The air battlefield situation provides a general description of the actions and states of all participants， while target combat intention recognition serves as a critical foundation for air battlefield posture assessment. To address the challenge of acquiring a large amount of labeled air target battlefield data amidst intense confrontation and rapidly evolving scenarios， a semi-supervised air targets combat intention recognition model is proposed based on metric learning. The model offers a method for uncovering potential patterns from unlabeled samples， thus reducing the reliance on extensive labeled data. In this model， the target time-series data encoder reduces the dimensionality of target data and produces an embedded representation. Based on this， loss values are calculated by measuring the similarity between labeled target sequences and intention types， as well as between labeled and unlabeled target sequences. The experimental results demonstrate that the model achieves combat intention recognition accuracy rates of 86%， 89%， and 91% with labeled sample rates of 30%， 40%， and 50%， respectively.

With the continuous evolution of war forms and the upgrading of weapons and equipment， the air battlefield situation is becoming more and more complex. Quickly and accurately identifying the combat intention of the target is an important content of battlefield situation assessment， which can provide auxiliary information for commanders to make decisions and help them to seize the initiative in the war. This paper firstly introduces the basic concepts and related models of target intention recognition， defines the concepts of target intention and intention recognition， and determines the status and importance of target intention recognition from two aspects of the operational command decision-making process and information fusion process. Secondly， target features and intention space， as input attributes and recognition framework for intention recognition respectively， are the basis of intention recognition and are reviewed. Then， five common target intention recognition methods such as rule and template matching， evidence theory， Bayesian network， traditional machine learning and neural network are reviewed. The basic mechanism and recognition process of each recognition method are introduced， and its advantages and disadvantages are summarized. Finally， the performance of five kinds of target combat intention recognition methods is compared and analyzed， and the future research direction is prospected.

Threats often have the characteristics of dynamic evolution with the development of bilateral confrontation. Traditional threat assessment methods are more based on static threats and lack of prediction of dynamic threats. To solve this problem， this paper proposes a threat assessment method for aerial targets under confrontational conditions. Setting the red side as the attacking side and the blueside as the defending side. The reinforcement learning model is established with the enemy target as the agent， and its state space， action space， transition function and reward function are designed. The threat assessment model is established， the threat element index is established， and the threat assessment method is designed. The model is trained， and the trained model can predict the enemy aerial target threat according to the confrontation situation. Through test and analysis， the method is more reasonable for threat assessment of enemy aerial targets under confrontation conditions.

With regard to the non-cooperative actively approaching spacecraft with maneuverability， a threat index calculation model to analyze the threat of the chaser to the target is proposed. We extract the impulse and the time required for rendezvous based on two impulses Lambert algorithm as threat elements after analyzing the mission mode of the chaser. The most threatening route is selected to weigh the impulse and time in order to calculate the comprehensive threat index. Simulation results show that the threat index of the chaser is very sensitive to the changes of orbital elements. Comparing the commonly used relative distance index， the threat index can identify the maneuver intention more quickly according to its trend. Under the complex space situation， the threat index analysis model is able to predict the threat of actively approaching spacecraft so as to take timely protective measures and improve the survivability of satellites.