Low-altitude aircraft path planning optimization method and system based on big data analysis

Abstract

The invention relates to the technical field of low-altitude aircrafts, and discloses a low-altitude aircraft path planning optimization method and system based on big data analysis, and the method comprises the steps: carrying out the preprocessing of original data, and generating standardized low-altitude data; inputting the standardized low-altitude data into a CNN-LSTM-DDPG model, extracting environment features through a CNN network, capturing an environment dynamic change trend in combination with an LSTM network, and outputting a path adjustment action based on a DDPG algorithm; generating an initial path based on the path adjustment action, and introducing a multi-target dynamic optimization cost function to evaluate and screen the initial path to obtain an optimal path; issuing the optimal path to a physical low-altitude aircraft for execution, collecting flight process data, monitoring the deviation between the flight process data and the optimal path in real time, and when the deviation exceeds a preset threshold value, generating a correction instruction and issuing the correction instruction to the low-altitude aircraft; the method improves the precision of path planning, greatly reduces the flight risk, and guarantees the efficient and safe completion of a task.

Description

Technical Field

[0001] This invention relates to the field of low-altitude aircraft technology, specifically to a method and system for optimizing low-altitude aircraft path planning based on big data analysis. Background Technology

[0002] With the booming development of the low-altitude economy, low-altitude aircraft are increasingly widely used in various fields. Path planning, as a core technology to ensure the safe and efficient operation of low-altitude aircraft, directly determines the quality of mission completion and airspace operation safety. Current technologies generally suffer from insufficient responsiveness. Traditional planning algorithms are mostly based on static data to generate paths, which makes it difficult to adapt to dynamic environmental variables in real time, resulting in lag in path adjustment. Moreover, multi-objective optimization is unbalanced. Most algorithms only consider the shortest distance or the lowest energy consumption, without fully integrating multi-dimensional constraints, and cannot meet the needs of large-scale, high-precision path planning in complex low-altitude environments. Therefore, there is an urgent need for a low-altitude aircraft path planning optimization method based on big data analysis to achieve integrated planning with dynamic adaptation and multi-objective global optimization. Summary of the Invention

[0003] The purpose of this invention is to solve the above problems by designing a low-altitude aircraft path planning optimization method and system based on big data analysis.

[0004] The first aspect of this invention provides a path planning optimization method for low-altitude aircraft based on big data analysis, the method comprising the following steps: Collect environmental data, aircraft status data, and mission requirement data to form raw data. Preprocess the raw data to generate standardized low-altitude data. Standardized low-altitude data is input into the CNN-LSTM-DDPG model. The CNN network extracts environmental features, and the LSTM network captures the dynamic changes in the environment. The DDPG algorithm outputs path adjustment actions. The CNN-LSTM-DDPG model is optimized for parameters based on the FGO algorithm. An initial path is generated based on the path adjustment action. A multi-objective dynamic optimization cost function is introduced to evaluate and filter the initial path to obtain the optimal path. The optimal path is sent to the physical low-altitude aircraft for execution, and flight process data is collected. The deviation between the flight process data and the optimal path is monitored in real time. When the deviation exceeds a preset threshold, a correction command is generated and sent to the low-altitude aircraft.

[0005] Optionally, in a first implementation of the first aspect of the present invention, the collected environmental data, aircraft status data, and mission requirement data form raw data; the raw data is preprocessed to generate standardized low-altitude data, including: First, multi-scale wavelet decomposition is performed on the original data to obtain low-frequency approximate components and high-frequency detail components. Then, noise components in the high-frequency detail components are suppressed. The processed high-frequency components and low-frequency components are subjected to inverse wavelet transform to obtain the denoised clean data. The collaborative Kalman filter algorithm is used to perform timestamp alignment and fusion processing on the denoised clean data to obtain the fused data. The min-max normalization algorithm is used to standardize the fused data, ultimately generating standardized low-altitude data.

[0006] Optionally, in a second implementation of the first aspect of the present invention, the step of inputting standardized low-altitude data into a CNN-LSTM-DDPG model, extracting environmental features through a CNN network, capturing dynamic trends of environmental changes through an LSTM network, and outputting path adjustment actions based on the DDPG algorithm includes: Standardized low-altitude data is reorganized according to time series and spatial dimensions to divide it into environmental characteristic data subsets and dynamic trend data subsets; First, a subset of environmental feature data is input into a CNN network. Then, a subset of dynamic trend data is concatenated with the high-dimensional environmental feature vector output by the CNN network and input into an LSTM network to output a time-series feature vector containing the dynamic trend of the environment. The temporal feature vector is input into the policy network of the DDPG algorithm. The policy network performs a nonlinear transformation on the temporal feature vector through a fully connected layer to generate an initial path adjustment action. The temporal feature vector and the initial path adjustment action are then input into the value network, which outputs an evaluation feedback value.

[0007] Optionally, in a third implementation of the first aspect of the present invention, the step of first inputting a subset of environmental feature data into a CNN network, then concatenating a subset of dynamic trend data with the high-dimensional environmental feature vector output by the CNN network and inputting the concatenated subset of dynamic trend data into an LSTM network, and outputting a temporal feature vector containing the dynamic trend of the environment, includes: The CNN network first performs sliding convolution operations using convolutional kernels of a preset size through convolutional layers to extract local spatial features. Then, it performs downsampling processing on the convolutional features through pooling layers. After multiple rounds of convolution and pooling operations, it outputs a high-dimensional environmental feature vector. LSTM networks process time-series data step by step through the combined action of the input gate, forget gate and output gate of LSTM units, and finally output a time-series feature vector containing the dynamic trend of the environment.

[0008] Optionally, in a fourth implementation of the first aspect of the present invention, the CNN-LSTM-DDPG model is optimized for parameters based on the FGO algorithm, including: Construct a function with the objective function of minimizing error, and set the iteration step size, maximum number of iterations, and error convergence threshold of the FGO algorithm; The gradient values ​​of the objective function with respect to each parameter to be optimized are calculated using the FGO algorithm. The direction of parameter adjustment is determined based on the gradient direction, and each parameter to be optimized is updated in combination with the preset iteration step size. After completing one round of parameter updates, the updated parameters are substituted into the CNN-LSTM-DDPG model to re-output the path adjustment action and calculate new error data; Repeat the parameter update and error calculation process. After each iteration, compare the current error data with the preset error convergence threshold to determine whether the error meets the convergence requirements. When the error data converges to the preset threshold or reaches the maximum number of iterations, the iteration process stops, and the final optimized parameters are substituted into the CNN-LSTM-DDPG model.

[0009] Optionally, in a fifth implementation of the first aspect of the present invention, the step of generating an initial path based on the path adjustment action, and introducing a multi-objective dynamic optimization cost function to evaluate and filter the initial path to obtain the optimal path includes: Using the task start point and target point as the base anchor points, the direction, speed, and height parameters in the action are adjusted according to the path to generate continuous path nodes segment by segment, and the nodes are connected in time sequence to form the initial path. The initial path length, aircraft operating energy consumption, safety risks in the areas traversed, and time required to complete the mission are evaluated sequentially to obtain cost assessment results for each dimension. Based on the cost evaluation results of each dimension, the cost evaluation results are weighted and integrated according to the preset weights to obtain the comprehensive evaluation value of the initial path. At the same time, multiple alternative initial paths are generated, the evaluation process is repeated, and finally the path with the best comprehensive evaluation value is selected as the optimal path.

[0010] Optionally, in a sixth implementation of the first aspect of the present invention, the step of sending the optimal path to a physical low-altitude aircraft for execution, collecting flight process data, monitoring the deviation between the flight process data and the optimal path in real time, and generating a correction command and sending it to the low-altitude aircraft when the deviation exceeds a preset threshold includes: The optimal path is encrypted using a symmetric encryption algorithm to generate an encrypted path data packet, which is then transmitted to the low-altitude aircraft. Collect real-time status data of the aircraft, dynamic environmental change data, and mission execution progress data during the flight process to form flight process data; The flight process data is compared with the preset standard data corresponding to the optimal path in each dimension, and the deviation value of each dimension of the data is calculated. The calculated deviation value is compared with the preset deviation threshold to determine whether there are any deviation items that exceed the threshold. If there are deviations exceeding the threshold, a path adjustment correction command is generated based on the magnitude and direction of the deviation value, and the correction command is sent to the low-altitude aircraft to execute the corresponding adjustment action.

[0011] A second aspect of the present invention provides a path planning and optimization system for low-altitude aircraft based on big data analysis, the system comprising: The preprocessing module is used to collect environmental data, aircraft status data, and mission requirement data to form raw data, and to preprocess the raw data to generate standardized low-altitude data. The model processing module is used to input standardized low-altitude data into the CNN-LSTM-DDPG model, extract environmental features through the CNN network, capture the dynamic change trend of the environment by combining the LSTM network, and output path adjustment actions based on the DDPG algorithm. The CNN-LSTM-DDPG model is optimized for parameters based on the FGO algorithm. The evaluation and screening module is used to generate an initial path based on the path adjustment action, and to evaluate and screen the initial path by introducing a multi-objective dynamic optimization cost function to obtain the optimal path. The instruction issuing module is used to issue the optimal path to the physical low-altitude aircraft for execution, collect flight process data, monitor the deviation between the flight process data and the optimal path in real time, and generate a correction instruction and issue it to the low-altitude aircraft when the deviation exceeds a preset threshold.

[0012] A third aspect of the present invention provides a low-altitude aircraft path planning optimization device based on big data analysis, the low-altitude aircraft path planning optimization device based on big data analysis including a memory and at least one processor, the memory storing instructions; the at least one processor calling the instructions in the memory to cause the low-altitude aircraft path planning optimization device based on big data analysis to perform the various steps of the low-altitude aircraft path planning optimization method based on big data analysis as described in any of the preceding claims.

[0013] A fourth aspect of the present invention provides a computer-readable storage medium storing instructions that, when executed by a processor, implement the steps of the low-altitude aircraft path planning optimization method based on big data analysis as described in any of the preceding claims.

[0014] In the technical solution provided by this invention, environmental data, aircraft status data, and mission requirement data are collected to form raw data. This raw data is preprocessed to generate standardized low-altitude data. The standardized low-altitude data is then input into a CNN-LSTM-DDPG model. The CNN network extracts environmental features, and the LSTM network captures dynamic environmental changes. A path adjustment action is output based on the DDPG algorithm. An initial path is generated based on the path adjustment action. A multi-objective dynamic optimization cost function is introduced to evaluate and filter the initial path to obtain the optimal path. The optimal path is then sent to the physical low-altitude aircraft for execution, and flight process data is collected. The deviation between the flight process data and the optimal path is monitored in real time. When the deviation exceeds a preset threshold, a correction command is generated and sent out. For low-altitude aircraft, this invention effectively removes noise from raw data and achieves standardized transformation, improving the reliability of basic data for path planning. Employing a CNN-LSTM-DDPG model combined with the FGO algorithm to optimize parameters, it can accurately extract environmental spatial features and capture dynamic change trends. The output path adjustment actions are more in line with actual flight requirements, improving the accuracy of path planning. A multi-objective dynamic optimization cost function is introduced to select the optimal path, comprehensively balancing multiple dimensions such as distance, energy consumption, and safety, enhancing the practicality and economy of the path. A closed loop for real-time monitoring and deviation correction during flight is constructed, enabling timely response to flight deviations and the issuance of correction commands, significantly reducing flight risks and ensuring efficient and safe mission completion, providing strong technical support for the large-scale application of low-altitude aircraft. Attached Figure Description

[0015] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the invention.

[0016] Figure 1 A flowchart of a low-altitude aircraft path planning optimization method based on big data analysis provided in an embodiment of the present invention; Figure 2 A schematic diagram of the structure of a low-altitude aircraft path planning and optimization system based on big data analysis provided in an embodiment of the present invention; Figure 3 A schematic diagram of the structure of a low-altitude aircraft path planning optimization device based on big data analysis provided in an embodiment of the present invention. Detailed Implementation

[0017] The terms “first,” “second,” “third,” “fourth,” etc. (if present) in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a particular order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms “comprising” or “having,” and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, apparatus, product, or device that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or devices.

[0018] For ease of understanding, the specific process of the embodiments of the present invention is described below. Please refer to [link / reference]. Figure 1 The flowchart of the low-altitude aircraft path planning optimization method based on big data analysis provided in this embodiment of the invention includes the following steps: Step 101: Collect environmental data, aircraft status data, and mission requirement data to form raw data. Preprocess the raw data to generate standardized low-altitude data. In this embodiment, the original data is first decomposed into low-frequency approximate components and high-frequency detail components. Then, the noise components in the high-frequency detail components are suppressed. The processed high-frequency components and low-frequency components are then subjected to inverse wavelet transform to obtain denoised clean data. The collaborative Kalman filter algorithm is used to perform timestamp alignment and fusion processing on the denoised clean data to obtain fused data. The min-max normalization algorithm is used to standardize the fused data to finally generate standardized low-altitude data.

[0019] Step 102: Input standardized low-altitude data into the CNN-LSTM-DDPG model, extract environmental features through the CNN network, capture the dynamic change trend of the environment by combining the LSTM network, and output path adjustment actions based on the DDPG algorithm. The CNN-LSTM-DDPG model is optimized for parameters based on the FGO algorithm. In this embodiment, standardized low-altitude data is reorganized according to time series and spatial dimensions to divide it into environmental feature data subsets and dynamic trend data subsets. The environmental feature data subset is first input into a CNN network, and then the dynamic trend data subset is concatenated with the high-dimensional environmental feature vector output by the CNN network and input into an LSTM network to output a time series feature vector containing environmental dynamic trends. The time series feature vector is input into the policy network of the DDPG algorithm. The policy network performs a nonlinear transformation on the time series feature vector through a fully connected layer to generate an initial path adjustment action. The time series feature vector and the initial path adjustment action are then input into the value network to output an evaluation feedback value.

[0020] In this embodiment, the CNN network first performs sliding convolution operations using convolutional kernels of a preset size through convolutional layers to extract local spatial features. Then, it performs downsampling processing on the convolutional features through pooling layers. After multiple rounds of convolution and pooling operations, it outputs a high-dimensional environmental feature vector. The LSTM network processes the temporal data step by step through the synergistic effect of the input gate, forget gate and output gate of the LSTM unit, and finally outputs a temporal feature vector containing the dynamic trend of the environment.

[0021] In this embodiment, an algorithm with error minimization as the objective function is constructed, and the iteration step size, maximum number of iterations, and error convergence threshold of the FGO algorithm are set. The gradient value of the objective function with respect to each parameter to be optimized is calculated using the FGO algorithm. The parameter adjustment direction is determined based on the gradient direction, and each parameter to be optimized is updated in combination with the preset iteration step size. After completing one round of parameter update, the updated parameters are substituted into the CNN-LSTM-DDPG model, the path adjustment action is re-output, and new error data is calculated. The parameter update and error calculation process is repeated. After each iteration, the current error data is compared with the preset error convergence threshold to determine whether the error meets the convergence requirements. When the error data converges to the preset threshold or reaches the maximum number of iterations, the iteration process is stopped, and the final optimized parameters are substituted into the CNN-LSTM-DDPG model.

[0022] In this embodiment, based on the core error data obtained in the early stage, an objective function with error minimization as its core is constructed. The function constraint is defined as the reasonable range of values ​​for each parameter to be optimized in the CNN-LSTM-DDPG model, avoiding parameter adjustments that exceed the model's operational limits. Simultaneously, considering model training requirements and computational resource allocation, key parameters of the FGO algorithm are set: the iteration step size is dynamically adapted based on parameter sensitivity, with smaller step sizes set for core parameters that significantly affect the accuracy of path action output to ensure optimization precision, and relatively larger step sizes set for auxiliary parameters to improve optimization efficiency; the maximum number of iterations is set based on historical optimization data and model convergence patterns to avoid insufficient optimization due to insufficient iterations or resource waste due to excessive iterations; the error convergence threshold is determined in conjunction with the task's accuracy requirements for path planning to ensure that the model error after convergence meets practical application standards; the FGO algorithm is launched to solve the gradient of the constructed objective function. Through the algorithm's built-in gradient calculation module, the gradient values ​​of the objective function with respect to each parameter to be optimized (CNN network convolutional kernel weights, LSTM network unit weights, DDPG policy and value network fully connected layer weights and bias terms) are calculated one by one. The direction of parameter adjustment is determined by the sign of the gradient value: a positive gradient indicates a decrease in parameter value, while a negative gradient indicates an increase in parameter value. Then, the specific adjustment amount for each parameter is calculated based on the preset iteration step size. Each parameter to be optimized is updated sequentially according to its importance priority, completing one round of parameter adjustment. All optimized parameters from the previous round are fully substituted back into the CNN-LSTM-DDPG model, keeping other input conditions (standardized low-altitude data, network structure configuration, etc.) consistent with those before optimization. The model is then restarted to output new path adjustment actions. Using a preset standard action as a benchmark, the same error calculation method as before is employed to compare the deviation between the newly output path adjustment actions and the standard actions, generating new error data after this round of parameter updates. This error data is then standardized to form an error index that can be directly used for convergence judgment. The new error data generated in this round is compared one by one with a preset error convergence threshold to determine whether the current error meets the convergence requirements: if the error data is less than or equal to the convergence threshold, it indicates that the model parameters have been optimized to the target state, and the iteration process can be terminated; if the error data is still greater than the convergence threshold, the current error data and parameter state must be recorded, and the process of gradient calculation, parameter update, and error recalculation is returned to step two to start the next iteration. The iteration count is recorded synchronously during each iteration to ensure traceability of the iteration process. The iterative process continues until one of two termination conditions is met: either the error data converges to a preset threshold, or the number of iterations reaches the set maximum. When the termination condition is met, the iteration process immediately stops, the final optimized parameters of the current model are extracted and their validity is verified to ensure they meet the basic requirements for model operation. After successful verification, the final optimized parameters are fixed in the parameter configuration file of the CNN-LSTM-DDPG model, overwriting the original initial parameters. This ensures that subsequent model operations are based on the optimized parameters for path planning, completing the entire FGO algorithm parameter optimization process.

[0023] Step 103: Generate an initial path based on the path adjustment action, and introduce a multi-objective dynamic optimization cost function to evaluate and filter the initial path to obtain the optimal path; In this embodiment, the mission start point and target point are used as the base anchor points. The direction, speed, and altitude parameters in the path adjustment action are adjusted to generate continuous path nodes segment by segment. The nodes are connected in time sequence to form the initial path. The length of the initial path, the energy consumption of the aircraft, the safety risks of the area it passes through, and the time required to complete the mission are evaluated in sequence to obtain the cost evaluation results of each dimension. Based on the cost evaluation results of each dimension, the cost evaluation results are weighted and fused according to the preset weights to obtain the comprehensive evaluation value of the initial path. At the same time, multiple alternative initial paths are generated, the evaluation process is repeated, and finally the path with the best comprehensive evaluation value is selected as the optimal path.

[0024] In this embodiment, the precise geographic coordinates of the mission's starting point and target point are used as the base anchor points. First, a positioning calibration algorithm is used to confirm the spatial accuracy of the anchor points, eliminating the impact of positioning errors on path generation. Then, the core parameters of direction, speed, and altitude from the output path adjustment actions are extracted. Combined with the dynamic performance parameters of the low-altitude aircraft and the basic requirements of mission execution, a generation interval threshold for path nodes is set. According to the set interval threshold, starting from the initial anchor point, the extension direction of each path segment is determined based on the direction parameters. The node generation time interval for each path segment is calculated based on the speed parameters. The spatial altitude coordinates of each node are determined based on the altitude parameters, generating continuous path nodes that conform to physical flight constraints segment by segment. Finally, a path smoothing algorithm is used to sequentially sort and connect the nodes, eliminating abrupt transitions between nodes and forming a continuous initial path that can be directly executed by the aircraft. Multi-objective cost evaluation is then performed. Finally, a multi-objective comprehensive evaluation algorithm is initiated to process the generated initial path. The comprehensive cost assessment first employs a path length calculation algorithm. Based on the coordinate information of each path node, the path length is calculated segment by segment using a spatial distance formula and then accumulated to obtain the initial path's total length assessment result. Next, combining the aircraft's dynamic model and energy consumption characteristic curves, energy consumption data for different flight stages is calculated based on the speed, altitude changes, and load weight of each segment of the path, resulting in a comprehensive aircraft operation energy consumption assessment result. Subsequently, the low-altitude environment risk assessment database is accessed to systematically examine the areas traversed by the initial path. Considering factors such as no-fly zone distribution, obstacle density, and weather conditions, the safety risk level of the traversed areas is quantitatively assessed, forming a safety risk cost assessment result. Finally, based on the initial path's total length and the speed parameters of each segment, the theoretical time required to complete the mission is calculated. Simultaneously, the airspace congestion probability of the traversed areas is considered to correct the theoretical time, resulting in a mission timeliness cost assessment result. Ultimately, these results are integrated to form a multi-dimensional cost assessment result set. Based on the preset weight allocation rules obtained from previous big data statistical training, the cost evaluation results of each dimension—path length, operational energy consumption, safety risk, and mission timeliness—are assigned corresponding weight coefficients. A weighted summation algorithm is used to fuse the cost evaluation results of each dimension, obtaining the comprehensive evaluation value of the initial path. The lower the comprehensive evaluation value, the better the overall performance of the path. Simultaneously, a candidate path generation mechanism is activated. Between the starting anchor point and the target anchor point, multiple candidate initial paths that meet basic flight constraints are generated by adjusting the parameter combinations of path adjustment actions. The above multi-objective cost evaluation process is repeated for each candidate initial path, calculating the comprehensive evaluation value of each candidate path one by one. Finally, a comprehensive evaluation value sorting algorithm is used to sort the comprehensive evaluation values ​​of the initial path and all candidate paths in descending order, selecting the path with the lowest comprehensive evaluation value as the optimal path, ensuring that this path achieves optimal overall performance under multi-dimensional constraints.

[0025] Step 104: Send the optimal path to the physical low-altitude aircraft for execution, collect flight process data, monitor the deviation between the flight process data and the optimal path in real time, and generate a correction command when the deviation exceeds a preset threshold and send it to the low-altitude aircraft.

[0026] In this embodiment, a symmetric encryption algorithm is used to encrypt the optimal path, generating an encrypted path data packet, which is then transmitted to the low-altitude aircraft. Real-time status data of the aircraft, dynamic environmental change data, and mission execution progress data are collected during the flight to form flight process data. The flight process data is compared dimension by dimension with the preset standard data corresponding to the optimal path, and the deviation value of each dimension is calculated. The calculated deviation value is compared with a preset deviation threshold to determine whether there are any deviation items exceeding the threshold. If there are deviation items exceeding the threshold, a path adjustment correction command is generated based on the magnitude and direction of the deviation value, and the correction command is sent to the low-altitude aircraft to execute the corresponding adjustment action.

[0027] Please see Figure 2 A schematic diagram of the structure of a low-altitude aircraft path planning optimization system based on big data analysis provided in this embodiment of the invention. The system includes: The preprocessing module is used to collect environmental data, aircraft status data, and mission requirement data to form raw data, and to preprocess the raw data to generate standardized low-altitude data. The model processing module is used to input standardized low-altitude data into the CNN-LSTM-DDPG model, extract environmental features through the CNN network, capture the dynamic change trend of the environment by combining the LSTM network, and output path adjustment actions based on the DDPG algorithm. The CNN-LSTM-DDPG model is optimized for parameters based on the FGO algorithm. The evaluation and screening module is used to generate an initial path based on path adjustment actions, and introduce a multi-objective dynamic optimization cost function to evaluate and screen the initial path to obtain the optimal path; The command issuance module is used to issue the optimal path to the physical low-altitude aircraft for execution, collect flight process data, monitor the deviation between the flight process data and the optimal path in real time, and generate correction commands and issue them to the low-altitude aircraft when the deviation exceeds a preset threshold. Figure 3This is a schematic diagram of a low-altitude aircraft path planning optimization device 300 based on big data analysis, provided in an embodiment of the present invention. This device 300 can vary significantly due to different configurations or performance characteristics. It may include one or more central processing units (CPUs) 310 (e.g., one or more processors) and a memory 320, and one or more storage media 330 (e.g., one or more mass storage devices) storing application programs 333 or data 332. The memory 320 and storage media 330 can be temporary or persistent storage. The program stored in the storage media 330 may include one or more modules (not shown in the diagram), each module including a series of instruction operations on the low-altitude aircraft path planning optimization device 300 based on big data analysis. Furthermore, the processor 310 may be configured to communicate with the storage media 330 and execute the series of instruction operations in the storage media 330 on the low-altitude aircraft path planning optimization device 300 based on big data analysis to implement the method provided in the above embodiment.

[0028] The low-altitude aircraft path planning and optimization device 300 based on big data analysis may also include one or more power supplies 340, one or more wired or wireless network interfaces 350, one or more input / output interfaces 360, and / or one or more operating systems 331, such as Windows Server, Mac OS X, Unix, Linux, FreeBSD, etc. Those skilled in the art will understand that... Figure 3 The structure of the low-altitude aircraft path planning optimization device based on big data analysis shown does not constitute a limitation on the computer device provided by the present invention. It may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0029] The present invention also provides a computer-readable storage medium, which can be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium, wherein the computer-readable storage medium stores instructions that, when executed on a computer, cause the computer to perform the various steps of the low-altitude aircraft path planning optimization method based on big data analysis provided in the above embodiments.

[0030] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working process of the above-described equipment, apparatus, or unit can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0031] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0032] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely preferred examples and are not intended to limit the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended technical solutions and their equivalents.

Claims

1. A path planning optimization method for low-altitude aircraft based on big data analysis, characterized in that, The method includes the following steps: Collect environmental data, aircraft status data, and mission requirement data to form raw data. Preprocess the raw data to generate standardized low-altitude data. Standardized low-altitude data is input into the CNN-LSTM-DDPG model. The CNN network extracts environmental features, and the LSTM network captures the dynamic changes in the environment. The DDPG algorithm outputs path adjustment actions. The CNN-LSTM-DDPG model is optimized for parameters based on the FGO algorithm. An initial path is generated based on the path adjustment action. A multi-objective dynamic optimization cost function is introduced to evaluate and filter the initial path to obtain the optimal path. The optimal path is sent to the physical low-altitude aircraft for execution, and flight process data is collected. The deviation between the flight process data and the optimal path is monitored in real time. When the deviation exceeds a preset threshold, a correction command is generated and sent to the low-altitude aircraft.

2. The low-altitude aircraft path planning optimization method based on big data analysis as described in claim 1, characterized in that, The collected environmental data, aircraft status data, and mission requirement data form raw data. This raw data is then preprocessed to generate standardized low-altitude data, including: First, multi-scale wavelet decomposition is performed on the original data to obtain low-frequency approximate components and high-frequency detail components. Then, noise components in the high-frequency detail components are suppressed. The processed high-frequency components and low-frequency components are subjected to inverse wavelet transform to obtain the denoised clean data. The collaborative Kalman filter algorithm is used to perform timestamp alignment and fusion processing on the denoised clean data to obtain the fused data. The min-max normalization algorithm is used to standardize the fused data, ultimately generating standardized low-altitude data.

3. The low-altitude aircraft path planning optimization method based on big data analysis as described in claim 1, characterized in that, The process involves inputting standardized low-altitude data into the CNN-LSTM-DDPG model, extracting environmental features through the CNN network, capturing dynamic changes in the environment using the LSTM network, and outputting path adjustment actions based on the DDPG algorithm, including: Standardized low-altitude data is reorganized according to time series and spatial dimensions to divide it into environmental characteristic data subsets and dynamic trend data subsets; First, a subset of environmental feature data is input into a CNN network. Then, a subset of dynamic trend data is concatenated with the high-dimensional environmental feature vector output by the CNN network and input into an LSTM network to output a time-series feature vector containing the dynamic trend of the environment. The temporal feature vector is input into the policy network of the DDPG algorithm. The policy network performs a nonlinear transformation on the temporal feature vector through a fully connected layer to generate an initial path adjustment action. The temporal feature vector and the initial path adjustment action are then input into the value network, which outputs an evaluation feedback value.

4. The low-altitude aircraft path planning optimization method based on big data analysis as described in claim 3, characterized in that, The process involves first inputting a subset of environmental feature data into a CNN network, then concatenating a subset of dynamic trend data with the high-dimensional environmental feature vector output by the CNN network, and finally inputting the concatenated data into an LSTM network. The output is a time-series feature vector containing the dynamic trends of the environment, including: The CNN network first performs sliding convolution operations using convolutional kernels of a preset size through convolutional layers to extract local spatial features. Then, it performs downsampling processing on the convolutional features through pooling layers. After multiple rounds of convolution and pooling operations, it outputs a high-dimensional environmental feature vector. LSTM networks process time-series data step by step through the combined action of the input gate, forget gate and output gate of LSTM units, and finally output a time-series feature vector containing the dynamic trend of the environment.

5. The low-altitude aircraft path planning optimization method based on big data analysis as described in claim 1, characterized in that, The CNN-LSTM-DDPG model is optimized for parameters based on the FGO algorithm, including: Construct a function with the objective function of minimizing error, and set the iteration step size, maximum number of iterations, and error convergence threshold of the FGO algorithm; The gradient values ​​of the objective function with respect to each parameter to be optimized are calculated using the FGO algorithm. The direction of parameter adjustment is determined based on the gradient direction, and each parameter to be optimized is updated in combination with the preset iteration step size. After completing one round of parameter updates, the updated parameters are substituted into the CNN-LSTM-DDPG model to re-output the path adjustment action and calculate new error data; Repeat the parameter update and error calculation process. After each iteration, compare the current error data with the preset error convergence threshold to determine whether the error meets the convergence requirements. When the error data converges to the preset threshold or reaches the maximum number of iterations, the iteration process stops, and the final optimized parameters are substituted into the CNN-LSTM-DDPG model.

6. The low-altitude aircraft path planning optimization method based on big data analysis as described in claim 1, characterized in that, The process of generating an initial path based on the path adjustment action, and then evaluating and filtering the initial path using a multi-objective dynamic optimization cost function to obtain the optimal path, includes: Using the task start point and target point as the base anchor points, the direction, speed, and height parameters in the action are adjusted according to the path to generate continuous path nodes segment by segment, and the nodes are connected in time sequence to form the initial path. The initial path length, aircraft operating energy consumption, safety risks in the areas traversed, and time required to complete the mission are evaluated sequentially to obtain cost assessment results for each dimension. Based on the cost evaluation results of each dimension, the cost evaluation results are weighted and fused according to the preset weights to obtain the comprehensive evaluation value of the initial path. At the same time, multiple alternative initial paths are generated, the evaluation process is repeated, and finally the path with the best comprehensive evaluation value is selected as the optimal path.

7. The low-altitude aircraft path planning optimization method based on big data analysis as described in claim 1, characterized in that, The process of sending the optimal path to the physical low-altitude aircraft for execution, collecting flight process data, monitoring the deviation between the flight process data and the optimal path in real time, and generating a correction command and sending it to the low-altitude aircraft when the deviation exceeds a preset threshold includes: The optimal path is encrypted using a symmetric encryption algorithm to generate an encrypted path data packet, which is then transmitted to the low-altitude aircraft. Collect real-time status data of the aircraft, dynamic environmental change data, and mission execution progress data during the flight process to form flight process data; The flight process data is compared with the preset standard data corresponding to the optimal path in each dimension, and the deviation value of each dimension of the data is calculated. The calculated deviation value is compared with the preset deviation threshold to determine whether there are any deviation items that exceed the threshold. If there are deviations exceeding the threshold, a path adjustment correction command is generated based on the magnitude and direction of the deviation value, and the correction command is sent to the low-altitude aircraft to execute the corresponding adjustment action.

8. A path planning and optimization system for low-altitude aircraft based on big data analysis, characterized in that, The system includes: The preprocessing module is used to collect environmental data, aircraft status data, and mission requirement data to form raw data, and to preprocess the raw data to generate standardized low-altitude data. The model processing module is used to input standardized low-altitude data into the CNN-LSTM-DDPG model, extract environmental features through the CNN network, capture the dynamic change trend of the environment by combining the LSTM network, and output path adjustment actions based on the DDPG algorithm. The CNN-LSTM-DDPG model is optimized for parameters based on the FGO algorithm. The evaluation and screening module is used to generate an initial path based on the path adjustment action, and to evaluate and screen the initial path by introducing a multi-objective dynamic optimization cost function to obtain the optimal path. The instruction issuing module is used to issue the optimal path to the physical low-altitude aircraft for execution, collect flight process data, monitor the deviation between the flight process data and the optimal path in real time, and generate a correction instruction when the deviation exceeds a preset threshold and issue it to the low-altitude aircraft.

9. A path planning and optimization device for low-altitude aircraft based on big data analysis, characterized in that, The low-altitude aircraft path planning optimization device based on big data analysis includes a memory and at least one processor, wherein the memory stores instructions; the at least one processor invokes the instructions in the memory to cause the low-altitude aircraft path planning optimization device based on big data analysis to perform each step of the low-altitude aircraft path planning optimization method based on big data analysis as described in any one of claims 1-7.

10. A computer-readable storage medium storing instructions thereon, characterized in that, When the instructions are executed by the processor, they implement the various steps of the low-altitude aircraft path planning optimization method based on big data analysis as described in any one of claims 1-7.

Images