Applying Knowledge-Based Engineering to Control Software Generation

Abstract

The development of control software for mechatronic systems requires data and information from all design domains in order to create the required integrated functionality. This paper proposes a method that combines function modeling and multi-domain modeling primitives to generate control software. Provided a function model based on the Function-Behavior-State modeling paradigm and performance requirements, a knowledge-based engineering application instantiates a virtual product model and control software for a target platform. Object-oriented modeling techniques provide for the development of primitive libraries, which represent both hardware and software components, while integrated design rules ensure that components are correctly placed and connected. Requirements are translated to software specific parameters. A case study of a mobile robot shows that for specific applications both target and simulation control software code can be generated from the same input, and a generated virtual product model can serve as a simulation model in order to validate functionality.