ABSTRACT
Modern cyber physical systems (CPS) are a complex mixture of cyber and physical components. This chapter focuses on two aspects of the development of a resilient sensor fusion algorithm. It considers a standard CPS model, in which multiple sensors measuring the same physical variable communicate with the system over a shared bus. The chapter analyzes the effect of the communication schedule on the impact of an attacker trying to increase the uncertainty in the system. It then investigates ways of using past measurements to improve the precision of the fusion algorithm at the current time. The chapter also explores how different schedules of sending measurements over a shared bus affect the uncertainty in the system, as reflected in the size of fusion interval. When designing a fusion algorithm, the first consideration is the underlying sensor model; it affects both the purpose and the approach of algorithm. The two most commonly used models can be broadly classified as probabilistic and abstract.
