A Multitasking Finite State Architecture for Computer Control of an Electric Powertrain

Finite state techniques provide a common design language between the control engineer and the computer engineer for event driven computer control systems. They simplify communication and provide a highly maintainable control system understandable by both.

This paper describes the development of a control system for an electric vehicle powertrain utilizing finite state concepts. The basics of finite state automata are provided as a framework to discuss a unique multitasking software architecture developed for this application. The architecture employs conventional time-sliced techniques with task scheduling controlled by a finite state machine representation of the control strategy of the powertrain. The complexities of excitation variable sampling in this environment are also considered.