基于动态博弈的空间目标行为预测与决策方法  

Actions Prediction and Decision Approach for Space Target Based on Dynamic Game

作　　者：张鹏[1,2] 王磊[1] 汪慧英[1] 

机构地区：[1]同济大学电子与信息工程学院,上海200092 [2]山东轻工业学院机械工程学院,济南250353

出　　处：《系统仿真学报》2009年第S2期49-51,共3页Journal of System Simulation

摘　　要：以多源信息融合为基础,将空间机动平台之间的攻防视作一种可观察行动的多阶段不完全信息博弈,建立了基于动态贝叶斯博弈的模型,结合空间平台的历史行为和机动能力的评估,对其类型信念进行合理的计算及修正。通过分析双方的收益,预测出理性的攻防双方在下个博弈阶段会选择攻击和防御的概率,从而使主动的有针对性的防御成为可能。Based on multi-source information fusion,the space countermeasure was described as a multistage game of the incomplete information with observable actions.The dynamic bayesian game model was given out.According to the historical actions and maneuver performance evaluation of the aerospace vehicles,the model reasonably updates the type probability of the aerospace vehicles.It can predict the probability of attacks or defenses that rational attacker or defender will take in the next stage of the game,in order to maximize their payoff.It may change the passive detection to the active protection for the defender.

关 键 词：动态博弈 空间博弈 行为决策 行为预测 

分 类 号：F224.32[经济管理—国民经济]