Search

In air combat, target intent recognition is the premise and foundation of battlefield situation awareness and intelligent decision-making. Aiming at the problem that traditional intention recognition methods cannot deal with a large amount of continuous target data, an air target combat intention recognition model based on one-dimensional convolutional neural networks and bidirectional long short-term memory (1DCNN-BiLSTM) is proposed. First, the target data is divided into fixed-size continuous subsequences by time sliding window on the basis of determining the target feature space and intention space. Second, the convolution operation is performed on the target sequence through the 1DCNN module as a means of extracting the features of the target attributes in the time dimension, and at the same time reducing the dimensionality of the target sequence, so as to facilitate the subsequent processing of the target data. Then, the BiLSTM module is utilized to capture the dependencies on the longer distance of the target sequence from both forward and reverse directions simultaneously. Finally, the optimal model structure and hyperparameters of 1DCNN-BiLSTM are not only determined through experiments, but also the validity of each part of the model is verified. Compared with the traditional methods, the model proposed in this paper effectively improves the accuracy of combat intention recognition of air targets, and provides the essential basis and auxiliary support for the decision-making of the commanders.

Most existing air target intention recognition methods use only single-moment information, risking failure when acquiring data containing noise and many outliers. The robustness of methods that utilize continuous moment information has yet to be explored. This paper designs a robust recognition method for air target intention to address the above problems. The method takes data with noise and outliers as the object, based on a parallel time-channel Transformer Encoder and a weight self-learning unit. First, a detailed introduction to air target intention recognition and robust recognition is given, and the intention space and feature space are defined. Subsequently, the data samples are reconstructed using a fixed-step sliding window to increase the information utilized with multi-moment information as input. Finally, step-wise and channel-wise correlations are extracted using a time-axis Transformer Encoder and a channel-axis Transformer Encoder, respectively, and the weights of the two branches’ outputs are automatically learnt using a weight self-learning unit. This enhanced self-attention network allocates attention weights between elements in the time and channel domain sequences to capture their long-range and short-range relationships and extract recognizable representations, making it robust to outliers and noise. The experimental results show that the model’s recognition accuracy and composite F1 score reach 96.9% and 0.9676, and its performance remains well when the noise level and outliers proportion increase. The ablation and comparison experiments show its advantage in accuracy over other models.

Abstract Aerial targets grouping is an important content of air battlefield situation awareness, and intelligent aerial targets grouping is an urgent need for intelligent air combat. Traditional aerial targets grouping methods have problems such as difficulty extracting air target features, difficulty designing formation similarity measures, and lack of data samples. To solve the above problems, this paper proposes a relational network based aerial targets grouping method to achieve end‐to‐end grouping under small sample conditions. Firstly, an adaptive clustering algorithm is proposed to preliminarily cluster the aerial targets in the air battlefield situation to obtain the group to be tested. Then, a representation of aerial target groups is designed, which unifies the multi‐dimensional features of the group's temporal, spatial, and attribute. Next, based on this representation, combining 3D convolutional neural networks (3DCNN) and long short‐term memory (LSTM), a neural network structure that can automatically extract multi‐dimensional features of groups is designed. After that, the relation module is used to learn the distance metric between groups. After training the relation network, the type of group to be tested can be identified, thereby achieving the goal of grouping. A dataset of 30 types of aerial target group patterns was established through simulation. The ablation experiments were conducted to verify the performance of each module, and the experimental results showed that the designed feature extraction module can effectively extract temporal sample features, while the designed relationship module performs better in extracting temporal sample information. By comparing several typical traditional aerial target grouping methods, the proposed method improves the grouping accuracy by 11.2% to 13.4%, achieving an accuracy of 91.1% in the 5‐way 1‐shot task and 95.6% in the 5‐way 5‐shot task.

Over-the-horizon radar (OTHR) is an important equipment for the ultralong-range early warning in the military, but the use of constant false-alarm rate (CFAR), which is a traditional detection method, makes it difficult in multi-aircraft formation recognition. To solve this problem, a multi-aircraft formation recognition method based on deep transfer learning in OTHR is proposed. First, the range-Doppler images of aircraft formation in OTHR are simulated, which are composed of four categories of samples. Secondly, a recognition model based on Convolutional Neural Network (CNN) and CFAR detection technology is constructed, whose training method is designed as a two-step transfer. Finally, the trained model can well distinguish the spectral characteristics of aircraft formation, and then recognize the aircraft number of a formation. Experiments show that the proposed method is better than the traditional CFAR detection method, and can detect the number of aircraft more accurately in the formation with the same false alarm rate.

In the application of aerial target recognition, on the one hand, the recognition error produced by the single measurement of the sensor is relatively large due to the impact of noise. On the other hand, it is difficult to apply machine learning methods to improve the intelligence and recognition effect due to few or no actual measurement samples. Aiming at these problems, an aerial target recognition algorithm based on self-attention and Long Short-Term Memory Network (LSTM) is proposed. LSTM can effectively extract temporal dependencies. The attention mechanism calculates the weight of each input element and applies the weight to the hidden state of the LSTM, thereby adjusting the LSTM’s attention to the input. This combination retains the learning ability of LSTM and introduces the advantages of the attention mechanism, making the model have stronger feature extraction ability and adaptability when processing sequence data. In addition, based on the prior information of the multidimensional characteristics of the target, the three-point estimation method is adopted to simulate an aerial target recognition dataset to train the recognition model. The experimental results show that the proposed algorithm achieves more than 91% recognition accuracy, lower false alarm rate and higher robustness compared with the multi-attribute decision-making (MADM) based on fuzzy numbers. [ABSTRACT FROM AUTHOR]

The frequent missing values in radar-derived time-series tracks of aerial targets (RTT-AT) lead to significant challenges in subsequent data-driven tasks. However, the majority of imputation research focuses on random missing (RM) that differs significantly from common missing patterns of RTT-AT. The method for solving the RM may experience performance degradation or failure when applied to RTT-AT imputation. Conventional autoregressive deep learning methods are prone to error accumulation and long-term dependency loss. In this paper, a non-autoregressive imputation model that addresses the issue of missing value imputation for two common missing patterns in RTT-AT is proposed. Our model consists of two probabilistic sparse diagonal masking self-attention (PSDMSA) units and a weight fusion unit. It learns missing values by combining the representations outputted by the two units, aiming to minimize the difference between the missing values and their actual values. The PSDMSA units effectively capture temporal dependencies and attribute correlations between time steps, improving imputation quality. The weight fusion unit automatically updates the weights of the output representations from the two units to obtain a more accurate final representation. The experimental results indicate that, despite varying missing rates in the two missing patterns, our model consistently outperforms other methods in imputation performance and exhibits a low frequency of deviations in estimates for specific missing entries. Compared to the state-of-the-art autoregressive deep learning imputation model Bidirectional Recurrent Imputation for Time Series (BRITS), our proposed model reduces mean absolute error (MAE) by 31%~50%. Additionally, the model attains a training speed that is 4 to 8 times faster when compared to both BRITS and a standard Transformer model when trained on the same dataset. Finally, the findings from the ablation experiments demonstrate that the PSDMSA, the weight fusion unit, cascade network design, and imputation loss enhance imputation performance and confirm the efficacy of our design. [ABSTRACT FROM AUTHOR]