Geometry-driven petri nets and a method for modelling mechatronic control systems

Abstract

The development process of mechatronic control systems often relies on physical prototypes to test the interactions between the control software and mechanical components. However, the logistics of synchronizing a concurrent development process and the risks of integrating only partially completed sub-systems often limits effective prototyping. The consequent lack of feedback can lead to overly complex and unreliable systems which may have to undergo expensive re-designs. The interactions between mechanical systems and control software can also be recreated artificially by combining a hybrid modelling language with computer graphics technology. A dynamic 3D environment can generate sensor telemetry for input to a control system, which in turn alters the state of the environment through virtual actuators. This kind of simulation allows engineers to explore a larger design space early during the development process without committing significant resources to physical prototypes. This dissertation introduces a method for simulating mechatronic systems using Petri Nets and Scene Trees. The following chapters formally define the modelling language and illustrate the software architecture and user interface of a novel simulation development environment. The research is validated through qualitative reasoning and by demonstrating a simulation that detects design flaws in a mechatronic system which may have otherwise lead to expensive redesigns in the physical system.