论文部分内容阅读
研究了基于图像理解和意图识别的多Agent自主协同方法,研究对象是空中目标,估计跟踪关系,帮助观察者有效地做出决策.列出了两种跟踪模型:一是通过估计敌方飞机的飞行路径来调整自己的前进方向;二是一直向敌方方向跟踪.基于两种方向,提出了一种复合函数来估计跟踪趋势的概率,同时假设Agent之间正态分布的关系来估计概率以便决策合作策略.将贝叶斯网络有效地应用在这两种跟踪模型中.所提出的算法均被应用于仿真序列图实验,证实了所提方法的有效性.
This paper studies a multi-agent autonomous collaborative method based on image understanding and intent recognition, which is aimed at air targets and tracking relationships to help observers make effective decision-making. Two types of tracking models are listed: Flight path to adjust their own forward direction; the second is to track the direction of the enemy.Based on two directions, a composite function to estimate the probability of tracking the trend, and the normal distribution of the relationship between Agents at the same time to estimate the probability of probability Decision-making and cooperation strategy, the Bayesian network is effectively applied to both tracking models.The proposed algorithms are applied to simulation sequence chart experiments, which proves the effectiveness of the proposed method.