基于贝叶斯信息更新的无人机目标搜索策略研究

Research on unmanned aerial vehicle target search strategy based on Bayesian information update

基于贝叶斯信息更新方法,提出了一个无人机目标搜索的动态策略模型,并给出了相应算法。该模型是在一般贝叶斯先验假设下给出的,不需要现有相关文献中均匀分布的假设。在无人机搜索行动中,可以实施贝叶斯干预,利用新获取的目标搜索区域的额外信息,实时地改进搜索策略。数值实验中,假设目标处于搜索区域的分布为正态分布,实验结果表明,相比较均匀分布假设,在正态分布假设下发现概率更大,而且贝叶斯干预后的累计发现概率不会降低。

Based on Bayesian information updating method,a dynamic strategy model of unmanned aerial vehicle target search was proposed and the corresponding algorithm was given.This model was based on a more general Bayesian prior hypothesis instead of uniform distribution assumption in some existing literature.In unmanned aerial vehicle search operations,Bayesian intervention could be implemented to improve the real time search strategy by utilizing the newly acquired information of the target search area.In the numerical experiment,the distribution of the target in the search area was assumed to be normal distribution.The experimental results show that the discovery probability under the normal distribution hypothesis is higher than that under the uniform distribution hypothesis,and the cumulative discovery probability will not decrease after Bayesian intervention.