight data, and an attention mechanism is employed to dynamically adjust the weights of the flight data to obtain a flight data representation. This representation is then mapped to different components of a multidimensional behavior expectation model to identify matching behavioral patterns. Based on these behavioral patterns, a flight intention model for the drone is constructed. Similarity metric learning methods are applied to evaluate the similarity between different flight intention models, determine the optimal parameter configuration, and generate a behavioral pattern configuration. Next, a Bayesian inference framework is used to calculate the probability distribution of each flight intention, and variational inference techniques are used to optimize the resulting probability distribution. Finally, based on the resulting probability distribution and behavioral pattern configuration, a Monte Carlo method is used to generate a series of simulated behavioral performances and their corresponding probability distributions.

[0056] Through the aforementioned methods, deep belief networks and attention mechanisms enhance the nonlinear representation of flight data, ensuring the effectiveness and flexibility of feature extraction. A multidimensional behavior expectation model and similarity metric learning methods enable accurate identification and prediction of drone flight intentions, improving both prediction accuracy and reliability. Combining a Bayesian inference framework with variational inference techniques further optimizes the probability distribution calculation process, enabling the system to rapidly respond to changes in complex environments. Ultimately, the simulated behavioral performance generated using the Monte Carlo method not only improves the system's adaptability and robustness but also provides more precise operational recommendations for the drone, significantly enhancing its efficiency and safety in actual missions. This suite of technical approaches works together to achieve highly automated and intelligent drone flight intention recognition, effectively ensuring its safe operation and successful mission completion.

[0057] These and other aspects of the present application will become more readily apparent from the description of the following embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0059] Figure 1 A flowchart of a method for identifying UAV flight intentions provided in an embodiment of the present application;

[0060] Figure 2 A schematic diagram of the structure of a drone flight intention recognition system provided in an embodiment of the present application;

[0061] Figure 3 A schematic diagram of the structure of a computing device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0062] In order to enable people skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0063] In some of the processes described in the specification and claims of this application and the above-mentioned figures, multiple operations that appear in a specific order are included, but it should be clearly understood that these operations may not be executed in the order in which they appear in this document or may be executed in parallel. The serial numbers of the operations, such as 101, 102, etc., are only used to distinguish between different operations, and the serial numbers themselves do not represent any order of execution. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this document are used to distinguish different messages, devices, modules, etc., and do not represent a sequential order, nor do they limit "first" and "second" to being different types.

[0064] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without making creative efforts are within the scope of protection of this application.

[0065] Figure 1 A flowchart of a method for identifying the flight intention of a drone is provided in an embodiment of the present application. Figure 1 As shown, the method includes:

[0066] Step 101: receiving flight data from a UAV, wherein the flight data includes flight parameters and environmental perception data that have been calibrated for spatiotemporal consistency;

[0067] In this step, flight data, including flight parameters (such as position, speed, and heading) and environmental perception data (such as weather conditions and obstacle information), undergoes spatiotemporal consistency calibration. This process ensures the temporal and spatial synchronization of these data, providing a reliable foundation for subsequent analysis. This step is a prerequisite for the operation of the entire system and ensures the quality and accuracy of the input data.

[0068] In practice, drones first collect flight parameters and environmental perception data through their various onboard sensors. This raw data is then calibrated for temporal and spatial consistency using a specific algorithm to eliminate inconsistencies caused by time delays or spatial variations. The calibrated data is then sent to a ground station or cloud server for further analysis and processing.

[0069] For example, in an urban logistics delivery project, drones are used to deliver packages from warehouses to customers. After takeoff, the drone continuously transmits a data stream containing its current position, flight speed, and surrounding environmental information (such as wind speed and building distribution). This data is calibrated for temporal and spatial consistency, ensuring high accuracy even in dynamically changing urban environments, providing a solid foundation for subsequent path planning.

[0070] Step 102: Using the flight data, a multi-dimensional behavior prediction model of the UAV is constructed according to a preset behavior pattern library and airspace usage rules;

[0071] In this step, the multi-dimensional behavior prediction model is a behavior prediction framework that comprehensively considers multiple factors, such as drone performance, mission requirements, and airspace regulations. Based on a library of preset behavior patterns and real-time flight data, it aims to simulate the drone's expected behavior in different scenarios and help identify and predict possible flight intentions.

[0072] In actual operation, the system first loads a preset behavioral model library, which contains templates for different types of mission behaviors (such as cruising, hovering, and obstacle avoidance). Then, based on currently received flight data and airspace usage rules, it adjusts model parameters to generate a multi-dimensional behavioral prediction model that reflects the current mission requirements. This step provides a theoretical basis and reference framework for subsequent flight intent recognition.

[0073] For example, continuing with the logistics delivery example above, when a drone begins a delivery mission, the system builds a multi-dimensional behavioral prediction model tailored to the mission, based on factors such as the distance between the drone's origin and destination, the estimated flight time, and real-time traffic conditions and weather forecasts. This model not only considers the drone's basic operational characteristics but also incorporates external environmental constraints, such as no-fly zones and inclement weather, to ensure the mission proceeds safely and smoothly.

[0074] Step 103: Using a deep belief network, the flight data is mapped to the multi-dimensional behavior expectation model to calculate the probability distribution of the UAV under different flight intentions, and generate simulated behavior performance and corresponding probability distribution results;

[0075] In this step, a Deep Belief Network (DBN) is used to map the calibrated flight data into a multi-dimensional behavior prediction model. This calculates the probability distribution of the drone under different flight intentions and generates corresponding simulated behavioral performance and its probability distribution. This process helps evaluate different possible action paths that the drone may take and their likelihood.

[0076] In practice, the DBN first performs a nonlinear transformation on the flight data and dynamically adjusts the weights using an attention mechanism to generate a representation of the flight data. This representation is then mapped to different components of a multidimensional behavior expectation model to identify matching behavioral patterns. Based on these patterns, the system constructs a model of the drone's flight intent and applies similarity metric learning to determine the optimal parameter configuration and generate a behavioral pattern configuration.

[0077] For example, during logistics delivery, as a drone approaches its target location, the system uses DBN to analyze its current flight state and environmental conditions, predicting the drone's likely actions, such as slowing down for landing or maneuvering around temporary obstacles. By generating multiple simulated behaviors and their probability distributions, the system can plan the optimal landing path in advance, optimizing mission execution efficiency.

[0078] Step 104: Based on the probability distribution results, the deviation metrics between the simulated behavior performance and the actual flight activity are compared through a Markov decision process to evaluate and predict the UAV's flight intention and obtain an optimal flight intention prediction, wherein the deviation metrics include position deviation, speed change, and time difference in response to environmental interactions;

[0079] In this step, based on the probability distribution results generated in the previous step, the Markov Decision Process (MDP) is used to compare the deviation metrics (such as position deviation, speed change, and response time difference) between the simulated behavior performance and the actual flight activity, evaluate and predict the UAV's flight intention, and obtain the optimal flight intention prediction.

[0080] In practice, the system uses the MDP to evaluate the deviation between each simulated behavior and actual flight activity, selecting the behavior with the smallest deviation as the optimal flight intention prediction. This process not only considers changes in physical distance and speed, but also incorporates the drone's response time to environmental changes, ensuring the comprehensiveness and accuracy of the prediction results.

[0081] For example, in a logistics delivery mission, as a drone approaches its target location, the system uses MDP to quickly evaluate the effectiveness of various response strategies, such as adjusting flight altitude to avoid sudden obstacles or increasing speed to cope with impending inclement weather. Ultimately, the system selects the most appropriate flight intent, enabling the drone to complete its delivery mission safely and efficiently.

[0082] Step 105: Based on the flight intention and the immediate airspace safety status, generate operation suggestions for the drone, send the operation suggestions to the drone, obtain feedback information after the drone executes the operation, and generate flight guidance suggestions based on the feedback information.

[0083] In this step, based on the predicted flight intent and the current airspace safety status, operational recommendations are generated for the drone and sent to the drone. The system also receives feedback from the drone after execution, using this information to generate further flight guidance recommendations and optimize subsequent operations.

[0084] In actual operation, the system comprehensively considers the drone's flight intentions and the current airspace safety status to formulate detailed operational recommendations, such as adjusting the flight path to avoid obstacles or enter a new monitoring area. After the recommendations are sent to the drone, the system monitors the implementation and continuously optimizes the flight guidance based on the feedback.

[0085] For example, as a logistics delivery mission neared its end, the system developed a detailed return plan based on the drone's flight intent and remaining mission load. Furthermore, considering an impending storm, the system recommended that the drone expedite its return to base. After the drone executed the command, the system adjusted subsequent mission schedules based on its feedback, ensuring the smooth completion of the entire project and accumulating valuable experience for future similar missions.

[0086] Through the above five steps, this method significantly improves the accuracy and real-time response capabilities of drone flight intention recognition. Specifically, spatiotemporal consistency calibration ensures the accuracy and consistency of input data; a multidimensional behavior expectation model enhances the system's adaptability and flexibility; the application of deep belief networks and Markov decision processes improves the reliability of flight intention prediction; and a closed-loop feedback mechanism optimizes the effectiveness of operational recommendations. Overall, this process not only ensures the safe operation of drones, but also greatly improves the success rate and efficiency of their mission execution, achieving efficient automated management in complex and changing environments.

[0087] To further improve the accuracy and adaptability of drone flight intention recognition, in some embodiments, the deep belief network used in step 103 maps the flight data to the multi-dimensional behavior expectation model to calculate the probability distribution of the drone under different flight intentions, generate simulated behavior performance and corresponding probability distribution results, including:

[0088] The flight data is nonlinearly transformed using a deep belief network, and the weight of the flight data is dynamically adjusted using an attention mechanism to obtain a flight data representation; the flight data representation is mapped to different parts of the multidimensional behavior expectation model to identify behavior patterns that match the flight data representation; based on the behavior pattern, a flight intention model of the UAV in the multidimensional behavior expectation model is constructed, and a similarity metric learning method is applied to evaluate the similarity between the flight intention models to determine the optimal parameter configuration of the flight intention model and generate a behavior pattern configuration; based on the behavior pattern configuration combined with a Bayesian reasoning framework, the probability distribution of the UAV executing each flight intention is calculated, and variational inference technology is used to obtain a probability distribution result; based on the probability distribution result combined with the behavior pattern configuration, a Monte Carlo method is used to generate a series of simulated behavior performances and corresponding probability distribution results.

[0089] In this embodiment, a DBN is used to perform nonlinear transformations on flight data. By introducing an attention mechanism, the weight of each feature is dynamically adjusted to achieve a more effective representation of flight data. The multidimensional behavior prediction model is a behavior prediction framework that comprehensively considers multiple factors (such as drone performance, mission requirements, and airspace regulations). It can identify behavioral patterns that match the current flight data representation. Based on these patterns, the system constructs a model of the drone's flight intention and uses similarity metric learning to determine the optimal parameter configuration, ultimately generating a behavioral pattern configuration. Combining a Bayesian inference framework with variational inference techniques further optimizes the probability distribution calculation process, ensuring the accuracy and reliability of predictions.

[0090] In an embodiment of the present application, first, the system uses DBN to perform nonlinear transformation processing on the flight data, and dynamically adjusts the weights through the attention mechanism to obtain a representation of the flight data. Then, the representation is mapped to different parts of the multi-dimensional behavior expectation model to identify the behavior patterns that match it. Based on these behavior patterns, the system constructs a flight intention model of the UAV, and applies a similarity metric learning method to evaluate the similarity between different flight intention models, determine the optimal parameter configuration, and generate a behavior pattern configuration. Next, the probability distribution of each flight intention is calculated in combination with the Bayesian reasoning framework, and the variational inference technique is used to optimize the probability distribution results. Finally, based on the obtained probability distribution results and behavior pattern configuration, the Monte Carlo method is used to generate a series of simulated behavior performances and their corresponding probability distribution results to provide more accurate operational recommendations.

[0091] Here's a concrete example:

[0092] In a city emergency response project, drones were deployed to quickly assess fire scene conditions and guide rescue operations. In step 103, the drone first used its sensors to collect various data from the fire scene, including temperature, smoke concentration, and obstacle locations. The system then used a DBN to perform nonlinear transformations on this raw data and dynamically adjusted the weight of each feature through an attention mechanism to obtain a more accurate representation of the flight data. The system then mapped this flight data representation to a pre-built multidimensional behavior expectation model, identifying behavioral patterns that matched the current environmental conditions, such as bypassing high-temperature areas or avoiding areas of dense smoke. Based on these behavioral patterns, the system constructed a flight intention model specifically for this emergency response mission. Using similarity metric learning methods, the system evaluated the similarity between different flight intention models, determined the optimal parameter configuration, and generated a behavioral pattern configuration.

[0093] The system then combined a Bayesian inference framework to calculate the probability distribution of each flight intention and used variational inference techniques to optimize the resulting probability distributions. Based on these probability distributions, the system employed Monte Carlo methods to generate a series of simulated behavioral performances and their corresponding probability distributions, providing multiple possible flight paths for selection. In actual operation, the system selected the safest and most efficient flight path, enabling the drone to quickly reach the core of the fire, obtain critical information, and provide real-time feedback to the ground command center. This approach not only improved the drone's mission execution efficiency but also significantly enhanced the overall effectiveness of the emergency response, ensuring the safety of people and property.

[0094] In order to further improve the accuracy of UAV flight intention recognition and the effectiveness of parameter optimization, in some embodiments, step 103 includes constructing a UAV flight intention model in the multi-dimensional behavior expectation model based on the behavior pattern, and applying a similarity metric learning method to evaluate the similarity between the flight intention models to determine the optimal parameter configuration of the flight intention model and generate a behavior pattern configuration, further comprising:

[0095] The flight intention of the behavior pattern is mapped to different parts of the multi-dimensional behavior expectation model to construct multiple flight intention models; the Bayesian optimization method is used to select the best initial parameter combination from the flight behavior pattern in the pre-training data set, and the Gaussian process is used in combination with the best initial parameter combination to initialize the parameter configuration of the flight intention model; a variety of distance measurement methods and kernel methods are used in combination with twin network learning to calculate the similarity between the flight data representation and the flight intention model, and the degree of matching between the flight data representation and the flight intention model is quantified to generate a similarity measurement result; based on the similarity measurement result and the initialization parameter configuration, the gradient descent algorithm and the reinforcement learning mechanism are applied to determine the optimal parameter configuration from the parameters of the flight intention model; based on the optimal parameter configuration, the situational awareness reasoning mechanism is used to generate the behavior pattern configuration. Optionally, the similarity measurement result is combined with the initialization parameter configuration, and the gradient descent algorithm and the reinforcement learning mechanism are applied to determine the optimal parameter configuration from the parameters of the flight intention model, including: setting an initial loss function according to the similarity measurement result and the initialization parameter configuration, and applying the gradient descent algorithm, by calculating the gradient of the initial loss function with respect to the initialization parameter configuration, adopting an adaptive learning rate strategy, to obtain the initial parameters of the flight intention model; using the reinforcement learning mechanism, defining a reward function, and dynamically adjusting the initial parameters of the flight intention model based on the reward function combined with actual flight data feedback to generate an intermediate parameter configuration; evaluating the intermediate parameter configuration according to the flight data to obtain a verification result, and adjusting the weights in the loss function and the reward function according to the verification result to generate a closed-loop feedback mechanism, and determining the optimal parameter configuration based on the closed-loop feedback mechanism.

[0096] In this embodiment, multiple flight intention models are constructed by mapping flight intention to different parts of a multi-dimensional behavior expectation model. These models are initialized with parameter configurations using the Bayesian optimization method and the Gaussian process, ensuring that the starting point of the model has a high degree of accuracy. Subsequently, a variety of distance measurement methods and kernel methods are combined with twin network learning to calculate the similarity between the flight data representation and the flight intention model, and to quantify the degree of matching. Based on the similarity measurement results and the initialization parameter configuration, the gradient descent method and the reinforcement learning mechanism are used to determine the optimal parameter configuration from the parameters of the flight intention model, and finally the situational awareness reasoning mechanism is used to generate the behavior pattern configuration. This method not only improves the accuracy of flight intention recognition, but also enhances the effectiveness of parameter optimization.

[0097] In an embodiment of the present application, first, the system maps the flight intention of the behavior pattern to different parts of the multi-dimensional behavior expectation model to construct multiple flight intention models. Then, the Bayesian optimization method is used to select the best initial parameter combination from the pre-training data set, and the Gaussian process is used to initialize the parameter configuration of the flight intention model. Next, a variety of distance measurement methods and kernel methods are combined with twin network learning to calculate the similarity between the flight data representation and the flight intention model to generate a similarity measurement result. Based on the similarity measurement result and the initialization parameter configuration, the gradient descent algorithm and reinforcement learning mechanism are applied to determine the optimal parameter configuration. Finally, based on the optimal parameter configuration, the situational awareness reasoning mechanism is used to generate the behavior pattern configuration.

[0098] Here's a concrete example:

[0099] In an urban firefighting and rescue project, drones were used to monitor fire scenes in real time and guide rescue operations. In step 103, the drone's sensors first collected various fire scene data, including temperature, smoke concentration, and obstacle locations. The system then fed this raw data into a deep belief network, which dynamically adjusted weights through nonlinear transformation processing and an attention mechanism to generate a flight data representation.

[0100] The system then maps the flight data representation to different components of a multidimensional behavioral expectation model, constructing multiple flight intent models, such as those for avoiding high-temperature areas, avoiding areas of dense smoke, or quickly reaching a designated location. Using Bayesian optimization, the optimal initial parameter combination is selected from the pretrained dataset, and the flight intent model parameters are initialized using a Gaussian process. Subsequently, the system utilizes various distance metrics and kernel methods combined with twin network learning to calculate the similarity between the flight data representation and the flight intent model, quantify the degree of match, and generate a similarity metric. Based on the similarity metric and the initial parameter configuration, a gradient descent algorithm and reinforcement learning are applied to determine the optimal parameter configuration. This process first sets an initial loss function and applies a gradient descent algorithm. By calculating the gradient of the initial loss function with respect to the initial parameter configuration, an adaptive learning rate strategy is employed to determine the initial parameters of the flight intent model. Then, a reward function is defined using reinforcement learning. Based on this reward function and actual flight data feedback, the initial parameters of the flight intent model are dynamically adjusted to generate an intermediate parameter configuration.

[0101] The system evaluates intermediate parameter configurations based on flight data, generates verification results, and adjusts the weights in the loss and reward functions based on these results, creating a closed-loop feedback mechanism. Based on this closed-loop feedback mechanism, the system ultimately determines the optimal parameter configuration and utilizes a context-aware reasoning mechanism to generate a behavioral pattern configuration. In actual operation, the system selected the safest and most efficient flight path, enabling the drone to quickly reach the core of the fire, obtain critical information, and provide real-time feedback to the ground command center. This approach not only improves the drone's mission execution efficiency but also significantly enhances the overall effectiveness of the emergency response, ensuring the safety of people and property.

[0102] To further improve the accuracy of the UAV flight intention prediction and the effectiveness of the matching evaluation, in some embodiments, the step 104 includes comparing the deviation measure between the simulated behavior performance and the actual flight activity based on the probability distribution result through a Markov decision process to evaluate and predict the UAV flight intention and obtain the optimal flight intention prediction, including:

[0103] Based on the probability distribution result, a state transition model is constructed. According to the state transition model, in combination with a preset behavior pattern library and airspace usage rules, a Markov decision process model is initialized, a reward function is defined, and based on the reward function, an initial Markov decision process configuration is generated; by comparing the deviation measurement between the simulated behavior performance and the actual flight activity, a series of deviation measurement results are obtained; based on the deviation measurement results and the initial Markov decision process configuration, a dynamic time warping algorithm is applied to evaluate the matching degree between the simulated behavior performance and the actual flight activity, and a matching degree evaluation result is generated; based on the matching degree evaluation result, the flight intention corresponding to the simulated behavior performance of the actual flight activity is selected as the prediction result, and the optimal flight intention prediction is obtained. Optionally, based on the deviation measurement result and the initial Markov decision process configuration, a dynamic time warping algorithm is applied to evaluate the matching degree between the simulated behavior performance and the actual flight activity, and generate a matching degree evaluation result, including: preprocessing the simulated behavior performance according to the deviation measurement result and the initial Markov decision process configuration to obtain the preprocessed simulated behavior performance; determining the parameter configuration of the dynamic time warping algorithm based on the characteristics of the preprocessed simulated behavior performance and the deviation measurement result; calculating the time series matching degree between the simulated behavior performance and the actual flight activity using the parameter configuration of the dynamic time warping algorithm to obtain a time series matching degree result; quantifying the similarity between the simulated behavior performance and the actual flight activity according to the time series matching degree result to generate a matching degree score; evaluating the matching degree of the simulated behavior performance based on the matching degree score and the Markov decision process configuration to generate a matching degree evaluation result.

[0104] In this embodiment, a state transition model is constructed based on probability distribution results and initialized using a pre-defined behavioral pattern library and airspace usage rules. The MDP model defines a reward function to guide flight intention prediction. Subsequently, a series of deviation measurement results are generated by comparing the deviation metrics between simulated behavioral performance and actual flight activities. Using the Dynamic Time Warping (DTW) algorithm, the system accurately assesses the degree of match between simulated behavioral performance and actual flight activities, ultimately generating a match evaluation result and selecting the flight intention that best matches the actual flight activities as the predicted result.

[0105] In an embodiment of the present application, first, a state transition model is constructed based on the probability distribution results, and the Markov decision process model is initialized in combination with the preset behavior pattern library and airspace usage rules. After the reward function is defined, an initial MDP configuration is generated. Next, the system compares the deviation measurement between the simulated behavior performance and the actual flight activity to obtain a series of deviation measurement results. Then, based on these deviation measurement results and the initial MDP configuration, the dynamic time warping algorithm is applied to pre-process the simulated behavior performance, and the DTW parameter configuration is determined according to its characteristics. Next, the DTW algorithm is used to calculate the time series matching between the simulated behavior performance and the actual flight activity to obtain the time series matching result. Finally, the degree of similarity between the two is quantified according to the time series matching result to generate a matching score, and based on this score combined with the MDP configuration to evaluate the matching of the simulated behavior performance to generate a final matching evaluation result.

[0106] Here's a concrete example:

[0107] In an urban logistics delivery project, drones are used to deliver packages from warehouses to customers. In step 104, the drones collect a large amount of flight data, including location, speed, and heading. The system first constructs a state transition model based on the probability distribution generated in the previous step. It then initializes a Markov decision process (MDP) model using a pre-defined behavioral pattern library and airspace usage rules. A reward function is defined, such as a positive reward for successfully completing a task and a negative reward for encountering an obstacle or straying from the path, to generate an initial MDP configuration.

[0108] The system then compares the simulated performance with actual flight activity by measuring deviations (e.g., position deviation, velocity change, and response time difference), generating a series of deviation metrics. Based on these deviation metrics and the initial MDP configuration, the system preprocesses the simulated performance using the Dynamic Time Warping (DTW) algorithm to ensure data consistency and accuracy. Based on the preprocessed simulated performance and its characteristics, the system determines the DTW algorithm's parameter configuration.

[0109] Then, using these parameter configurations, the system calculated the time series matching between the simulated behavior and the actual flight activities, generating a time series matching result. Based on these matching results, the system quantified the similarity between the simulated behavior and the actual flight activities, generating a matching score. Based on the matching score combined with the MDP configuration, the system evaluated the matching of the simulated behavior and generated a final matching evaluation result. In actual operation, the system selected the flight intention corresponding to the simulated behavior that best matched the actual flight activities as the prediction result, thereby optimizing the drone's mission execution path. This not only improves the drone's mission execution efficiency, but also significantly enhances its ability to cope with complex environments, ensuring the safe and efficient delivery of packages. In this way, the entire system achieves more accurate flight intention predictions, improving the overall performance and safety of drone operations.

[0110] To further improve the accuracy of drone flight behavior recognition and the effectiveness of safety assessment, in some embodiments, step 102 utilizes the flight data to construct a multi-dimensional drone behavior prediction model based on a preset behavior pattern library and airspace usage rules, including:

[0111] The flight data is subjected to feature extraction processing to obtain a set of key feature sets; based on a preset behavior pattern library, the key feature set is compared with the flight behavior patterns in the preset behavior pattern library to identify patterns related to the UAV flight behavior and generate behavior pattern results; using the behavior pattern results, airspace usage rules are applied to evaluate whether the UAV flight behavior complies with predicted aviation management regulations and safety requirements to obtain a flight behavior pattern; based on the flight behavior pattern, a multi-dimensional behavior expectation model is constructed.

[0112] In this embodiment, feature extraction processing extracts a set of key features from raw flight data. These features may include position, speed, and heading. These key features are then compared with flight behavior patterns in a pre-defined behavioral pattern library to identify the pattern that best matches the current drone's flight behavior. Based on these behavioral pattern results and in conjunction with airspace usage rules (such as no-fly zones and route restrictions), the system assesses whether the drone's flight behavior complies with aviation regulations and safety requirements. Ultimately, a multidimensional behavioral prediction model is constructed based on the evaluation results, which is then used to predict flight intentions and generate operational recommendations.

[0113] In the embodiment of the present application, the system first performs feature extraction on the flight data, extracting a set of key features. Next, it compares these key features with flight behavior patterns in a pre-set behavior pattern library, identifying patterns related to drone flight behavior and generating behavior pattern results. Subsequently, using these behavior pattern results, the system applies airspace usage rules to assess whether the drone's flight behavior complies with pre-determined aviation regulations and safety requirements, thereby generating flight behavior patterns. Finally, based on these flight behavior patterns, the system constructs a multi-dimensional behavior prediction model, providing a foundation for subsequent flight intent identification and operational recommendations.

[0114] Here's a concrete example:

[0115] In an urban logistics delivery project, drones are used to deliver packages from warehouses to customers. In step 102, the drone first collects a large amount of flight data, including location, speed, and heading. The system then performs feature extraction on this flight data, extracting a set of key features, such as the distance from the current location to the target point, the current flight speed, and the location of surrounding obstacles.

[0116] Next, the system compares these key characteristics with flight behavior patterns in a pre-set behavioral pattern library. This library contains a variety of typical flight behavior patterns, such as cruising, hovering, and obstacle avoidance. Through this comparison, the system identifies the pattern that best matches the current drone's flight behavior, such as the behavior of a drone slowing down and preparing to land when approaching a target point, and generates a behavioral pattern result. The system then uses these behavioral pattern results, combined with airspace usage rules, to assess whether the drone's flight behavior complies with anticipated aviation management regulations and safety requirements. For example, the system checks whether the drone has entered a no-fly zone or is following a prescribed route. If any non-compliant behavior is detected, the system will issue a warning and recommend adjusting the flight path.

[0117] Ultimately, based on these flight behavior patterns, the system constructed a multidimensional behavior prediction model. This model not only considers the drone's basic operating characteristics but also incorporates external environmental constraints, such as no-fly zones and inclement weather, to ensure the safe and smooth execution of missions. In actual operations, this model helped the drone select the optimal flight path, avoiding potential risks and ensuring the safe and efficient delivery of packages. Furthermore, the model provided a solid foundation for subsequent flight intent identification and action recommendations, enhancing the reliability and flexibility of the entire system. In this way, drones can achieve smarter and safer operations in complex urban environments.

[0118] This application considers that accurately analyzing the flight intention of drones is crucial for ensuring their safe and efficient operation in the field of drone flight intention recognition and behavior prediction. However, traditional flight intention recognition methods often face problems such as high data complexity, inaccurate feature extraction, and poor model adaptability. Therefore, a new alternative solution is proposed, which includes:

[0119] The deep belief network is used to map the flight data to the multi-dimensional behavior expectation model to calculate the probability distribution of the drone under different flight intentions, generate simulated behavior performance and corresponding probability distribution results, including:

[0120] The flight data is processed nonlinearly using a deep belief network, and the weight of the flight data is dynamically adjusted using an attention mechanism to obtain a flight data representation. The weight of each feature in the flight data is calculated using the following formula: :

[0121] ;

[0122] in, represents the weight of each feature in the flight data, is the temperature parameter, is the number of features, and is the regulating factor, Representation characteristics time delay or spatial distance, and are the weight matrix and bias term in the attention mechanism, It is feature vectors, and is an additional tuning parameter;

[0123] The following is a detailed explanation of each parameter:

[0124] Indicates the flight data The weight of a feature reflects the importance of the feature in identifying the flight intention of the drone.

[0125] and are the weight matrix and bias term in the attention mechanism, respectively. These parameters are obtained through training and are used to adjust the importance of each feature. Their design goal is to enable the model to adaptively focus on features that are more important for the current task.

[0126] Indicates the This is one of the key features extracted from the original flight data, such as position, speed, etc.

[0127] Represents the temperature parameter, which is used to control the smoothness of the softmax function. A value of will make all weights equal, while a lower Smaller values ​​of α increase the variance between the weights. The optimal value is usually determined through experimentation or cross-validation.

[0128] and represents additional tuning parameters. is a global adjustment factor, and These are adjustment parameters for each feature. These parameters are set through training or experience to ensure that the model can flexibly adjust feature weights in different situations.

[0129] and Represents the adjustment factor, used to control time delay or spatial distance The impact on feature weights. These parameters are tuned through experiments or based on domain knowledge to optimize model performance.

[0130] Representation characteristics This reflects the temporal or spatial characteristics of the feature and helps the model understand the temporal dynamics or spatial distribution of the feature.

[0131] The following is an introduction to the design reasons of each sub-item:

[0132] This is a standard linear transformation used to initially adjust the importance of features. The weight matrix learned through training and bias It can help the model better capture the relationship between features and targets.

[0133] This part of the design is to introduce nonlinear components so that the model can handle more complex feature dependencies. Using the hyperbolic sine function (sinh) can enhance the model's expressiveness and enable it to capture subtle differences between features.

[0134] This section considers the impact of a feature's time delay or spatial distance on its weight. By introducing a combination of logarithmic and exponential functions, we can effectively balance the impact of a feature's temporal or spatial characteristics on its weight. This approach prevents certain features from having excessively low weights due to excessive time delay or spatial distance.

[0135] By summing the individual components, we combine them in an additive manner, primarily to comprehensively consider the impact of multiple factors on feature weights. Linear transformations, nonlinear components, and temporal or spatial effects are all considered independent but correlated factors that collectively determine feature importance.

[0136] mapping the flight data representation to different portions of the multi-dimensional behavior expectation model to identify behavior patterns that match the flight data representation;

[0137] Based on the behavior pattern, the flight intention model of the UAV in the multi-dimensional behavior expectation model is constructed, and the similarity between the flight intention models is evaluated by applying the similarity metric learning method to determine the optimal parameter configuration of the flight intention model and generate the behavior pattern configuration. The similarity between the flight data representation and the behavior pattern is evaluated by the following similarity metric formula :

[0138] ;

[0139] in, represents the UAV flight data representation vector, Indicates the The vector representation of the behavior pattern, represents the Euclidean norm, express The transpose of Represents a vector The Euclidean norm of , Represents a vector The Euclidean norm of , is the regularization parameter, express and The Kullback-Leibler divergence between the probability distributions of ;

[0140] The following is a detailed explanation of each parameter:

[0141] Represents the UAV flight data representation vector With the Vector representation of behavioral patterns The higher the value, the more similar the two are.

[0142] and Here is a representation vector of UAV flight data, which is usually extracted from the raw flight data through a deep learning model (such as a deep belief network). A behavior pattern vector is a pre-set behavior pattern library. These patterns are usually constructed based on historical data or expert knowledge and represent the typical behavior of a drone in a specific situation.

[0143] This is a vector and The dot product (inner product) is used to measure the directional consistency of two vectors. The larger the dot product, the closer the directions of the two vectors are.

[0144] and These two symbols represent vectors and The Euclidean norm of (i.e., the length of the vector). The norm is used to normalize the dot product result so that the similarity measure is not affected by the length of the vector.

[0145] Represents the regularization parameter, which is used to balance the weights of the dot product part and the Kullback-Leibler divergence (KL divergence). This parameter is usually set through cross-validation or experience.

[0146] This is and The Kullback-Leibler divergence between the probability distributions of is used to measure the difference between the two distributions. KL divergence is asymmetric, indicating that the difference between the two distributions is not significant. to distribution The relative extraction.

[0147] The following is an introduction to the design reasons of each sub-item:

[0148] This is the formula for calculating cosine similarity. Cosine similarity measures the angular difference between two vectors, rather than their absolute distance. This metric is well-suited for comparing vectors in high-dimensional spaces because it ignores vector magnitude and focuses solely on direction. The rationale behind this is that cosine similarity effectively captures directional consistency between flight data representations and behavioral patterns, which is crucial for identifying drone flight intent.

[0149] The KL divergence is a measure of the difference between two probability distributions. Here, it is used to quantify the probability distribution represented by the flight data. Probability distribution of behavior patterns The reason for this design is that while cosine similarity is a good measure of the directional consistency of vectors, it cannot capture subtle differences between probability distributions. Introducing KL divergence can help the system more comprehensively evaluate the similarity between two distributions, especially when dealing with scenarios with high uncertainty.

[0150] The purpose of summing these components is to comprehensively consider two different similarity measures. Cosine similarity focuses on directional consistency between vectors, while KL divergence focuses on differences between probability distributions. Adding these two components provides a more comprehensive and robust similarity measure.

[0151] Here's a specific example:

[0152] In a city logistics distribution project, suppose we have the following specific values:

[0153] Flight data representation vector , behavior pattern vector , the regularization parameter .

[0154] Calculation steps:

[0155] Calculate cosine similarity:

[0156] Dot Product ;

[0157] Euclidean norm ;

[0158] Euclidean norm ;

[0159] Cosine similarity ;

[0160] Calculating KL divergence: Assumptions and The probability distributions of and , and calculate Since the specific probability distribution needs to be defined according to the actual application background, it is assumed here that .

[0161] Calculate the final similarity: Final Similarity .

[0162] From the calculation results, we can see that the value close to 1 indicates that the flight data represents the vector and behavioral pattern vectors The close proximity in direction indicates that the current flight data is highly consistent with the pre-set behavioral pattern. The small KL divergence value indicates that the difference between the two probability distributions is small, further confirming the consistency between the flight data and the behavioral pattern. The high final similarity of 0.9805 indicates that the current flight data and the behavioral pattern match well, allowing the system to accurately identify the drone's flight intent and make appropriate operational recommendations accordingly. In this way, the system not only accurately identifies the drone's flight intent but also dynamically adjusts flight strategies based on real-time data, ensuring safe and efficient mission completion.

[0163] Based on the above behavior pattern configuration and the Bayesian reasoning framework, the probability distribution of the drone executing each flight intention is calculated, and the variational inference technique is used to obtain the probability distribution result. The probability distribution result is calculated using the following Bayesian formula:

[0164] ;

[0165] in, Indicates the flight intentions, is the mixing coefficient, and Flight Intention The mean vector and covariance matrix of the flight data are represented below, is the number of mixture components, is the data dimension, represents the prior probability distribution, is the total number of flight intentions;

[0166] The following is a detailed explanation of each parameter:

[0167] Indicates that in a given flight data In the case of Flight intention This is the core output of the Bayesian formula and is used to identify the most likely flight intention.

[0168] Representation vectors representing drone flight data are usually extracted from raw flight data through deep learning models such as deep belief networks.

[0169] Indicates the Various flight intentions, such as cruising, hovering or obstacle avoidance.

[0170] Indicates the number of mixture components, indicating that each flight intention can be composed of a mixture of multiple Gaussian distributions. This allows the model to adapt more flexibly to complex data distributions.

[0171] Represents the mixing coefficient, which represents the weight of each mixture component in the overall model. These coefficients are usually estimated through the training process and satisfy .

[0172] and They are Under the flight intention The mean vector and covariance matrix of the mixture components. These parameters describe the center position of each mixture component and the shape of the data distribution. They are also estimated during the training process.

[0173] Represents the data dimension, that is, the feature vector length.

[0174] represents the prior probability distribution, which represents the initial estimate of each flight intention before any data is observed. These probabilities can be obtained through historical data or expert knowledge.

[0175] Represents the likelihood function, which means that under the assumption that the drone performs the In the case of flight intention, flight data is observed probability.

[0176] The following is an introduction to the design reasons of each sub-item:

[0177] This is the core part of the probability density function of the multidimensional Gaussian distribution, which measures the data points Relative to the mean vector The smaller the index part, the and The design rationale is that using a Gaussian distribution can effectively model the probability distribution of continuous data, especially in the presence of noise. This approach is well-suited to handling the uncertainty in flight data.

[0178] This is the standard normalization factor of the multidimensional Gaussian distribution, ensuring that the integral of the probability density function over the entire space is equal to 1. The design reason is that the normalization factor ensures the validity of the probability density function, so that it can correctly reflect the probability distribution of the data points.

[0179] is the mixing coefficient, which balances the contributions of different mixture components. The importance of each component is determined by its corresponding mixing coefficient. The rationale behind this design is that the mixing coefficient allows the model to better fit complex real-world data distributions, rather than being limited to a single Gaussian distribution.

[0180] It represents the prior probability, reflecting the initial estimate of each flight intention before any data is observed. The design reason is that the prior probability provides important background information, helping the system make reasonable judgments when the data is insufficient or ambiguous.

[0181] This is the denominator of the Bayesian formula, which is used to normalize the posterior probabilities to ensure that the sum of all posterior probabilities is 1. The design reason is that the normalization factor ensures the validity and consistency of the posterior probabilities, making them directly usable for comparison and decision-making.

[0182] Based on the probability distribution results In combination with the behavioral pattern configuration, the Monte Carlo method is used to generate a series of simulated behavioral performances and corresponding probability distribution results.

[0183] Here's a specific example:

[0184] In a city logistics distribution project, suppose we have the following specific values:

[0185] Flight data representation vector , flight intention , respectively cruise and hover , the number of mixed ingredients , data dimension .

[0186] Calculation steps:

[0187] Compute the probability density function of each mixture component:

[0188] Assumptions about flight intention , the mean vector and covariance matrix of the mixture components are:

[0189] ;

[0190] The mixing coefficients are ;

[0191] For flight intention , assuming similar parameter settings.

[0192] Calculate the likelihood function for each flight intention:

[0193] Using the formula Compute the probability density function of each mixture component.

[0194] Calculate the posterior probability:

[0195] Combine the likelihood function of each flight intention with the prior probability Combined, substitute into the Bayesian formula to calculate the posterior probability .

[0196] Example calculation:

[0197] Assume that after calculation, we get the following results:

[0198] For flight intention , the weighted sum of the likelihood functions is 0.75.

[0199] For flight intention , the weighted sum of the likelihood functions is 0.25.

[0200] The final posterior probabilities are:

[0201] ;

[0202] ;

[0203] Assume prior probability ,but:

[0204] ;

[0205] ;

[0206] From the calculation results, it can be seen that the flight intention The posterior probability of (cruise) is significantly higher than that of flight intention (Hovering). This means that based on the current flight data, the system is more likely to believe that the drone is performing a cruise behavior. In this way, the system can not only accurately identify the drone's flight intentions, but also dynamically adjust the flight strategy based on real-time data to ensure the safe and efficient completion of the mission.

[0207] Figure 2 A schematic diagram of a UAV flight intention recognition system is provided for an embodiment of the present application. Figure 2 As shown, the system includes:

[0208] A receiving module 21 is configured to receive flight data from a UAV, wherein the flight data includes flight parameters and environmental perception data that have been calibrated for spatiotemporal consistency;

[0209] A construction module 22 is configured to utilize the flight data to construct a multi-dimensional behavior prediction model of the UAV according to a preset behavior pattern library and airspace usage rules;

[0210] A calculation module 23 is configured to map the flight data to the multi-dimensional behavior expectation model using a deep belief network to calculate the probability distribution of the UAV under different flight intentions and generate simulated behavior performance and corresponding probability distribution results;

[0211] a comparison module 24 for comparing, based on the probability distribution result, a deviation metric between the simulated behavior performance and the actual flight activity through a Markov decision process to evaluate and predict the flight intention of the UAV and obtain an optimal flight intention prediction, wherein the deviation metric includes position deviation, speed change, and time difference in response to environmental interaction;

[0212] The generation module 25 is used to generate operation suggestions for the drone based on the flight intention and the immediate airspace safety status, send the operation suggestions to the drone, obtain feedback information after the drone executes the operation, and generate flight guidance suggestions based on the feedback information.

[0213] Figure 2 The UAV flight intention recognition system can perform Figure 1 The implementation principle and technical effects of the method for identifying drone flight intentions described in the illustrated embodiment will not be elaborated on here. The specific manner in which each module and unit performs operations in the drone flight intention identification system in the above embodiment has been described in detail in the embodiments of the method and will not be elaborated on here.

[0214] In one possible design, Figure 2 A drone flight intention recognition system of the embodiment shown can be implemented as a computing device, such as Figure 3 As shown, the computing device may include a storage component 31 and a processing component 32;

[0215] The storage component 31 stores one or more computer instructions, wherein the one or more computer instructions are called and executed by the processing component 32 .

[0216] The processing component 32 is used for the above Figure 1 The embodiment provides a method for identifying the flight intention of a UAV.

[0217] The processing component 32 may include one or more processors to execute computer instructions to complete all or part of the steps in the above method. Of course, the processing component may also be implemented as one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the above method.

[0218] The storage component 31 is configured to store various types of data to support operations at the terminal. The storage component can be implemented by any type of volatile or non-volatile memory device, or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk, or optical disk.

[0219] Of course, a computing device may also include other components, such as input / output interfaces, display components, communication components, etc.

[0220] The input / output interface provides an interface between the processing component and the peripheral interface module, which can be an output device, an input device, etc.

[0221] The communication component is configured to facilitate, among other things, wired or wireless communications between the computing device and other devices.

[0222] Among them, the computing device can be a physical device or an elastic computing host provided by a cloud computing platform, etc. In this case, the computing device can refer to a cloud server, and the above-mentioned processing components, storage components, etc. can be basic server resources rented or purchased from the cloud computing platform.

[0223] The present application also provides a computer storage medium storing a computer program, wherein the computer program can achieve the above-mentioned Figure 1 A method for identifying UAV flight intentions according to the illustrated embodiment.

[0224] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0225] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, i.e., they may be located in one location or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of the present embodiment. Persons of ordinary skill in the art will be able to understand and implement the present invention without inventive effort.

[0226] Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented using software plus a necessary general-purpose hardware platform, or of course, hardware. Based on this understanding, the essence of the above technical solution, or the portion that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, or an optical disk, and includes a number of instructions for causing a computer device (such as a personal computer, server, or network device) to execute the methods described in each embodiment or certain portions of the embodiments.

[0227] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for identifying UAV flight intention, characterized in that: include: Receiving flight data from the UAV, wherein the flight data includes flight parameters and environmental perception data that have been calibrated for spatiotemporal consistency; Using the flight data, a multi-dimensional behavior prediction model of the UAV is constructed based on a preset behavior pattern library and airspace usage rules; Using a deep belief network, the flight data is mapped to the multi-dimensional behavior expectation model to calculate the probability distribution of the drone under different flight intentions, and generate simulated behavior performance and corresponding probability distribution results; Based on the probability distribution results, a Markov decision process is used to compare the deviation metrics between the simulated behavior performance and the actual flight activity to evaluate and predict the UAV's flight intention and obtain an optimal flight intention prediction, wherein the deviation metrics include position deviation, speed change, and time difference in response to environmental interactions; Generate operational suggestions for the drone based on the flight intent and the immediate airspace safety status, send the operational suggestions to the drone, obtain feedback information after the drone executes the suggestions, and generate flight guidance suggestions based on the feedback information; The deep belief network is used to map the flight data to the multi-dimensional behavior expectation model to calculate the probability distribution of the drone under different flight intentions, generate simulated behavior performance and corresponding probability distribution results, including: Using a deep belief network, performing nonlinear transformation processing on the flight data, and dynamically adjusting the weight of the flight data using an attention mechanism to obtain a flight data representation; mapping the flight data representation to different portions of the multi-dimensional behavior expectation model to identify behavior patterns that match the flight data representation; Based on the behavior pattern, a flight intention model of the UAV in the multi-dimensional behavior expectation model is constructed, and a similarity metric learning method is applied to evaluate the similarity between the flight intention models to determine the optimal parameter configuration of the flight intention model and generate a behavior pattern configuration; Based on the behavior pattern configuration and the Bayesian reasoning framework, the probability distribution of the drone executing each flight intention is calculated, and the variational inference technique is used to obtain the probability distribution result; Based on the probability distribution results and the behavior pattern configuration, a Monte Carlo method is used to generate a series of simulated behavior performances and corresponding probability distribution results.

2. The method according to claim 1, characterized in that The method of constructing a flight intention model of the UAV in the multi-dimensional behavior expectation model based on the behavior pattern and applying a similarity metric learning method to evaluate the similarity between the flight intention models to determine the optimal parameter configuration of the flight intention model and generate a behavior pattern configuration includes: Mapping the flight intention of the behavior pattern to different parts of the multi-dimensional behavior expectation model to construct multiple flight intention models; Using the Bayesian optimization method, the optimal initial parameter combination is selected from the flight behavior patterns in the pre-training dataset, and the Gaussian process is used to initialize the parameter configuration of the flight intention model in combination with the optimal initial parameter combination; Utilizing multiple distance measurement methods and kernel methods in combination with twin network learning, the similarity between the flight data representation and the flight intention model is calculated, and the degree of matching between the flight data representation and the flight intention model is quantified to generate a similarity measurement result; Determining an optimal parameter configuration from the parameters of the flight intention model by applying a gradient descent algorithm and a reinforcement learning mechanism based on the similarity measurement result and the initialization parameter configuration; Based on the optimal parameter configuration, a behavior pattern configuration is generated using a context-aware reasoning mechanism.

3. The method according to claim 2, characterized in that Determining the optimal parameter configuration from the parameters of the flight intention model by applying a gradient descent algorithm and a reinforcement learning mechanism based on the similarity measurement result in combination with the initialization parameter configuration includes: According to the similarity measurement result and the initialization parameter configuration, an initial loss function is set, and a gradient descent algorithm is applied to obtain initial parameters of the flight intention model by calculating the gradient of the initial loss function with respect to the initialization parameter configuration and adopting an adaptive learning rate strategy; Using a reinforcement learning mechanism, a reward function is defined, and based on the reward function and actual flight data feedback, initial parameters of the flight intention model are dynamically adjusted to generate an intermediate parameter configuration; Based on the flight data, the intermediate parameter configuration is evaluated to obtain a verification result, and the weights in the loss function and the reward function are adjusted according to the verification result to generate a closed-loop feedback mechanism. Based on the closed-loop feedback mechanism, the optimal parameter configuration is determined.

4. The method according to claim 1, wherein Based on the probability distribution result, the deviation measure between the simulated behavior performance and the actual flight activity is compared through a Markov decision process to evaluate and predict the flight intention of the UAV and obtain the optimal flight intention prediction, including: Based on the probability distribution result, a state transition model is constructed; according to the state transition model, in combination with a preset behavior pattern library and airspace usage rules, a Markov decision process model is initialized, a reward function is defined, and an initial Markov decision process configuration is generated based on the reward function; By comparing the deviation measures between the simulated behavioral performance and the actual flight activities, a series of deviation measure results are obtained; Based on the deviation measurement result and the initial Markov decision process configuration, applying a dynamic time warping algorithm to evaluate the match between the simulated behavior performance and the actual flight activity, and generating a match evaluation result; According to the matching evaluation result, the flight intention corresponding to the simulated behavior performance of the actual flight activity is selected as the prediction result to obtain the optimal flight intention prediction.

5. The method according to claim 4, characterized in that The step of applying a dynamic time warping algorithm based on the deviation measurement result and the initial Markov decision process configuration to evaluate the degree of match between the simulated behavior performance and the actual flight activity and generating a degree of match evaluation result includes: Preprocessing the simulated behavior performance according to the deviation measurement result and in combination with the initial Markov decision process configuration to obtain a preprocessed simulated behavior performance; Determining parameter configuration of a dynamic time warping algorithm based on the characteristics of the simulated behavior performance after the preprocessing and the deviation measurement result; Calculating the time series matching between the simulated behavior performance and the actual flight activity using the parameter configuration of the dynamic time warping algorithm to obtain a time series matching result; quantifying the similarity between the simulated behavior performance and the actual flight activity based on the time series matching result to generate a matching score; Based on the matching score combined with the Markov decision process configuration, the matching of the simulated behavior performance is evaluated to generate a matching evaluation result.

6. The method according to claim 1, wherein The flight data is used to construct a multi-dimensional behavior prediction model of the UAV according to a preset behavior pattern library and airspace usage rules, including: Performing feature extraction processing on the flight data to obtain a set of key feature sets; According to a preset behavior pattern library, the key feature set is compared with the flight behavior patterns in the preset behavior pattern library, the pattern related to the UAV flight behavior is identified, and a behavior pattern result is generated; Using the behavior pattern results, applying airspace use rules, evaluating whether the UAV flight behavior complies with the predicted aviation management regulations and safety requirements, and obtaining a flight behavior pattern; Based on the flight behavior pattern, a multi-dimensional behavior expectation model is constructed.

7. A UAV flight intention recognition system, characterized in that: include: A receiving module, configured to receive flight data from the UAV, wherein the flight data includes flight parameters and environmental perception data that have been calibrated for spatiotemporal consistency; A construction module is used to use the flight data to build a multi-dimensional behavior expectation model of the UAV based on a preset behavior pattern library and airspace usage rules; a computing module, configured to map the flight data to the multi-dimensional behavior expectation model using a deep belief network to calculate the probability distribution of the UAV under different flight intentions and generate simulated behavior performance and corresponding probability distribution results; a comparison module, configured to compare, based on the probability distribution result, a deviation metric between the simulated behavior performance and the actual flight activity through a Markov decision process to evaluate and predict the flight intention of the UAV and obtain an optimal flight intention prediction, wherein the deviation metric includes position deviation, speed change, and time difference in response to environmental interaction; a generation module, configured to generate an operation suggestion for the UAV based on the flight intention and the instantaneous airspace safety status, send the operation suggestion to the UAV, obtain feedback information after the UAV executes the operation suggestion, and generate a flight guidance suggestion based on the feedback information; The deep belief network is used to map the flight data to the multi-dimensional behavior expectation model to calculate the probability distribution of the drone under different flight intentions, generate simulated behavior performance and corresponding probability distribution results, including: Using a deep belief network, performing nonlinear transformation processing on the flight data, and dynamically adjusting the weight of the flight data using an attention mechanism to obtain a flight data representation; mapping the flight data representation to different portions of the multi-dimensional behavior expectation model to identify behavior patterns that match the flight data representation; Based on the behavior pattern, a flight intention model of the UAV in the multi-dimensional behavior expectation model is constructed, and a similarity metric learning method is applied to evaluate the similarity between the flight intention models to determine the optimal parameter configuration of the flight intention model and generate a behavior pattern configuration; Based on the behavior pattern configuration and the Bayesian reasoning framework, the probability distribution of the drone executing each flight intention is calculated, and the variational inference technique is used to obtain the probability distribution result; Based on the probability distribution results and the behavior pattern configuration, a Monte Carlo method is used to generate a series of simulated behavior performances and corresponding probability distribution results.

8. A computing device, characterized in that It includes a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement a drone flight intention recognition method as described in any one of claims 1 to 6.

9. A computer storage medium, characterized in that A computer program is stored, and when the computer program is executed by a computer, a method for identifying the flight intention of a drone as described in any one of claims 1 to 6 is implemented.

Images