J. Grantner and G. Fodor
An autonomous control system working in a real-world environment may encounter events that are outside the bounds of the established safety criteria. In the case of a hybrid system, unexpected situations in its environment may violate assumptions that have been used to design the continuous and/or the discrete segment. Hence, it is difficult to design a control unit that can detect and recover from such errors. An execution monitoring algorithm for this type of faults will draw practical interest only if it is application independent. This paper presents an execution monitoring and fault detection method for hybrid systems along with its limitations. This approach addresses a new class of problems encountered with hybrid systems that needs to be considered in the design and analysis phases.