The identification of Unmanned Aerial Vehicles (UAVs) and flying birds based on radar track sequences is crucial for air safety supervision. In practical applications, it is essential to achieve accurate and rapid classification of UAVs/flying birds with the continuous arrival of track data. A short-medium-long multi-scale dynamic classification mechanism characterized by ‘rapid multi-feature synthesis, multi-likelihood sequential decision, and multi-factor long-term precise classification’ is proposed. In rapid multi-feature synthesis, input track vectors are categorized based on physical features: position-related features (representing the target's situation), velocity-related features (representing target's situation changes), and radiation-related features (representing target's material structure). These features are then fed into a short-term multi-head one-dimensional Convolutional Neural Network (1D-CNN) and synthesized using a channel attention mechanism, enabling real-time measurement of target attribute confidence. In multi-likelihood sequential decision-making, the likelihood distribution of target attribute confidence is statistically analyzed, and a multi-leveldecision logic incorporating both short-term and long-term confidence likelihood is designed to achieve comprehensive inference of target attributes over a longer time span. In multi-factor long-term precise classification, multiple factors measurements including velocity/heading angle changes and velocity/head angle trends are proposed, and the random forest algorithm is then employed to accurately classify hard-to-distinguish samples with multiple features over a long period of time. The proposed algorithm outperforms the existing algorithms in terms of classification accuracy, false positive rate, and false negative rate in real radar track data, verifying its effectiveness.

Analysis of cluster behaviors of the on-coming cluster is an essential measure for high value locations on the battlefield. Unlike target tracking and clustering analysis, cluster behaviors analysis under cluster security requirements remains an open issue, which is a joint analysis of clustering, intent reasoning and activity regions. To address this issue, a framework for cluster behaviors analysis is proposed by incorporating expert knowledge and domain knowledge, and a knowledge-assisted score function with is designed to improve the accuracy of intent reasoning network, overcoming the effects of possible knowledge errors. The framework consists of three modules for cluster analysis, intent reasoning and activity region analysis for typical tasks, in which an intent reasoning network is constructed to obtain cluster intents by using a hybrid knowledge and data driven approach. Furthermore, considering the complexity of the battlefield environment, different tasks and corresponding activity region optimization functions are designed for cluster activity region analysis, which are vital elements of cluster behaviors analysis. Simulations demonstrate the effectiveness of the proposed cluster behaviors analysis framework.