Maneuvering Pattern Recognition of Aerial Targets Based on SCFG Model

Accurate identification of the maneuvering type of target aircraft is a prerequisite for the automatic judgment of target combat intention. Traditional target tracking algorithms adhere to the Markov property of "memorylessness", assuming that the state transition of the system depends only on the current state. However, in actual air combat, the inference of aircraft maneuvering patterns and operational intent is not an instantaneous process; rather, it requires analysis based on the temporal features of long-term continuous trajectories. This strong dependence on historical states conflicts with the single-step recursive mechanism of traditional algorithms, rendering them ineffective in identifying maneuvering patterns. To address this issue, a maneuvering pattern recognition method is proposed in this paper that transforms trajectory recognition into a syntactic parsing problem. The proposed method introduces stochastic context free grammar and the Cocke-Younger-Kasami (CYK) algorithm. Specifically, "maneuver primitives" are first defined to constitute the alphabet of the grammar, and a corresponding categorization mechanism is developed to map continuous motion trajectories into discrete syntactic strings. Furthermore, dedicated grammar models are established for typical aerial maneuvering patterns. By deriving and parsing the string of maneuver primitives using the CYK algorithm, the system can reversely match the specific grammar model to which the target belongs, thereby achieving accurate discrimination of the maneuvering pattern. Simulation results demonstrate that the proposed method effectively meets the requirements for automatic maneuvering target recognition in complex air combat scenarios.