In this paper a new partitioning approach called CASA (Correlate And Soak Allocation) was presented in order to allocated application software fulfilling the AUTOSAR standard during mapping process. Moreover a CAN-Model implemented in Matlab/Simulink was implemented to validated test data. To fulfill real-time constraints within a priority based network as a CAN network, a method was described adapting priorities due to response times of certain software modules. Therefore an algorithm based on the Max-Plus-Algebra was depicted by an example task graph. It was shown how to derive response times as well as priorities in order to configure the system. Finally CASA was validated using the CAN-Model. First, five models were randomly generated using Matlab. Partitioning these models confirmed the power of CASA. Furthermore a vehicle suspension model was partitioned leading to a reduction of busload of about 5.31 % in contrast to partitioning results obtained by Kernighan-Lin and Simulated-Annealing. However, it has to be emphasized that results obtained by Kernighan-Lin as well as Simulated-Annealing are based on an intelligent starting initialization requiring expertises of the network to be implemented. These expertises are not needed using CASA. CASA determines an optimized number of electronic control units (ECUs) in order to treat the problem of space limitations in vehicles. Applying this knowledge for the initialization of the commonly used partitioning methods leads still to a amelioration of 5.31 % as presented above. Using this approach makes it possible to optimize the mapping process of AUTOSAR standardized networks. It will lead to a reduction of external communication via the bus as well as a reduction of the number of ECUs used. So space limitations in vehicles are faced as well as latency times in safety-relevant systems are reduced by less communication via the bus. This makes it to possible to implement further electronically based systems like X-by-wire without breaking realtime constraints.
Configuration of standardized automotive real-time networks by automated software allocation
2007
12 Seiten, 5 Bilder, 3 Tabellen, 5 Quellen
Conference paper
Storage medium
English
A model-based configuration approach for automotive real-time operating systems
Automotive engineering | 2015
|A Model-Based Configuration Approach for Automotive Real-Time Operating Systems
SAE Technical Papers | 2015
|A Model-Based Configuration Approach for Automotive Real-Time Operating Systems
British Library Conference Proceedings | 2015
|SOLUTIONS - Software - Automated Real-Time Spacecraft Navigation System
Online Contents | 2001
SAE Technical Papers | 2008
|