Design patterns for safety-critical embedded systems

Over the last few years, embedded systems have been increasingly used in safety-critical applications where failure can have serious consequences. The design of these systems is a complex process, which is requiring the integration of common design methods both in hardware and software to fulfill functional and non-functional requirements for these safety-critical applications. Design patterns, which give abstract solutions to commonly recurring design problems, have been widely used in the software and hardware domain. In this thesis, the concept of design patterns is adopted in the design of safety-critical embedded system. A catalog of design patterns was constructed to support the design of safety-critical embedded systems. This catalog includes a set of hardware and software design patterns which cover common design problems such as handling of random and systematic faults, safety monitoring, and sequence control. Furthermore, the catalog provides a decision support component that supports the decision process of choosing a suitable pattern for a particular problem based on the available resources and the requirements of the applicable patterns. As non-functional requirements are an important aspect in the design of safety-critical embedded systems, this work focuses on the integration of implications on non-functional properties in the existing design pattern concept. A pattern representation is proposed for safety-critical embedded application design methods by including fields for the implications and side effects of the represented design pattern on the non-functional requirements of the systems. The considered requirements include safety, reliability, modifiability, cost, and execution time. Safety and reliability represent the main non-functional requirements that should be provided in the design of safety-critical applications. Thus, reliability and safety assessment methods are proposed to show the relative safety and reliability improvement which can be achieved when using the design patterns under consideration. Moreover, a Monte Carlo based simulation method is used to illustrate the proposed assessment method which allows comparing different design patterns with respect to their impact on safety and reliability