In this chapter, the creation of "virtual prototypes" of complex micro-electromechanical transducers is presented. Creating these behavioural models can be partially automatised using a reduced-order modelling (ROM) method. It uses modal decomposition to represent the movement of flexible structures. Shape functions model the energy conservation and full coupling between the different physical domains. Both modal shapes and shape functions for strain energy and lumped capacitances of the structure can be derived in a highly automated way from a detailed 3D finite element (FE) model available from earlier design stages. Separating the generation of the ROM from the same FE model but for different operation directions circumvents the current limitations of the used ROM method. These submodels are integrated into a full model of the transducer. VHDL-AMS system is used to describe additional strong coupling effects 135 A. Vachoux (ed.), Applications of Specification and Design Languages for SoCs -Selected papers from FDL 2005, 135-153. © 2006 Springer. 136Torsten Mähne et al. between the different operation directions, which are not considered by the used ROM method itself. The application of this methodology on a commercially available yaw rate sensor as an example for a complex transducer demonstrates the practical suitability of this approach.

This paper describes ModelLib, a web-based platform for collecting models from different domains (e.g. electrical, mechanical) and levels of abstractions. Use cases for this tool are presented, which show how it can support the design process of complex AMS systems through better reuse of existing models for tasks like architecture exploration, system validation, and creation of more and more elaborated models of the system. The current state of the implemented ModelLib prototype is described and an outlook on its further development is given.

No abstract

No abstract

Abstract-Systems-on-Chips (SoCs) integrate more and more heterogeneous components: analog/RF/digital circuits, sensors, actuators, software. For the design of these systems very different description formalisms, or Models of Computation (MoCs), and tools are used for the different subblocks and design stages, which often create interoperability problems. Additionally the verification of a complete SoC is difficult due to huge performance problems. The goal of this Ph.D. work is to develop an efficient modeling and simulation platform that supports the design of mixed-signal SoCs using component models written in different design languages and using different MoCs. One component of this work is the development of a web-based platform for collecting behavioral models and supporting the design of Analog and Mixed-Signal (AMS) SoCs. Its current state and an outlook on its further development is the focus of this paper. I. IThe design of SoCs has currently to address a number of significant issues, namely: increasing complexity (computing and communication capabilities), significant heterogeneity (analog, RF and digital hardware, embedded software, sensors, actuators), increasing environmental awareness (energy saving, environmental monitoring and interaction), increasing sensitivity to silicon technologies (deep sub-micron technological processes), and increasing reuse of subsystems (ever shrinking time to market). One difficulty in solving these issues while designing a mixed-signal SoC is the usage of a diversity of specialized EDA tools, design languages and design formats that are usually efficiently supporting only one aspect, i.e., RF, analog, or digital, of the complete mixed-signal system. Furthermore, the design's heterogeneity requires a diversity of description formalisms, also called Models of Computation (MoCs), analysis and simulation methods. It is still difficult to handle all the different design aspects simultaneously. Designers are forced to bridge the gaps between tools and methodologies using manual conversion of models, proprietary tool couplings, and tool integrations. This makes the design process overly complicated, error-prone and time consuming. Another very important issue is the capability to perform efficient overall system verification in the early phases of the design process. System verification is based on the development of virtual prototypes [1] of the complete heterogeneous system. Its main goals are to support architecture exploration, performance estimations, validation of reused parts, verification of the interfaces between MEMS, RF, analog, and digital parts, This work has been funded by the Hasler Stiftung under project № 2161. the interoperability with other systems, and the assessment of the impacts of the future working environment and the used manufacturing technologies. Modeling tasks are therefore at the heart of the V-shaped SoC design process (Fig. 1). The management of the created models of a device, component, or the whole system on different abstraction...