To date the complexity of software (SW) in automobiles has increased significantly. This impressive growth of SW content in the last decade represents a significant investment of OEM's (Organisation of Engine Manufacturers) as well as suppliers. Therefore, explicit and vital platform concepts are a key factor for reusability and resulting cost reduction. Further more, tomorrow's automotive SW world will be more open (e.g. AUTOSAR) to migrate functions of different origin. Besides legal and commercial implications, platforms offer the long-term frame to practice SW-sharing. Taking a module as the smallest functional unit within an engine control unit (ECU), then an ECU today requires about 500 modules to achieve a control of engine and vehicle functions, including various levels of functional and electrical diagnoses and backup control strategies. Engine and vehicle functions on the one hand satisfy legal requirements, such as emissions and safety related issues, and on the other hand satisfy the manufacturers' requirements, such as performance and comfort functions. The 500 modules are represented by more than 3000 A4 pages of SW specification which are eventually coded and compiled to result in several Megabyte of code. The interactions of the 500 modules are very complex. It is safe to say, that our human cognitive ability is swiftly overstrained with respect to understanding these interactions. For this and for reasons of improving quality and optimising synergies between the diesel and gasoline engine development sections, a platform concept has been developed at Siemens VDO Powertrain reducing the apparent complexity by aggregating modules to larger functional blocks, with stable interfaces, configurable, scaleable, which can be shared between the diesel and gasoline world and which allows the integration of functionality developed by the customers. In this Paper a tool for the analysis, visualisation and optimisation is introduced supporting the partitioning of the modules into the aforementioned larger functional blocks (Aggregates), taking also into account the minimisation of the signal flow between Aggregates and the maximisation of reuse of Aggregates. The effect of partitioning on the signal flow is effectively represented in a visual form. Due to this visual representation everyone involved will more easily understand the implications for the SW-supplier as well as for the customer supplying his own functionality, potentially resulting in an early collaboration with customers and thus being able to respond more effectively to the customers' needs. As an example the paper includes a visual representation of the differences between the un-partitioned structure and a possible partitioning structure (Aggregates). The optimisation strategy utilising genetic algorithms is not described in this paper.The advantage of the Aggregate concept is that it is flexible but stable, adaptable within limits, reduces the logistic overhead of changes, eases system understanding and maximises reuse on an Aggregate level. In this paper PAVO (Platform Analysis Visualisation and Optimisation)was used to analyse and visualise, how a well designed aggregation reduces the number of externally exchanged signals by 15%, while doubling the number of Aggregates. At the same time inter-Aggregate connectivity was not increased. Compared to the non-aggregated case the external exchanged sig nals were reduced from 2213 to 1359 signals, a reduction of 40% and the number of packages to be handled at integration was re duced from 484 modules to 44 Aggregates, an order of magnitude reduction. The advantages of the Siemens VDO Aggregates are visible in terms of a clearer system understanding, reduced apparent complexity and reduced overheads when handling Aggregates as compared to modules. As continuous improvement Siemens VDO periodically revises the concrete aggregation. From this approach it is anticipated that further to the functional cohesion, external signal exchange minimisation leads to new concrete strategies for designing Aggregates.
PAVO - platform analysis, visualisation and optimisation
PAVO - Plattformanalyse, Visualisierung und Optimierung
2004
7 Seiten, 15 Bilder, 1 Quelle
Aufsatz (Konferenz)
Englisch
FLIGHT TRAJECTORY OPTIMISATION AND VISUALISATION TOOL
Europäisches Patentamt | 2017
|FLIGHT TRAJECTORY OPTIMISATION AND VISUALISATION TOOL
Europäisches Patentamt | 2016
|FLIGHT TRAJECTORY OPTIMISATION AND VISUALISATION TOOL
Europäisches Patentamt | 2016
|FLIGHT TRAJECTORY OPTIMISATION AND VISUALISATION TOOL
Europäisches Patentamt | 2023
|