Software-Based Cross-Platform Automated FlexRay Physical Layer Conformance Test Suite 2020-01-5129
Over the years FlexRay has achieved a significant position among automotive communication networks by satisfying the ever-growing needs for high data rate, advanced security, and performance robustness. Though widely accepted by vehicle manufactures, development of new vehicle variants demand conformance to physical layer protocol due to change in network topology, number of network nodes, cable type/length, and EMC/EMI influences. Irrespective of the platform, real vehicle or simulated (hardware-in-loop or vehicle-in-loop), physical layer conformance tests ensure network and signal integrity in the most fundamental physical layer and thereby help to avoid any discrepancy in the application layers. The proposed method discusses a MATLAB-based, measurement device-independent, automated test suite for physical layer conformance tests integrated with network configuration parameters. This method helps to reduce the need for expensive hardware test facilities by splitting the whole verification process into two phases: logging and analysis. Independent logging and analysis helps eliminate dependencies on specific scoping devices and device locked software licenses while keeping the same test standards. Additionally, it also helps to integrate real-time network configuration parameters, thereby achieving robust test coverage. This method also helps achieve a high level of integrity by bringing in pulse code modulation-based decoding and cross verifying the same against the Field Bus Exchange File format information. The proposed method, when applied to real vehicle/simulated platforms, yielded reduction effort and time reduction around one-third compared to the traditional method and due to automation achieved quicker processing of a large amount of network data. Also, by reducing the requirement of hardware by one-third compared to traditional method, the proposed method helps to reduce overall process cost. This method also lays a foundation for deep learning on much larger data to identify random distortions, anomalies, and unwanted hidden patterns in the real-time network.