Target situation awareness method and system based on low-altitude intelligent network

Abstract

The present invention relates to the technical field of low-altitude intelligent networking, and specifically provides a target situation perception method and system based on the low-altitude intelligent networking. First, a target perception data set including three-dimensional position data and environmental disturbance data collaboratively collected by multiple nodes in the low-altitude intelligent networking is obtained. Then, feature extraction is performed on the target perception data set to generate spatial distribution features and continuous change features of the target object. The spatial distribution features and continuous change features are then input into a pre-trained situation perception model to generate situation correlation features. The current situation attributes and situation development trend information of the target object are obtained through a situation classification layer. Finally, a collaborative perception instruction including a target positioning identifier is generated according to the current situation attributes and situation development trend information and is sent to a low-altitude intelligent networking management terminal, thereby enabling comprehensive and accurate perception of the low-altitude target situation, providing effective collaborative perception instructions for the low-altitude intelligent networking, and improving safety and management efficiency in the low-altitude field.

Description

Technical Field

[0001] The present invention relates to the technical field of low-altitude intelligent networking, and in particular to a target situation awareness method and system based on a low-altitude intelligent networking. Background Art

[0002] In the low-altitude sector, the increasing presence of drones and other low-altitude aircraft has significantly increased the complexity and uncertainty of the low-altitude environment. As a new network architecture, the Low-Altitude Intelligent Network (LAI) provides a data foundation for low-altitude target situational awareness by collaboratively collecting target perception data from multiple nodes. However, existing low-altitude target situational awareness technologies have numerous shortcomings.

[0003] On the one hand, existing technologies often only focus on the location information of the target, while ignoring the important role of environmental disturbance data on the target situation. In fact, environmental factors such as wind speed, wind direction, and air pressure will have a significant impact on the flight status of low-altitude targets. It is difficult to accurately grasp the true situation of the target by relying solely on location information. On the other hand, in terms of data processing, existing technologies lack effective feature extraction of target perception data and are unable to fully explore the characteristic information of the target object in the spatial and temporal dimensions, resulting in insufficient accuracy and comprehensiveness of situation perception. In addition, existing situation perception models have limitations in generating situation-related features and situation classification, making it difficult to accurately judge the current situation attributes and situation development trends of the target object, and unable to provide effective collaborative perception instructions for the low-altitude intelligent network management terminal, thereby affecting the timeliness and accuracy of the linkage response operation. Summary of the Invention

[0004] In view of the above-mentioned problems, in combination with the first aspect of the present invention, the present invention provides a target situation awareness method based on a low-altitude intelligent network, the method comprising:

[0005] Acquire a target perception data set collaboratively collected by multiple nodes in a low-altitude intelligent network, wherein the target perception data set includes three-dimensional position data and environmental disturbance data collected by different nodes in continuous time periods;

[0006] Performing feature extraction processing on the target perception data set to generate spatial distribution features and continuous change features of the target object, wherein the spatial distribution features represent the position correlation relationship of the target object in the low-altitude area, and the continuous change features represent the state change law of the target object in a continuous time period;

[0007] Inputting the spatial distribution features and the continuous change features into a pre-trained situation awareness model to generate situation-related features of the target object;

[0008] Inputting the situation association features into the situation classification layer of the situation awareness model to obtain the current situation attributes and situation development trend information of the target object;

[0009] A collaborative perception instruction including a target positioning identifier is generated according to the current situation attributes and the situation development trend information, and the collaborative perception instruction is sent to the low-altitude intelligent network management terminal to trigger a linkage response operation.

[0010] On the other hand, the present invention also provides a target situation awareness system based on a low-altitude intelligent network, including a processor and a machine-readable storage medium, wherein the machine-readable storage medium is connected to the processor, the machine-readable storage medium is used to store programs, instructions or codes, and the processor is used to execute the programs, instructions or codes in the machine-readable storage medium to implement the above method.

[0011] Based on the above aspects, the present invention obtains a target perception data set containing three-dimensional position data and environmental disturbance data collected by multiple nodes in a collaborative manner, comprehensively considers various factors affecting the target situation, performs feature extraction processing on the target perception data set, generates spatial distribution characteristics and continuous change characteristics of the target object, and can accurately characterize the state of the target object from two dimensions of space and time. The spatial distribution characteristics and continuous change characteristics are input into the pre-trained situation perception model to generate situation correlation characteristics, and further obtain the current situation attributes and situation development trend information of the target object through the situation classification layer, thereby realizing accurate judgment and prediction of the target situation. Based on this information, a collaborative perception instruction containing a target positioning identifier is generated and sent to the low-altitude intelligent network management terminal to trigger a linkage response operation, so that the low-altitude intelligent network can make corresponding adjustments and response measures in a timely manner according to the real-time situation of the target, thereby improving the safety and management efficiency of the low-altitude field and enhancing the intelligence level and collaborative combat capability of the low-altitude intelligent network. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 It is a schematic diagram of the execution flow of the target situation awareness method based on the low-altitude intelligent network provided by an embodiment of the present invention.

[0013] Figure 2 It is a schematic diagram of exemplary hardware and software components of a target situation awareness system based on a low-altitude intelligent network provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0014] The present invention will be described in detail below with reference to the accompanying drawings. Figure 1 This is a flow chart of a target situation awareness method based on a low-altitude intelligent network provided by an embodiment of the present invention. The target situation awareness method based on a low-altitude intelligent network is introduced in detail below.

[0015] Step S110: Acquire a target perception data set collaboratively collected by multiple nodes in the low-altitude intelligent network, where the target perception data set includes three-dimensional position data and environmental disturbance data collected by different nodes in continuous time periods.

[0016] In this embodiment, in the application scenario of the low-altitude intelligent network, in order to accurately perceive the target's situation, a target perception data set is first acquired. This target perception data set includes three-dimensional position data and environmental disturbance data collected by different nodes over a continuous period of time. For example, in a low-altitude area, there are multiple targets. The motion state of these targets will be affected by the surrounding environment. By collaboratively collecting data, different nodes can comprehensively and continuously record the relevant information of the target objects.

[0017] Step S111: Determine multiple sensing nodes covering the target monitoring area in the low-altitude intelligent network, where the sensing nodes include positioning sensing equipment deployed on low-altitude aircraft and environmental monitoring equipment deployed on ground sites.

[0018] In the low-altitude intelligent network, in order to achieve comprehensive coverage of the target monitoring area, it is necessary to rationally arrange sensing nodes. Among them, the positioning sensing equipment deployed on low-altitude aircraft is characterized by mobility and flexibility, and can follow the movement of the target object in real time and collect its location information. The environmental monitoring equipment deployed at the ground station can stably monitor the environmental conditions around the target object. For example, in a vast low-altitude area, there are multiple low-altitude aircraft flying at different altitudes and areas. Each aircraft is equipped with positioning sensing equipment. At the same time, multiple environmental monitoring stations are set up at different locations on the ground. These sensing nodes together form a data collection network covering the target monitoring area.

[0019] Step S112: collecting three-dimensional position data of the target object in continuous time periods through the positioning sensor device, wherein the three-dimensional position data includes the spatial coordinate points of the target object in each time period.

[0020] Then, positioning sensing equipment is used to collect the target object's three-dimensional position data over a continuous period of time. Positioning sensing equipment is typically based on advanced positioning technologies, such as the Global Positioning System (GPS) or other high-precision positioning algorithms, and can accurately obtain the target object's position in three-dimensional space. Over a continuous period of time, the positioning sensing equipment continuously records the target object's spatial coordinate points, forming a continuous sequence of position data. For example, over a period of time, the positioning sensing equipment will collect the target object's three-dimensional position data at set intervals. This three-dimensional position data can be represented as a series of coordinate points (x1, y1, z1), (x2, y2, z2), etc. These coordinate points can clearly depict the target object's movement trajectory.

[0021] Step S113: collecting environmental disturbance data around the target object within the same continuous time period through the environmental monitoring device, wherein the environmental disturbance data includes airflow direction information and wind speed fluctuation information.

[0022] At the same time, environmental monitoring equipment collects environmental disturbance data around the target object over the same continuous time period. Environmental monitoring equipment is typically equipped with multiple sensors, such as wind direction sensors and wind speed sensors, which can accurately measure airflow direction and wind speed. Over a continuous period of time, environmental monitoring equipment continuously records airflow direction information and wind speed fluctuation information, which is important for analyzing the motion state and posture of the target object. For example, when the target object is in an area with strong airflow, its motion trajectory may be significantly affected. By analyzing environmental disturbance data, we can better understand the target object's motion patterns.

[0023] Step S114: performing synchronous alignment processing on the three-dimensional position data and the environmental disturbance data according to the acquisition timestamp to generate a perception data unit with a time correspondence relationship.

[0024] Afterwards, the collected three-dimensional position data and environmental disturbance data need to be synchronized and aligned according to the acquisition timestamp. Since the three-dimensional position data and environmental disturbance data are collected by different devices, there may be certain differences in their acquisition times. To ensure the accuracy and validity of the data, these data need to be aligned according to the acquisition timestamp so that each set of three-dimensional position data can correspond to the corresponding environmental disturbance data. For example, the positioning sensor device collects the three-dimensional position data (x1, y1, z1) of the target object at time t1, and the environmental monitoring device collects the airflow direction information A1 and wind speed fluctuation information V1 at time t1. By aligning these data according to the timestamp, a perception data unit with a time correspondence can be generated (t1, (x1, y1, z1), A1, V1).

[0025] Step S115: Classify and integrate the perception data units according to the node identifiers to generate perception data subsets corresponding to different nodes, wherein the perception data subsets include the three-dimensional position data and environmental disturbance data collected by the corresponding perception nodes.

[0026] In this embodiment, each sensing node has a unique node identifier. This node identifier can be used to classify sensing data units, allowing data belonging to the same node to be integrated. For example, sensing data units collected by positioning sensor equipment deployed on low-altitude aircraft A are grouped together to form a sensing data subset corresponding to that sensing node. Similarly, sensing data units collected by environmental monitoring equipment deployed at ground station B are also grouped together to form a sensing data subset corresponding to another sensing node. Thus, the sensing data subsets corresponding to different sensing nodes contain the three-dimensional position data and environmental disturbance data collected by that sensing node.

[0027] Step S116: Perform cross-node information fusion processing on all perception data subsets to generate a target perception data set containing multi-node collaborative information.

[0028] In the low-altitude intelligent network, data collected by different nodes is complementary. Cross-node information fusion can fully leverage this data to generate a target perception data set that incorporates multi-node collaborative information. For example, fusing three-dimensional position data collected by positioning sensors on low-altitude aircraft with environmental disturbance information collected by ground-based environmental monitoring equipment can provide a more comprehensive understanding of the target object's motion state and the impact of its surroundings. During the fusion process, various fusion algorithms, such as data association methods, can be employed to ensure that the fused data accurately reflects the target object's true condition.

[0029] Step S120: Perform feature extraction processing on the target perception data set to generate spatial distribution features and continuous change features of the target object. The spatial distribution features represent the position correlation relationship of the target object in the low-altitude area, and the continuous change features represent the state change law of the target object in a continuous time period.

[0030] After acquiring the target perception data set, this step requires feature extraction to extract the spatial distribution characteristics and continuous change characteristics of the target objects. Spatial distribution characteristics can reflect the positional relationships of target objects in the low-altitude area, such as the distance and relative position between target objects. Continuous change characteristics reflect the state changes of target objects over consecutive time periods, such as changes in speed and direction.

[0031] Step S121: Perform spatial relationship analysis on the three-dimensional position data in the target perception data set, and calculate the relative distance parameters and azimuth angle parameters between different target objects. The relative distance parameter represents the straight-line distance of the target object in the three-dimensional space, and the azimuth angle parameter represents the direction angle of the target object relative to the reference point.

[0032] First, the 3D position data in the target perception data set is spatially analyzed. In 3D space, the positional relationships between different target objects can be described by relative distance parameters and azimuth angle parameters. To calculate these parameters, the 3D position data of multiple target objects within the same time period must be extracted from the target perception data set.

[0033] Step S1211: extracting three-dimensional position data of multiple target objects in the same time period from the target perception data set, wherein the three-dimensional position data includes the spatial coordinate point of each target object.

[0034] Next, the 3D position data for multiple target objects within the same time period is extracted from the target perception data set. Since the target perception data set is collected over a continuous period of time, accurate analysis of the spatial relationships between target objects requires selecting data from the same time period. For example, at time t, multiple target objects are active in a low-altitude area. The positioning sensor equipment records the 3D position data for these target objects. Each 3D position data point contains the spatial coordinates (x, y, z) of the target object.

[0035] Step S1212: Select a spatial coordinate point of one of the target objects as a reference point.

[0036] In this embodiment, the selection of the reference point can be random or determined according to actual needs. For example, the coordinate point of a representative target object is selected as the reference point to better analyze the positional relationship of other target objects relative to the reference point.

[0037] Step S1213: For each other target object, extract its spatial coordinate point, and calculate the three-dimensional Euclidean distance between the spatial coordinate point and the reference point, and use the calculation result as the relative distance parameter between the target object and the reference point.

[0038] For each target object other than the one corresponding to the reference point, its spatial coordinates must be extracted and the 3D Euclidean distance between that coordinate point and the reference point calculated. 3D Euclidean distance is a common method for measuring the straight-line distance between two points in 3D space. For example, let the spatial coordinates of the reference point be (x0, y0, z0) and the spatial coordinates of another target object be (x1, y1, z1). Using the corresponding calculation method, the 3D Euclidean distance between these two points can be calculated. This distance is the relative distance parameter between the target object and the reference point.

[0039] Step S1214: Calculate the horizontal projection coordinates of the spatial coordinate point of the target object relative to the reference point, where the horizontal projection coordinates represent the position of the target object on the horizontal plane.

[0040] In this embodiment, the horizontal projection coordinates are the coordinates obtained by projecting the three-dimensional coordinates of the target object onto a horizontal plane. These coordinates can represent the position of the target object on the horizontal plane. For example, by ignoring the z coordinate in the three-dimensional coordinates (x, y, z) of the target object, its projection coordinates (x, y) on the horizontal plane are obtained. These projection coordinates are the horizontal projection coordinates of the target object relative to the reference point.

[0041] Step S1215: Calculate the azimuth angle between the horizontal projection coordinates, and use the calculation result as the azimuth angle parameter of the target object relative to the reference point.

[0042] Next, the azimuth angle between the horizontal projection coordinates is calculated. The azimuth angle represents the angle between the target object's orientation and the reference point. By performing corresponding calculations on the horizontal projection coordinates, the azimuth angle parameter of the target object relative to the reference point can be obtained. For example, using trigonometric functions or other methods, the azimuth angle can be calculated from the difference in the horizontal projection coordinates. This azimuth angle accurately describes the target object's orientation relative to the reference point.

[0043] Step S1216: Summarize and organize the relative distance parameters and azimuth angle parameters of all target objects to generate a spatial relationship analysis result including the spatial relationship of multiple target objects.

[0044] In this embodiment, the relative distance parameters and azimuth angle parameters of each target object relative to the reference point can be collected and integrated to form a spatial relationship analysis result containing the spatial relationship of multiple target objects. This spatial relationship analysis result can clearly show the positional relationship of different target objects in three-dimensional space.

[0045] Step S122: constructing a spatial distribution description structure based on the relative distance parameter and the azimuth angle parameter, wherein the spatial distribution description structure includes a position coordinate set and a position correlation matrix of the target object.

[0046] After calculating the relative distance parameters and azimuth angle parameters between different target objects, this step needs to construct a spatial distribution description structure based on the relative distance parameters and azimuth angle parameters. The spatial distribution description structure can more comprehensively represent the positional association relationship of target objects in the low-altitude area. Among them, the position coordinate set contains the three-dimensional position coordinates of all target objects. These coordinate points determine the specific position of the target object in the three-dimensional space; the position association matrix reflects the relative position relationship between the target objects. The numerical values of the matrix elements can represent the distance, azimuth and other information between the target objects. For example, the elements in the position association matrix can be assigned according to the relative distance parameters and azimuth angle parameters, so that the matrix can accurately describe the positional association between the target objects.

[0047] Step S123: Perform time period difference analysis on the three-dimensional position data in the target perception data set, and calculate the displacement and displacement direction change of the target object in adjacent time periods. The displacement represents the moving distance of the target object in unit time, and the displacement direction change represents the adjustment amplitude of the target object's moving direction.

[0048] In this embodiment, by comparing the three-dimensional position data of the target object in adjacent time periods, the target object's displacement and change in displacement direction can be calculated. The displacement reflects the distance the target object moves per unit time and can be calculated by the difference in position coordinates between adjacent time periods. The change in displacement direction reflects the adjustment in the target object's movement direction and can be determined, for example, by calculating the angle between the displacement vectors in adjacent time periods. For example, between time periods t1 and t2, the target object's position moves from (x1, y1, z1) to (x2, y2, z2). Using corresponding calculation methods, the target object's displacement and change in displacement direction during this period can be calculated.

[0049] Step S124: performing time period trend extraction processing on the environmental disturbance data to extract the change period of the airflow direction information within a continuous time period and the fluctuation frequency of the wind speed fluctuation information.

[0050] In this embodiment, the airflow direction and wind speed fluctuation information in the environmental disturbance data changes over time. By extracting the periodicity and frequency of these fluctuations within a continuous time period, we can understand the changing patterns of these environmental factors. For example, for airflow direction information, its periodicity can be analyzed over a period of time to determine its periodicity; for wind speed fluctuation information, its frequency can be calculated to understand wind speed variations. The changing patterns of these environmental factors can affect the motion state of the target object, so extracting this information is crucial for analyzing the target object's situation.

[0051] Step S125: performing feature fusion processing on the displacement amount, displacement direction change amount, change period and fluctuation frequency to generate a continuous change description structure, which includes the state change rate of the target object and the environmental impact correlation.

[0052] In this embodiment, the displacement, displacement direction change, change period, and fluctuation frequency reflect the state change of the target object and the influence of environmental factors from different perspectives. A more comprehensive continuous change description structure can be generated through fusion processing. During the fusion process, a weighted fusion method can be used to assign different weights according to the importance of each feature. For example, a higher weight can be assigned to the displacement and displacement direction change because they directly reflect the change in the motion state of the target object; for the change period and fluctuation frequency, appropriate weights can be assigned according to the degree of influence of environmental factors on the target object. The continuous change description structure generated after fusion includes the state change rate of the target object and the environmental influence correlation. The state change rate reflects the speed of the state change of the target object in a continuous period of time, and the environmental influence correlation indicates the degree of influence of environmental factors on the state change of the target object.

[0053] Step S126: using the spatial distribution description structure and the continuous change description structure as spatial distribution features and continuous change features respectively.

[0054] Finally, the spatial distribution description structure and the continuous change description structure are used as the target object's spatial distribution characteristics and continuous change characteristics, respectively. The spatial distribution characteristics can intuitively demonstrate the locational relationships of the target object in the low-altitude area; the continuous change characteristics reflect the state changes of the target object over a continuous period of time and the influence of environmental factors, which helps predict the target object's future situation. Through the spatial distribution characteristics and continuous change characteristics, a more comprehensive understanding of the target object's situation information can be obtained.

[0055] Step S130: inputting the spatial distribution features and the continuous change features into a pre-trained situation awareness model to generate situation association features of the target object.

[0056] After generating the spatial distribution and continuous variation features of the target object, this step requires inputting these features into the pre-trained situation awareness model to generate the target object's situation association features. These situation association features provide a deeper understanding of the target object's situational information. By combining spatial distribution and continuous variation features, the situation awareness model can process and analyze these features to uncover potential situational relationships within the target object.

[0057] Step S131: Input the spatial distribution features into the first coding layer of the situation awareness model, and perform dimensionality unification processing on the position coordinate set and position correlation matrix in the spatial distribution features through the first coding layer to generate spatial coding features with standard dimensional representation.

[0058] In this embodiment, the primary function of the first coding layer is to unify the dimensions of the location coordinate set and the location association matrix in the spatial distribution features. Because the original dimensions of the location coordinate set and the location association matrix may differ, they need to be unified to a standard dimension to facilitate model processing and analysis. For example, the location coordinate set can be dimensionalized by expanding or compressing it to match the dimensions of the location association matrix; the location association matrix can also be normalized to ensure that the value range and dimensionality of its elements meet the model requirements. After dimensional unification, spatial coding features with a standard dimensional representation are generated, which can be more effectively utilized by the model.

[0059] Step S132: Input the continuously changing feature into the second coding layer of the situation awareness model, and perform time alignment processing on the state change rate and environmental impact correlation in the continuously changing feature through the second coding layer to generate a change coding feature with a time correspondence.

[0060] In this embodiment, the second coding layer needs to perform time alignment processing on the state change rate and environmental impact correlation in the continuously changing features. Since the state change rate and environmental impact correlation are collected in a continuous time period, there may be differences in their time series. In order to ensure the accuracy and consistency of the data, they need to be aligned in chronological order so that the state change rate and environmental impact correlation have a corresponding relationship at the same time point. For example, through interpolation, resampling and other methods, the time series of the state change rate and environmental impact correlation are adjusted to synchronize them in time. After the time alignment processing, a change coding feature with a time correspondence is generated, which can better reflect the state change and environmental impact of the target object in a continuous time period.

[0061] Step S133: Input the spatial coding features and the change coding features into the weight allocation unit of the association analysis module of the situation awareness model to dynamically calculate the association weights of the spatial coding features and the change coding features, and the association weights are used to represent the contribution of the spatial coding features and the change coding features to the situation association features.

[0062] In this embodiment, the role of the weight allocation unit is to dynamically calculate the association weights of the spatial coding features and the change coding features. The association weights represent the contribution of the spatial coding features and the change coding features to the situation association features, and different weight values reflect the importance of these two features in the situation association feature generation process. For example, in some cases, the spatial coding features may contribute more to the situation association features, and the weight allocation unit will assign a higher weight to them accordingly; in other cases, the change coding features may be more critical, and the weight allocation unit will assign a higher weight to them. The weight allocation unit determines the appropriate association weight by analyzing and processing the input spatial coding features and change coding features.

[0063] Step S1331: inputting the spatial coding features and the change coding features into the weight distribution unit, wherein the weight distribution unit includes a fully connected layer and an activation function layer.

[0064] In this embodiment, the weight distribution unit is composed of a fully connected layer and an activation function layer. The fully connected layer can perform a linear transformation on the input features and map the input features to a new feature space; the activation function layer performs a nonlinear transformation on the output of the fully connected layer to make it have stronger expressive power. For example, the fully connected layer linearly combines the spatial encoding features and the change encoding features according to the preset weight matrix, and performs a weighted summation on each dimension of the input features to obtain a new feature representation. Then, the activation function layer processes the output of the fully connected layer. Common activation functions include Sigmoid function, ReLU function, etc. Through the action of the activation function, the value range and distribution of the features are more in line with the requirements of the model.

[0065] Step S1332: performing linear transformation processing on the spatial coding feature and the change coding feature through the fully connected layer to generate a feature importance evaluation vector.

[0066] When the fully connected layer performs a linear transformation on the input spatial encoding features and change encoding features, it can perform a weighted summation of each dimension of the features according to pre-set weight parameters. Assume that the spatial encoding feature is represented as A and the change encoding feature is represented as B, and the weight matrices of the fully connected layer are W_A and W_B, respectively. For each dimension of the spatial encoding feature A, you can multiply it by the corresponding element in the weight matrix W_A, and then add all the products; the same operation is performed for the change encoding feature B. Through this linear combination, a new vector is obtained, which contains preliminary information on the importance of each dimension of the spatial encoding feature and the change encoding feature. This vector is defined as the feature importance evaluation vector. For example, the spatial encoding feature A has multiple dimensions (a_1, a_2, …, a_n), and the weight matrix W_A also has multiple elements (w_a1, w_a2, …, w_an). Then the result after linear transformation is a_1*w_a1+a_2*w_a2+ …+a_n*w_an. Similar calculations are performed on the change encoding feature B, and finally the two results are merged into a feature importance evaluation vector.

[0067] Step S1333: Normalize the feature importance evaluation vector through the activation function layer to generate a first weight value corresponding to the spatial coding feature and a second weight value corresponding to the change coding feature, where the sum of the first weight value and the second weight value is 1.

[0068] After the activation function layer receives the feature importance evaluation vector, it can be normalized. The purpose of normalization is to make the generated weight value within a reasonable range and satisfy the sum of the first weight value corresponding to the spatial coding feature and the second weight value corresponding to the change coding feature is 1. For example, if the Sigmoid activation function is used, it can map each element in the feature importance evaluation vector to a range between 0 and 1. Then, through the set calculation method, the result after the Sigmoid function processing is adjusted so that the sum of the first weight value and the second weight value is equal to 1. Specifically, assuming that the value related to the spatial coding feature obtained after the Sigmoid function processing is S_A and the value related to the change coding feature is S_B, then the first weight value can be calculated by S_A / (S_A+S_B), and the second weight value can be calculated by S_B / (S_A+S_B), so that the sum of the two weight values is guaranteed to be 1.

[0069] Step S1334: Using the first weight value as the associated weight of the spatial coding feature, and using the second weight value as the associated weight of the change coding feature; the weight allocation unit adjusts the contribution ratio of the spatial coding feature and the change coding feature in the fusion process according to the first weight value and the second weight value.

[0070] The calculated first weight value is used as the associated weight of the spatial coding feature, and the second weight value is used as the associated weight of the change coding feature. In the subsequent feature fusion process, the weight allocation unit will adjust the contribution ratio of the spatial coding feature and the change coding feature according to these two weight values. For example, if the first weight value is larger, it means that the spatial coding feature is more important in the generation process of the situation-related feature, then the spatial coding feature will be given a higher weight during fusion, so that it has a greater impact on the final result; conversely, if the second weight value is larger, the contribution of the change coding feature will be greater. In this way, the role of the two features in the fusion process can be dynamically adjusted according to different situations to generate more accurate situation-related features.

[0071] Step S134: performing weighted fusion processing on the spatial coding feature and the change coding feature based on the association weight to generate a fused association feature.

[0072] After obtaining the association weights for the spatial coding features and the change coding features, they need to be weightedly fused. The weighted fusion method, whether weighted addition or weighted concatenation, depends on the nature of the features and actual needs. If the spatial coding features and the change coding features are complementary in terms of dimensionality and meaning, and their values can be combined within the same dimension, weighted addition can be used. For example, the value of each dimension of the spatial coding feature is multiplied by its corresponding association weight (first weight value), and the value of each dimension of the change coding feature is multiplied by its corresponding association weight (second weight value). The values of the corresponding dimensions are then added together to obtain the fused feature value. If the dimensionality and meaning of the two features differ significantly, making it inappropriate to add them within the same dimension, weighted concatenation is more appropriate. This involves concatenating the weighted spatial coding features and change coding features in a predetermined order to form a fused association feature. Weighted fusion effectively integrates the information of the spatial coding features and the change coding features, highlighting the role of important features.

[0073] Step S135: performing information enhancement processing on the fused association features through the feature enhancement unit of the association analysis module to generate situation association features of the target object.

[0074] The feature enhancement unit of the association analysis module will perform information enhancement processing on the fused association features. The feature enhancement unit can use a variety of methods to achieve information enhancement, such as using convolution operations, pooling operations, etc. The convolution operation can perform local feature extraction on the fused association features, and extract local feature information of different scales and directions by sliding different convolution kernels on the features. The pooling operation can perform dimensionality reduction processing on the features while retaining important feature information and reducing data redundancy. For example, by using the maximum pooling operation, the maximum value can be selected in the local area of the feature as the representative value of the area, which can highlight the important information in the feature. Through the processing of the feature enhancement unit, the potential information in the fused association features can be further excavated, the expressive ability of the features can be enhanced, and the generated situation-related features of the target object can be more accurate and rich, and can better reflect the real situation information of the target object.

[0075] Step S140: inputting the situation association features into the situation classification layer of the situation awareness model to obtain the current situation attributes and situation development trend information of the target object.

[0076] After generating the target object's situation-related features, this step requires inputting these features into the situation-awareness model's situation classification layer to obtain information about the target object's current situation attributes and its development trends. The situation classification layer analyzes and determines the situation-related features, determining the target object's current situation type and predicting its future development trends.

[0077] Step S141: Input the situation association features into the situation classification layer of the situation awareness model, perform attribute matching processing on the situation association features through the situation classification layer, generate a situation attribute candidate set of the target object and the matching confidence corresponding to each situation type identifier in the situation attribute candidate set, wherein the situation attribute candidate set contains multiple candidate situation type identifiers.

[0078] After the situation-related features are input into the situation classification layer, the situation classification layer will perform attribute matching on them. The situation classification layer pre-stores a variety of situation type identifiers and the feature patterns corresponding to each situation type. By comparing and matching the input situation-related features with these pre-stored feature patterns, the situation type that is most similar to the situation-related features is found, forming a candidate set of situation attributes of the target object. At the same time, the matching confidence is calculated for each candidate situation type identifier, and the matching confidence indicates the degree of matching between the situation type and the input situation-related features. For example, a similarity calculation method is used to calculate the similarity between the situation-related features and each pre-stored feature pattern, and the similarity value is used as the matching confidence. If the feature pattern of a situation type has a high degree of similarity with the situation-related features, then its corresponding matching confidence will be larger, indicating that the situation type is more likely to be the situation type that the target object is currently in.

[0079] Step S142: Select the situation type identifier with the highest matching confidence as the current situation attribute of the target object.

[0080] After obtaining the candidate set of situation attributes and the matching confidence levels corresponding to each situation type identifier, the situation type identifier with the highest matching confidence level is selected as the current situation attribute of the target object. This is because the highest matching confidence level indicates that the situation type most closely matches the input situation-related features and most accurately reflects the current situation of the target object. For example, the situation attribute candidate set contains several different situation type identifiers, such as Type A, Type B, and Type C, and their corresponding matching confidence levels are R_A, R_B, and R_C, respectively. By comparing these matching confidence levels, the situation type identifier corresponding to the highest matching confidence level is selected as the current situation attribute of the target object.

[0081] Step S143: Input the situation-related features into the tendency prediction layer of the situation awareness model, perform historical information backtracking processing on the situation-related features through the tendency prediction layer, extract the situation change pattern of the target object in the historical period, and perform tendency prediction processing on the current situation attributes for the future period based on the situation change pattern, and generate situation development tendency information including direction indication and rate indication, wherein the direction indication indicates the possible evolution direction of the target situation, and the rate indication indicates the speed of the target situation evolution.

[0082] Situation-related features are input into the tendency prediction layer of the situation awareness model. The tendency prediction layer first performs a retrospective analysis of the situation-related features. Because situation-related features contain information about the target object's situation over a period of time, by analyzing and mining this information, it can extract patterns of situation changes within the target object's historical period. For example, the tendency prediction layer performs time series analysis on the situation-related features within the historical period, observing information such as feature change trends and periodicity to determine the historical evolution of the target object's situation. Then, based on the extracted situation change patterns, it performs a trend prediction for the target object's current situation attributes for future periods. The tendency prediction layer combines the current situation-related features with historical situation change patterns to predict the likely direction and speed of the target object's situation evolution in the future. The direction indicator indicates the likely direction of the target's situation development, such as strengthening, weakening, or transition. The rate indicator indicates the speed of the target's situation evolution, such as rapid or slow evolution.

[0083] Step S1431: extracting a correlation feature sequence of a historical period according to the situation correlation feature, wherein the correlation feature sequence includes situation correlation features of the target object in multiple historical periods.

[0084] First, based on the situation correlation features, a correlation feature sequence for each historical period is extracted. Situation correlation features are collected and generated over a period of time and contain information about the target object's situation during different historical periods. By dividing and extracting the situation correlation features along the time dimension, we can obtain situation correlation features for the target object across multiple historical periods. These features are then arranged in chronological order to form a correlation feature sequence. For example, assuming the situation correlation features were collected over a long period of time, this period can be divided into multiple sub-periods, each corresponding to a situation correlation feature. Arranging the situation correlation features of these sub-periods in sequence yields a correlation feature sequence. This correlation feature sequence can reflect the changes in the target object's situation over the historical period.

[0085] Step S1432: Inputting the associated feature sequence into the tendency prediction layer, the tendency prediction layer comprises a long short-term memory network structure and a pattern extraction structure.

[0086] The associated feature sequence is input into the trend prediction layer. The trend prediction layer consists of a long-short-term memory (LSTM) network structure and a pattern extraction structure. The long-short-term memory (LSTM) network structure is a special recurrent neural network architecture that can process sequential data and effectively address the vanishing and exploding gradient problems found in traditional recurrent neural networks. The LSTM structure has memory cells that retain long-term information when processing sequential data, making it very effective for extracting information about trend changes over a historical period. The pattern extraction structure is responsible for further analyzing and processing the LSTM structure's output to extract patterns of trend changes in the target object over a historical period.

[0087] Step S1433: performing time series information storage processing on the associated feature sequence through the long short-term memory network structure to generate a memory state vector containing historical situation information.

[0088] The Long Short-Term Memory (LSTM) network architecture processes temporal information stored in the input feature sequence. The LSTM architecture consists of an input gate, a forget gate, an output gate, and a memory cell. When processing the feature sequence, the input gate determines which new information enters the memory cell, the forget gate determines which old information is forgotten, and the output gate determines which information is output from the memory cell. Through the control of these gates, the LSTM architecture can store historical situation information in the memory cells while processing sequential data. As the feature sequence is processed, the LSTM architecture continuously updates the state of the memory cells, ultimately generating a memory state vector containing historical situation information. This memory state vector contains important situation information about the target object during the historical period.

[0089] Step S1434: The memory state vector is subjected to feature decomposition processing through the pattern extraction structure to separate the stable situation pattern and the fluctuating situation pattern of the target object in the historical period. The stable situation pattern represents the situation characteristics maintained for a long time by the target object, and the fluctuating situation pattern represents the situation characteristics of the target object with short-term changes.

[0090] The pattern extraction structure performs feature decomposition on the memory state vector. This is because the memory state vector contains a variety of situational information about the target object over a historical period, including both relatively stable long-term situational features and short-term fluctuations. The pattern extraction structure uses appropriate methods, such as principal component analysis and cluster analysis, to decompose the memory state vector. Principal component analysis reduces the dimensionality of the features in the memory state vector and extracts the main characteristic components, which reflect the stable situational patterns of the target object. Cluster analysis classifies the features in the memory state vector, grouping similar features together and separating stable and fluctuating situational patterns. Stable situational patterns represent situational features that persist over a long period of time, such as the target object's basic motion patterns and long-term situational states. Fluctuating situational patterns represent short-term changes in the target object's situation, such as sudden fluctuations and temporary adjustments. By separating stable and fluctuating situational patterns, we can more clearly understand the target object's situational changes over a historical period.

[0091] Step S1435: taking the stable situation pattern and the fluctuating situation pattern as the situation change pattern of the target object in the historical period.

[0092] The isolated stable and fluctuating situation patterns are used as patterns of the target object's situation changes over a historical period. These two patterns comprehensively reflect the target object's situation evolution over that period. The stable situation pattern reflects the target object's long-term situation characteristics, while the fluctuating situation pattern reflects its short-term situation changes. By comprehensively considering both stable and fluctuating situation patterns, the trend prediction layer can more accurately predict the trend of the target object's current situation attributes for future periods, generating more reliable information on its situation development trends.

[0093] Step S144: taking the current situation attributes and the situation development tendency information as the target situation determination result.

[0094] Finally, the target's current situation attributes and situation development trends are used as the target situation determination result. This target situation determination result includes the target's current situation type and possible future development trends. For example, in the application scenario of low-altitude intelligent networking, the target situation determination result can be used to promptly implement appropriate measures, such as adjusting monitoring strategies and issuing early warnings, to ensure effective management and control of the target.

[0095] Step S150: Generate a collaborative perception instruction containing a target positioning identifier based on the current situation attributes and the situation development trend information, and send the collaborative perception instruction to the low-altitude intelligent network management terminal to trigger a linkage response operation.

[0096] After obtaining the target's current situation attributes and development trends, this step requires generating a collaborative sensing instruction containing the target's location identifier based on this information. This instruction is then sent to the low-altitude intelligent network management terminal to trigger a coordinated response. The generation of collaborative sensing instructions requires comprehensive consideration of the target's current situation and future development trends, and includes an accurate target location identifier, allowing the management terminal to accurately monitor and respond to the target according to the instruction.

[0097] Step S151: parse the preset response rule library corresponding to the current situation attribute, and extract the response level identifier and collaborative strategy code associated with the current situation attribute. The response level identifier is used to indicate the urgency of the response to be triggered, and the collaborative strategy code is used to indicate the node type that needs to be linked.

[0098] First, parse the preset response rule library corresponding to the current situation attributes. The preset response rule library is pre-established and contains response rules and strategies corresponding to different situation attributes. By matching and parsing the current situation attributes, the response level identifier and collaborative strategy code associated with the situation attribute can be extracted from the preset response rule library. The response level identifier indicates the urgency of the response that needs to be triggered. For example, it may be divided into different levels, such as high urgency, medium urgency, low urgency, etc. Different urgency levels correspond to different response measures and resource allocation. The collaborative strategy code is used to indicate the type of node that needs to be linked, such as whether it is necessary to link the positioning sensor equipment on the low-altitude aircraft, the environmental monitoring equipment on the ground station, or both. By extracting the response level identifier and collaborative strategy code, clear response strategy and linkage information can be provided for the subsequent generation of collaborative perception instructions.

[0099] Step S152: Based on the three-dimensional position data in the target perception data set, calculate the position centroid coordinates and position coverage area boundaries of the target object in the current time period, the position centroid coordinates represent the center position of the target object, and the position coverage area boundaries represent the spatial distribution range of the target object.

[0100] Based on the 3D position data in the target perception data set, the position centroid coordinates and position coverage area boundaries of the target object in the current time period are calculated. The position centroid coordinates can be obtained by taking a weighted average of all the 3D position data of the target object in the current time period. Assuming that the target object has multiple 3D position coordinates in the current time period (x_1, y_1, z_1), (x_2, y_2, z_2), …, (x_n, y_n, z_n), the x, y, and z values of each coordinate can be weighted summed and then divided by the number of coordinates n to obtain the position centroid coordinates (x_c, y_c, z_c). The position coverage area boundaries can be calculated using various methods. For example, a bounding box is defined by determining the maximum and minimum values of all position coordinates. The boundaries of this bounding box can approximate the position coverage area boundaries of the target object. Alternatively, more complex methods, such as the convex hull algorithm, can be used to obtain a more accurate position coverage area boundary by finding the convex hull of all position coordinates. The position centroid coordinates represent the center position of the target object, and the position coverage area boundary represents the spatial distribution range of the target object.

[0101] Step S153: Process the coordinate system of the position centroid and the boundary of the position coverage area to generate a standardized positioning identifier consistent with the low-altitude intelligent network geographic coordinate system, and the standardized positioning identifier includes a center coordinate point and a boundary coordinate point sequence of the target object.

[0102] The calculated position centroid coordinates and position coverage area boundaries are processed by coordinate system one. Since the three-dimensional position data in the target perception data set may be collected based on different coordinate systems, and the low-altitude intelligent network has its own geographic coordinate system, in order to ensure the accuracy and consistency of the positioning identification, the position centroid coordinates and position coverage area boundaries need to be converted to a coordinate system consistent with the low-altitude intelligent network geographic coordinate system. Coordinate system one processing can be achieved through a series of coordinate transformation operations, such as translation, rotation, and scaling. First, determine the transformation parameters between the two coordinate systems. These parameters can be calculated using known control points. Control points are points that have accurate coordinates in both coordinate systems. Through the coordinate correspondence of these control points, the translation parameters, rotation parameters, and scaling parameters can be calculated. Then, each coordinate point of the position centroid coordinates and the position coverage area boundary is transformed accordingly according to these transformation parameters to obtain new coordinates in the low-altitude intelligent network geographic coordinate system. After processing the coordinate system, a standardized positioning identifier is generated that is consistent with the low-altitude intelligent network geographic coordinate system. This standardized positioning identifier contains the target object's center coordinate point (i.e., the coordinates after the location centroid coordinates are converted) and a sequence of boundary coordinate points (i.e., the coordinate sequence after the location coverage area boundaries are converted). This standardized positioning identifier can accurately represent the target object's location and spatial distribution range in the low-altitude intelligent network geographic coordinate system.

[0103] Step S154: combining the direction indication and the speed indication in the situation development tendency information, performing position prediction processing on the standardized positioning identifier for future time periods, and generating a predicted positioning identifier sequence of the target object in subsequent time periods.

[0104] The standardized positioning identifier is used to perform position prediction for future time periods, combining the direction and rate indicators in the situation development trend information. The direction indicator in the situation development trend information indicates the possible direction of the target situation evolution, while the rate indicator indicates the speed of the target situation evolution. Based on this information, the target object's movement direction and speed in the future time period can be inferred. For the sequence of center coordinate points and boundary coordinate points in the standardized positioning identifier, the direction of each coordinate point's possible movement in the future time period is determined based on the direction indicator, and the distance each coordinate point is likely to move per unit time is determined based on the rate indicator. For example, assuming the center coordinate point is (x0, y0, z0), the direction indicator indicates that the target object will move in a specific direction, and the rate indicator indicates that its movement speed is a certain value. Then, within a future time period t, the new coordinates (x1, y1, z1) of the center coordinate point after that time period can be calculated based on the direction and rate. The same process is performed for each point in the boundary coordinate point sequence. By continuously predicting the position of each time period, a predicted positioning identifier sequence for the target object in subsequent time periods is generated. This predicted positioning identifier sequence contains the predicted center coordinate point and predicted boundary coordinate point sequences for the target object in multiple future time periods.

[0105] Step S155: The response level identifier, collaborative strategy code, standardized positioning identifier, and predicted positioning identifier sequence are integrated and processed to generate a collaborative perception instruction containing a positioning identifier chain with timestamp alignment, wherein the coordinate dimension of each identifier node in the positioning identifier chain is consistent with the coordinate dimension of the standardized positioning identifier.

[0106] The response level identifier, collaborative strategy code, standardized positioning identifier, and predicted positioning identifier sequence are integrated. The purpose of this information integration is to combine these different types of information into a complete collaborative perception instruction, enabling the low-altitude intelligent network management terminal to accurately execute coordinated response operations based on this instruction. First, a timestamp is added to each identifier node in the predicted positioning identifier sequence, indicating the future time period corresponding to that identifier node. Then, the response level identifier, collaborative strategy code, standardized positioning identifier, and the timestamp-added predicted positioning identifier sequence are integrated according to a predefined format. During the integration process, the coordinate dimensions of each identifier node in the positioning identifier chain must be consistent with those of the standardized positioning identifier to ensure information consistency and accuracy. For example, the response level identifier and collaborative strategy code can be used as the instruction header information, the standardized positioning identifier as the current positioning information, and the predicted positioning identifier sequence, arranged in chronological order, as the positioning information for the future time period. This forms a collaborative perception instruction containing a timestamp-aligned positioning identifier chain. This collaborative perception instruction clearly conveys information such as the target object's current location, its possible future locations, the required response strategy, and the type of coordinated node.

[0107] Step S156: Send the collaborative sensing instruction to the management terminal, so that the management terminal schedules the corresponding node to perform target monitoring or response operations according to the collaborative sensing instruction.

[0108] The generated collaborative sensing command is sent to the low-altitude intelligent network management terminal. In the low-altitude intelligent network, the management terminal is responsible for the management and scheduling of the entire network. Upon receiving the collaborative sensing command, the management terminal can perform appropriate processing based on the information contained in the collaborative sensing command. First, the management terminal parses the collaborative sensing command, extracting information such as the response level identifier, collaborative strategy code, standardized positioning identifier, and predicted positioning identifier sequence. The response level identifier is then used to determine the urgency of the response to be triggered, and the collaborative strategy code is used to determine the node types to be coordinated. For example, if the response level is high urgency, the management terminal will quickly deploy more resources and nodes to participate in target monitoring and response operations. If the collaborative strategy code indicates the need to coordinate positioning sensors on low-altitude aircraft and environmental monitoring equipment at ground stations, the management terminal will send corresponding commands to these nodes. Next, based on the standardized positioning identifier and predicted positioning identifier sequence, the management terminal will dispatch the corresponding nodes to accurately monitor and respond to the target object. At the current moment, the node will locate and monitor the target object based on the standardized positioning identifier. For future periods, the node will make advance preparations based on the predicted positioning identifier sequence to promptly track changes in the target object's position. Through the above method, the low-altitude intelligent network management terminal can effectively dispatch the corresponding nodes to perform target monitoring or response operations according to the collaborative perception instructions, thereby realizing situational awareness and effective management of low-altitude targets.

[0109] Furthermore, in order for the situation awareness model to accurately classify and predict the target object's situation, it needs to be trained. The training process of the situation awareness model is a complex process, involving the adjustment and optimization of multiple modules and layers.

[0110] First, collect a large amount of sample data. This sample data should contain various situational information about the target object. This sample data can be selected from historical target perception data sets or generated through simulation experiments. The sample data must include the target object's spatial distribution characteristics, continuous change characteristics, and corresponding true situation attributes and situation development trends. For example, collect motion data of different target objects in different environments, calculate their spatial distribution characteristics and continuous change characteristics, and determine their true situation attributes (such as normal flight, abnormal flight, etc.) and situation development trends (such as situation strengthening, situation weakening, etc.) through manual annotation or other reliable methods. Then, divide the sample data into training, validation, and test sets. The training set is used to learn model parameters, the validation set is used to adjust model hyperparameters during training, and the test set is used to evaluate the performance of the trained model. The proportion of these divisions can be adjusted based on actual conditions. Generally, the training set accounts for a larger proportion, while the validation and test sets account for smaller proportions.

[0111] On this basis, a situational awareness model is constructed, comprising multiple essential modules and layers. The situational awareness model primarily consists of a first encoding layer, a second encoding layer, an association analysis module, a situation classification layer, and a tendency prediction layer. The first encoding layer unifies the dimensions of spatially distributed features. It can employ structures such as fully connected layers to convert the input location coordinate set and location association matrix of spatially distributed features into spatially encoded features with a standard dimensional representation. The second encoding layer performs temporal alignment on continuously changing features. It can also employ structures such as fully connected layers or recurrent neural network layers to convert the state change rate and environmental impact correlation of the input continuously changing features into change encoding features with a time-corresponding relationship. The association analysis module includes a weight allocation unit and a feature enhancement unit. The weight allocation unit dynamically calculates the association weights between spatial encoding features and change encoding features. It can be composed of fully connected layers and activation function layers. The feature enhancement unit enhances the information of the fused association features and can employ structures such as convolutional layers and pooling layers. The situation classification layer performs attribute matching on the situation association features to determine the current situation attributes of the target object. It can employ structures such as fully connected layers and a softmax activation function. The tendency prediction layer is used to process historical information backtracking and future tendency prediction based on situation-related features. It can include long-short-term memory network structures and pattern extraction structures. Each module and layer is connected according to the predefined connection relationships. For example, the output of the first and second coding layers serves as the input of the association analysis module, which in turn serves as the input of the situation classification layer and the tendency prediction layer.

[0112] The training set data is then input into the constructed situation awareness model for training. During training, an appropriate loss function is used to measure the difference between the model's output and the true labels. For the situation classification layer, the cross-entropy loss function can be used. This loss function measures the difference between the model's predicted situation attributes and the true situation attributes, ensuring that the model's predictions are closer to the ground truth. For the tendency prediction layer, the mean squared error loss function can be used. This loss function measures the difference between the model's predicted situation development tendency information (such as direction and rate) and the true situation development tendency information, enabling the model to more accurately predict future situation changes of the target object. An optimization algorithm is then used to update the model's parameters. Common optimization algorithms include stochastic gradient descent and the Adam optimization algorithm. Based on the gradient information of the loss function, the optimization algorithm continuously adjusts the model's parameters to gradually reduce the loss function value, thereby narrowing the difference between the model's output and the true labels. During the training process, several training parameters must be set, such as the learning rate, batch size, and number of training rounds. The learning rate controls the step size of parameter updates, the batch size indicates the number of samples fed into the model at a time, and the number of training epochs indicates how many times the model is trained on the entire training set. Proper training parameter settings are crucial to model training effectiveness, and determining the optimal parameter combination requires continuous experimentation and adjustment.

[0113] During training, the model is regularly validated using validation data. The purpose of validation is to evaluate the model's performance on unseen data and prevent overfitting. Overfitting occurs when a model performs well on the training set but poorly on the validation or test set. The validation set evaluation results can be used to adjust model hyperparameters, such as the learning rate and batch size, to optimize performance. For example, if the model's loss function value on the validation set stops decreasing or begins to increase, this indicates overfitting. In this case, methods such as reducing the learning rate or increasing the regularization term can be used to mitigate the overfitting issue. Furthermore, based on the validation set evaluation results, the model parameters that performed best during training can be selected as the final model parameters.

[0114] After the model training is completed, the test set data is used to conduct a final test on the model. The test set data is data that the model has never seen during the training and verification process, and can more realistically evaluate the generalization ability of the model. Through the evaluation results of the test set, the model's performance indicators such as accuracy, recall rate, F1 value, etc. can be obtained. These indicators can comprehensively reflect the accuracy and reliability of the model's classification and prediction of the target object situation. If the test performance of the model meets the requirements, the trained model can be applied to the actual low-altitude intelligent network target situation awareness mission; if the test performance of the model does not meet the requirements, it is necessary to readjust the model structure, training parameters or increase sample data, etc., and train and test again until the model performance reaches a satisfactory level.

[0115] Figure 2 A schematic diagram illustrates exemplary hardware and software components of a low-altitude intelligent network-based target situational awareness system 100 that can implement the concepts of the present invention, as provided in some embodiments of the present invention. For example, a processor 120 can be used in the low-altitude intelligent network-based target situational awareness system 100 to perform the functions of the present invention.

[0116] The target situation awareness system 100 based on the low-altitude intelligent network can be a general-purpose server or a special-purpose server, both of which can be used to implement the target situation awareness method based on the low-altitude intelligent network of the present invention. Although the present invention only shows a single server, for convenience, the functions described in the present invention can be implemented in a distributed manner on multiple similar platforms to balance the processing load.

[0117] For example, the target situation awareness system 100 based on the low-altitude intelligent network may include a network port 110 connected to the network, one or more processors 120 for executing program instructions, a communication bus 130, and storage media 140 in different forms, such as a disk, ROM, or RAM, or any combination thereof. Exemplarily, the target situation awareness system 100 based on the low-altitude intelligent network may also include program instructions stored in ROM, RAM, or other types of non-transitory storage media, or any combination thereof. The method of the present invention can be implemented according to these program instructions. The target situation awareness system 100 based on the low-altitude intelligent network also includes an I / O interface 150 between the computer and other input and output devices.

[0118] For ease of explanation, only one processor is described in the target situation awareness system 100 based on the low-altitude intelligent network. However, it should be noted that the target situation awareness system 100 based on the low-altitude intelligent network in the present invention can also include multiple processors, so the steps performed by one processor described in the present invention can also be performed jointly or individually by multiple processors. For example, if the processor of the target situation awareness system 100 based on the low-altitude intelligent network executes step A and step B, it should be understood that step A and step B can also be executed jointly by two different processors or individually in one processor. For example, the first processor executes step A, the second processor executes step B, or the first processor and the second processor execute steps A and B together.

[0119] In addition, an embodiment of the present invention also provides a readable storage medium, in which computer-executable instructions are preset. When the processor executes the computer-executable instructions, the above-mentioned target situation awareness method based on the low-altitude intelligent network is implemented.

[0120] It should be noted that in order to simplify the description of the present invention and thus help understand one or more embodiments of the invention, in the foregoing description of the embodiments of the present invention, multiple features are sometimes combined into one embodiment, figure or description thereof.

Claims

1. A target situation awareness method based on low-altitude intelligent network, characterized in that: The method comprises: Acquire a target perception data set collaboratively collected by multiple nodes in a low-altitude intelligent network, wherein the target perception data set includes three-dimensional position data and environmental disturbance data collected by different nodes in continuous time periods; Performing feature extraction processing on the target perception data set to generate spatial distribution features and continuous change features of the target object, wherein the spatial distribution features represent the position correlation relationship of the target object in the low-altitude area, and the continuous change features represent the state change law of the target object in a continuous time period; Inputting the spatial distribution features and the continuous change features into a pre-trained situation awareness model to generate situation-related features of the target object; Inputting the situation association features into the situation classification layer of the situation awareness model to obtain the current situation attributes and situation development trend information of the target object; Generate a collaborative sensing instruction including a target positioning identifier according to the current situation attributes and the situation development trend information, and send the collaborative sensing instruction to the low-altitude intelligent network management terminal to trigger a linkage response operation; The performing feature extraction processing on the target perception data set to generate spatial distribution features and continuous change features of the target object includes: Performing spatial relationship analysis on the three-dimensional position data in the target perception data set to calculate relative distance parameters and azimuth angle parameters between different target objects, wherein the relative distance parameter represents the straight-line distance between the target objects in three-dimensional space, and the azimuth angle parameter represents the direction angle of the target objects relative to the reference point; Constructing a spatial distribution description structure based on the relative distance parameter and the azimuth angle parameter, wherein the spatial distribution description structure includes a position coordinate set and a position correlation matrix of the target object; Performing time period difference analysis on the three-dimensional position data in the target perception data set to calculate the displacement and displacement direction change of the target object in adjacent time periods, wherein the displacement represents the movement distance of the target object per unit time, and the displacement direction change represents the adjustment amplitude of the target object's movement direction; Performing time period trend extraction processing on the environmental disturbance data to extract the change period of the airflow direction information within a continuous time period and the fluctuation frequency of the wind speed fluctuation information; Performing feature fusion processing on the displacement, displacement direction change, change period, and fluctuation frequency to generate a continuous change description structure, wherein the continuous change description structure includes the state change rate of the target object and the environmental impact correlation; The spatial distribution description structure and the continuous change description structure are used as spatial distribution features and continuous change features respectively.

2. The target situation awareness method based on low-altitude intelligent network according to claim 1 is characterized in that: The method of obtaining a target perception data set collaboratively collected by multiple nodes in the low-altitude intelligent network includes: Determine multiple sensing nodes in the low-altitude intelligent network that cover the target monitoring area, wherein the sensing nodes include positioning sensing equipment deployed on low-altitude aircraft and environmental monitoring equipment deployed on ground stations; Collecting three-dimensional position data of the target object in continuous time periods by the positioning sensing device, wherein the three-dimensional position data includes the spatial coordinate point of the target object in each time period; Collecting environmental disturbance data around the target object by the environmental monitoring device within the same continuous time period, wherein the environmental disturbance data includes airflow direction information and wind speed fluctuation information; Synchronously aligning the three-dimensional position data and the environmental disturbance data according to acquisition timestamps to generate perception data units with a time correspondence relationship; Classifying and integrating the perception data units according to node identifiers to generate perception data subsets corresponding to different nodes, wherein the perception data subsets include three-dimensional position data and environmental disturbance data collected by the corresponding perception nodes; All perception data subsets are subjected to cross-node information fusion processing to generate a target perception data set containing multi-node collaborative information.

3. The target situation awareness method based on low-altitude intelligent network according to claim 1 is characterized in that: The performing spatial relationship analysis on the three-dimensional position data in the target perception data set to calculate relative distance parameters and azimuth angle parameters between different target objects includes: Extracting three-dimensional position data of multiple target objects within the same time period from the target perception data set, wherein the three-dimensional position data includes a spatial coordinate point of each target object; Select one of the target object's spatial coordinate points as the reference point; For each other target object, extract its spatial coordinate point, and calculate the three-dimensional Euclidean distance between the spatial coordinate point and the reference point, and use the calculation result as the relative distance parameter between the target object and the reference point; Calculating the horizontal projection coordinates of the spatial coordinate point of the target object relative to the reference point, wherein the horizontal projection coordinates represent the position of the target object on the horizontal plane; Calculating the azimuth angle between the horizontal projection coordinates, and using the calculation result as the azimuth angle parameter of the target object relative to the reference point; The relative distance parameters and azimuth angle parameters of all target objects are summarized and sorted to generate a spatial relationship analysis result containing the spatial relationships of multiple target objects.

4. The target situation awareness method based on low-altitude intelligent network according to claim 1 is characterized in that: The step of inputting the spatial distribution features and the continuous change features into a pre-trained situation awareness model to generate situation-related features of the target object includes: Inputting the spatial distribution features into a first coding layer of the situation awareness model, performing dimensionality unification processing on a position coordinate set and a position correlation matrix in the spatial distribution features through the first coding layer to generate spatial coding features with a standard dimensional representation; Inputting the continuously changing feature into a second coding layer of the situation awareness model, and performing time alignment processing on the state change rate and the environmental impact correlation degree in the continuously changing feature through the second coding layer to generate a change coding feature with a time correspondence relationship; Inputting the spatial coding feature and the change coding feature into the weight allocation unit of the association analysis module of the situation awareness model to dynamically calculate the association weight of the spatial coding feature and the change coding feature, wherein the association weight is used to represent the contribution of the spatial coding feature and the change coding feature to the situation association feature; Performing weighted fusion processing on the spatial coding feature and the change coding feature based on the association weight to generate a fused association feature; The feature enhancement unit of the association analysis module performs information enhancement processing on the fused association features to generate situation association features of the target object.

5. The target situation awareness method based on low-altitude intelligent network according to claim 4 is characterized in that: The weight allocation unit of the association analysis module inputting the spatial coding feature and the change coding feature into the situation awareness model dynamically calculates the association weight of the spatial coding feature and the change coding feature, including: Inputting the spatial coding feature and the change coding feature into the weight distribution unit, wherein the weight distribution unit includes a fully connected layer and an activation function layer; Performing linear transformation processing on the spatial coding feature and the change coding feature through the fully connected layer to generate a feature importance evaluation vector; Normalizing the feature importance evaluation vector through the activation function layer to generate a first weight value corresponding to the spatial coding feature and a second weight value corresponding to the change coding feature, where the sum of the first weight value and the second weight value is 1; The first weight value is used as the associated weight of the spatial coding feature, and the second weight value is used as the associated weight of the change coding feature; the weight allocation unit adjusts the contribution ratio of the spatial coding feature and the change coding feature in the fusion process according to the first weight value and the second weight value.

6. The target situation awareness method based on low-altitude intelligent network according to claim 1 is characterized in that: Inputting the situation association features into the situation classification layer of the situation awareness model to obtain the current situation attributes and situation development trend information of the target object includes: Inputting the situation association features into the situation classification layer of the situation awareness model, performing attribute matching processing on the situation association features through the situation classification layer, generating a situation attribute candidate set of the target object and a matching confidence corresponding to each situation type identifier in the situation attribute candidate set, wherein the situation attribute candidate set includes multiple candidate situation type identifiers; Select the situation type identifier with the highest matching confidence as the current situation attribute of the target object; Inputting the situation-related features into the tendency prediction layer of the situation awareness model, performing historical information backtracking processing on the situation-related features through the tendency prediction layer, extracting the situation change pattern of the target object in the historical period, and performing tendency prediction processing on the current situation attributes for the future period based on the situation change pattern, thereby generating situation development tendency information including a direction indication and a rate indication, wherein the direction indication indicates the possible evolution direction of the target situation and the rate indication indicates the speed of the target situation evolution; The current situation attributes and the situation development tendency information are used as target situation determination results.

7. The target situation awareness method based on low-altitude intelligent network according to claim 6 is characterized in that: Inputting the situation association features into the tendency prediction layer of the situation awareness model, performing historical information backtracking processing on the situation association features by the tendency prediction layer, and extracting the situation change pattern of the target object within the historical period, includes: Extracting a correlation feature sequence of a historical period based on the situation correlation feature, wherein the correlation feature sequence includes situation correlation features of the target object in multiple historical periods; Inputting the associated feature sequence into the tendency prediction layer, wherein the tendency prediction layer comprises a long short-term memory network structure and a pattern extraction structure; Performing time series information storage processing on the associated feature sequence through the long short-term memory network structure to generate a memory state vector containing historical situation information; Performing feature decomposition processing on the memory state vector through the pattern extraction structure to separate the stable situation pattern and the fluctuating situation pattern of the target object in the historical period, wherein the stable situation pattern represents the situation characteristics maintained for a long time by the target object, and the fluctuating situation pattern represents the situation characteristics of the target object changing in the short term; The stable situation pattern and the fluctuating situation pattern are used as situation change patterns of the target object in a historical period.

8. The target situation awareness method based on low-altitude intelligent network according to claim 1 is characterized in that: The generating of a collaborative sensing instruction including a target positioning identifier according to the current situation attributes and the situation development tendency information, and sending the collaborative sensing instruction to the low-altitude intelligent network management terminal to trigger a linkage response operation, includes: Parsing the preset response rule library corresponding to the current situation attribute, extracting the response level identifier and the coordination strategy code associated with the current situation attribute, wherein the response level identifier is used to indicate the urgency of the response to be triggered, and the coordination strategy code is used to indicate the node type to be coordinated; Calculate the centroid coordinates and the position coverage area boundary of the target object in the current time period based on the three-dimensional position data in the target perception data set, where the centroid coordinates represent the center position of the target object and the position coverage area boundary represents the spatial distribution range of the target object; Processing the coordinates of the centroid of the location and the boundary of the location coverage area into a coordinate system to generate a standardized positioning identifier consistent with the geographic coordinate system of the low-altitude intelligent network, wherein the standardized positioning identifier includes a sequence of the center coordinate point and boundary coordinate points of the target object; Combined with the direction indication and the speed indication in the situation development tendency information, the standardized positioning identifier is processed for position prediction in the future time period to generate a predicted positioning identifier sequence of the target object in the subsequent time period; Integrate the response level identifier, collaborative strategy code, standardized positioning identifier, and predicted positioning identifier sequence to generate a collaborative sensing instruction including a positioning identifier chain with timestamp alignment, wherein the coordinate dimension of each identifier node in the positioning identifier chain is consistent with the coordinate dimension of the standardized positioning identifier; The collaborative sensing instruction is sent to the management terminal, so that the management terminal schedules the corresponding node to perform target monitoring or response operations according to the collaborative sensing instruction.

9. A target situation awareness system based on low-altitude intelligent network, characterized in that: It includes a processor and a memory, the memory is connected to the processor, the memory is used to store programs, instructions or codes, and the processor is used to execute the programs, instructions or codes in the memory to implement the target situation awareness method based on the low-altitude intelligent network as described in any one of claims 1 to 8.

Images