Application-Driven Evaluation of AUTOSAR Basic Software on Modern ECUs

Abstract

When integrating AUTOSAR software on an automotive ECU, errors may occur due to the large number of modules involved and/or improper timing. These errors manifest at the application level complicating the test and verification process. Since AUTOSAR has a layered architecture, it is often cumbersome to identify sources of errors. In this paper, to help integrating software on an ECU, we propose a technique to verify functionality and timing of generated AUTOSAR modules in a semi-automated manner. Our technique consists in defining test cases based on the interface descriptions of AUTOSAR modules and application software. This allows reliably identifying AUTOSAR modules affected by functional and/or timing errors and simplifies the test and verification process. We illustrate the benefits by our technique by means of a case study performed on the real hardware.