Property Enforcement for Partially-Observed Discrete-Event Systems

Discrete Event Systems;Electrical Engineering;Engineering;Electrical Engineering: Systems

Yin, XiangTeneketzis, Demosthenis ;

University of Michigan

关键词: Discrete Event Systems;    Electrical Engineering;    Engineering;    Electrical Engineering: Systems;   

Others  :  https://deepblue.lib.umich.edu/bitstream/handle/2027.42/137097/xiangyin_1.pdf?sequence=1&isAllowed=y

瑞士|英语

来源: The Illinois Digital Environment for Access to Learning and Scholarship

【 摘 要 】

Engineering systems that involve physical elements, such as automobiles, aircraft, or electric power pants, that are controlled by a computational infrastructure that consists of several computers that communicate through a communication network, are called Cyber-Physical Systems. Ever-increasing demands for safety, security, performance, and certi cation of these critical systems put stringent constraints on their design and necessitate the use of formal model-based approaches to synthesize provably-correct feedback controllers. This dissertation aims to tackle these challenges by developing a novel methodology for synthesis of control and sensing strategies for Discrete Event Systems (DES), an important class of cyber-physical systems. First, we develop a uniform approach for synthesizing property enforcing supervisors for a wide class of properties called information-state-based (IS-based) properties. We then consider the enforcement of non-blockingness in addition to IS-based properties. We develop a nite structure called the All Enforcement Structure (AES) that embeds all valid supervisors. Furthermore, we propose novel and general approaches to solve the sensor activation problem for partially-observed DES. We extend our results for the sensor activation problem from the centralized case to the decentralized case. The methodology in the dissertation has the following novel features: (i) it explicitly considers and handles imperfect state information, due to sensor noise, and limited controllability, due to unexpected environmental disturbances; (ii) it is a uniform information-state-based approach that can be applied to a variety of user-speci ed requirements; (iii) it is a formal model-based approach, which results in provably correct solutions; and (iv) the methodology and associated theoretical foundations developed are generic and applicable to many types of networked cyber-physical systems with safety-critical requirements, in particular networked systems such as aircraft electric power systems and intelligent transportation systems.