Randomized Algorithms for Platform-based Design

The design of automotive control systems is becoming increasingly complex as the level of performance required by car manufactures grows continuously and the constraints on cost and development time imposed by the market become tighter. A successful design, without costly and time consuming re-desing cycles, can be achieved only be using an efficient design methodology that allows for component re-use and evaluation of platform requirements at the early stages of the design flow. In this paper, we illustrate a control-implementation design methodology for the development of embedded controllers by composition of algorithms picked up from libraries. Randomized algorithms and hybrid system theory are used to develop techniques for functional and architecture evaluations, which are implemented in a prototype tool.