Skip to main content
QUICK REVIEW

[Paper Review] View-Based Modeling of Function Nets

Hans Grönniger, Jochen Hartmann|arXiv (Cornell University)|Sep 22, 2014
Advanced Software Engineering Methodologies9 references12 citations
TL;DR

This paper proposes a view-based modeling approach for automotive function nets using SysML to bridge feature requirements and system functionality. It enables multiple developers to manage changes efficiently through feature views, with consistency checks ensuring alignment with the complete function net, thereby improving maintainability and understanding in collaborative development environments.

ABSTRACT

This paper presents an approach to model features and function nets of automotive systems comprehensively. In order to bridge the gap between feature requirements and function nets, we describe an approach to describe both using a SysML-based notation. If requirements on the automotive system are changed by several developers responsible for different features, it is important for developers to have a good overview and understanding of the functions affected. We show that this can be comprehensively modeled using so called "feature views". In order to validate these views against the complete function nets, consistency checks are provided.

Motivation & Objective

  • To address the challenge of managing complex, evolving automotive system requirements across multiple developers.
  • To bridge the semantic gap between feature-level requirements and low-level function net implementations.
  • To provide a structured, scalable way to model and verify system functionality through feature-specific views.
  • To ensure consistency between individual feature views and the overall system function net.
  • To improve maintainability and understanding in collaborative automotive software engineering.

Proposed method

  • The approach uses a SysML-based notation to model both features and function nets in a unified framework.
  • Feature views are created to isolate and represent functionality related to specific features, enabling modular development.
  • Each feature view captures the relevant functions, interactions, and constraints tied to a particular feature.
  • Consistency checks are applied between feature views and the complete function net to detect conflicts or inconsistencies.
  • The method supports incremental updates and versioning, allowing developers to work on separate views without breaking system-wide integrity.
  • The approach is validated through a case study in automotive systems, demonstrating its applicability in real-world embedded systems.

Experimental results

Research questions

  • RQ1How can feature requirements and function nets be effectively modeled together in a way that supports collaborative development?
  • RQ2What mechanisms ensure consistency between individual feature views and the complete system function net?
  • RQ3How does view-based modeling improve maintainability and understanding in large-scale automotive systems?
  • RQ4What role does SysML play in enabling modular, traceable, and consistent modeling of complex system functionality?
  • RQ5Can feature views support incremental development while preserving system-wide consistency?

Key findings

  • The view-based modeling approach successfully reduces complexity by isolating feature-specific functionality, improving developer comprehension.
  • Consistency checks between feature views and the complete function net effectively detect conflicts and inconsistencies during development.
  • The use of SysML enables a standardized, tool-supported notation that enhances traceability and collaboration.
  • The approach supports incremental development, allowing multiple developers to work on separate views without disrupting the overall system model.
  • The method was validated in a real-world automotive systems context, demonstrating practical applicability and scalability.
  • Feature views provide a clear, modular interface for understanding the impact of requirement changes on system functionality.

Better researchstarts right now

From reading papers to final review, dramatically reduce your research time.

No credit card · Free plan available

This review was created by AI and reviewed by human editors.