Synthesis of Control Systems based on Incomplete Observed Information

本课题研究如何利用随机优化和随机控制器设计不完全观测数据下的高性能控制系统。首先，粒子滤波器可以有效地构建系统模型。然后给出了一种基于模型的从低分辨率信号中重建高分辨率信号的方法，并对其有效性进行了评价。此外，通过多智能体系统和生物系统等实例，验证了控制中的随机行为对不完全观测数据的有效补偿。

This project studies how to design high performance control systems under incom-plete observed data by exploiting stochastic optimizations and stochastic controllers. First,particle fil-ters are shown to be effective to construct system models. Then a model-based method to reconstruct the high resolution signals from the low resolution ones is given and its effectiveness is evaluated. Further-more, it is confirmed that the stochastic behaviors in control are effective to compensate the incomplete observed data through various cases including multi-agent systems and biological ones.