Formal methods for analysis of heterogeneous models of embedded systems

Technological developments in micro-electronics have made digital control an indispensable component in all engineering systems. The rapid pace of development and the demands on modern systems in terms of novel functions and shorter development cycles has led to many challenges in system design and verification. The down side of improved functionality is the unmanaged complexity: never have we had systems built with so many different disciplines simultaneously at work-each with their own collection of conceptual and concrete tools. To manage complexity in this setting it is essential to recognise and accommodate the diversities as early as they arise. For most application domains this results in a multi-paradigm development process, and is most visible in the design modelling stage. We discuss how mathematical modelling and analysis of system properties is affected by having several disciplines at work. We show that soundness in design models can be obtained both through static analysis based on properties defined for a meta-model, and through formal verification of an instance of a model-the latter being defined in terms of conformance to a requirements specification.