A Formal High-level Modeling Approach to Develop Reliable Components in Vision-based Robotics

摘要

This paper proposes the use of the control-state Abstract State Machines for a rigorous foundation in highlevel modeling and validating component-based applications in Vision-Based Robotics. In particular, an extension of the classical flowchart notation for control-state ASMs is proposed to support modularization and reuse in a direct way. The resulting ASM models are to be intended as “ground models” that can be used as basis or patterns to practically model and formally validate the behavior of typical robotic control tasks, and to link (via successive refinements) these high-level models of components to their implementation code by making their functional correctness mathematically controllable. The proposed flowchart extension and the availability of reusable and validated ground models allow a better system design and speed up the development of the system.