Controller Design and Verification Using Qualitative and Semi-Quantitative Reasoning

本研究将侧重于开发一种方法和软件工具，用于设计和验证由多个简单控制器组成的异构控制器。组件控制器可以在不同的操作区域或在不同的时间尺度上操作。设计工具将使用QSIM和NSIM模拟器生成的植物和控制器的定性和半定量模拟来绘制可能的行为，从而生成自动稳定性证明。（定性模型指定变量之间的单调关系，而半定量模型指定约束变量相关函数的包络。）故障和干扰模型将用于测试所提出的控制器的鲁棒性，并证明其稳定性或建议对控制律进行修改。这是一个与德克萨斯大学奥斯汀分校计算机科学系Benjamin Kuipers教授合作的项目。//

This research will focus on developing a methodology and a software tool for the design and verification of heterogeneous controllers composed of multiple simple controllers. The component controllers may operate either on different operations regions or on different time scales. The design tool will use qualitative and semiquantitative simulations of the plants and controllers as produced by the QSIM and NSIM simulators to map out possible behaviors, thus producing automatic stability proofs. (Qualitative models specify monotonic relationships between variables, while semiquantitative models specify envelopes which bound the function relating the variables.) Models of faults and disturbances will be used to test the robustness of proposed controllers, and either prove their stability or suggest modifications to the control law. This is a collaborative project with Prof. Benjamin Kuipers, Department of Computer Sciences, University of Texas at Austin, Austin, Texas. //