关键词: 再入滑翔飞行器, 禁飞区, 代价函数, 贝叶斯定理, 意图推断

Abstract:

To address the issues of high misjudgment rates and excessively long inference time in intent inference for reentry glide vehicle under multiple no fly zone constraints, an intent inference method based on an adaptive cost function is proposed. Firstly, the effectiveness of intent inference using conventionally feature-based cost functions in multiple no fly zone scenarios is analyzed, and the confidence levels of these functions are evaluated to screen out the intent cost function with the best intent inference. Secondly, a robust prior mixing factor is introduced into Bayesian recursion. In light of the trajectory characteristics of the vehicle, it dynamically adjusts the uniform distribution and minimum probability, effectively preventing probability collapse. Finally, the relative situational relationships among the vehicle, no fly zones, and intended key regions of intent are analyzed in depth. The intent cost functions from angular and distance dimensions are adaptively fused, enabling the combined cost function to leverage superior intent inference capabilities. Simulation results demonstrate that compared to existing methods, the proposed approach achieves faster intent inference for reentry glide vehicle and delivers more reliable intent inference results.

Key words: reentry glide vehicle, no fly zone, cost function, Bayesian theorem, intent inference