Addressing Cross-Tool Semantic Ambiguities in Behavior Modeling for Vehicle Motion Control

  • Authors:
  • Sandeep Neema;Sushil Birla;Shige Wang;Tripti Saxena

  • Affiliations:
  • Vanderbilt University, Nashville TN 37203;General Motors Corporation, Warren MI 48090;General Motors Corporation, Warren MI 48090;Vanderbilt University, Nashville TN 37203

  • Venue:
  • Model-Driven Development of Reliable Automotive Services
  • Year:
  • 2006

Quantified Score

Hi-index 0.00

Visualization

Abstract

Emerging model-based development methods in the Automotive Vehicle Motion Control (VMC) domain are using different tools at various stages of the engineering process. Behavioral models created in various forms of finite state machines have to be exchanged across these tools, but semantic unknowns in modeling environments and semantic variations across tools preclude automated correct interpretation. This research presents an approach to address this issue through an unambiguous, math-based, tool-neutral extended finite state machine metamodel (eFSM) for behavior specifications in the automotive VMC domain. The semantics of the metamodel are anchored to formal specifications in a mathematical framework. Our approach requires modeling with commercial tool environments conforming to the eFSM. The conformance is enforced by exporting the tool native models into eFSM-conformant models and checking them against the well-formed rules encoded as OCL constraints in the eFSM. We have performed "proof of concept" exercises with two commercial tools in transforming their native models into eFSM-conformant forms, and have been able to show that certain ambiguities in both tools can be prevented through the eFSM, promising higher confidence software engineering for the VMC domain.