Acausal Modeling and Code Generation for AUTOSAR Applications Based on Modelica

Mainstream AUTomotive Open System ARchitecture (AUTOSAR) tools are typically closed-source and rely on causal signal-flow modeling, which limits cross-domain integration, extensibility, and support for early-stage co-design. To overcome these limitations, we propose a Modelica-based framework that enables acausal, graphical modeling of AUTOSAR application-layer components in Modelica and automated generation of ARXML and C code for embedded targets. A lightweight and extensible XML configuration scheme is introduced to manage nongraphical AUTOSAR elements and to ensure accurate semantic mapping between Modelica classes and AUTOSAR software components. A rule-driven transformation pipeline generates standard-compliant ARXML files, while a structured code generation process produces prototypical AUTOSAR-compliant C implementations. We evaluate the framework on eight representative automotive use cases. Experimental results demonstrate that the generated ARXML and C code are functionally equivalent to those produced by the MATLAB/Simulink AUTOSAR Blockset, while achieving faster generation speed, reduced modeling complexity, and improved flexibility for multidomain design. The open-source availability of the framework ensures reproducibility, facilitates collaborative development, and supports industrial adoption. This work bridges the gap between Modelica-based modeling and industrial AUTOSAR deployment, enabling seamless integration of multidomain cyber-physical systems in automotive software engineering.