Domain Controlled Architecture
A New Approach for Large Scale Software Integrated Automotive Systems
Dominik Reinhardt
1
and Markus Kucera
2
1
BMW AG, Munich, Germany
2
University of Applied Science Regensburg, Regensburg, Germany
Dominik.Reinhardt@bmw.de, Markus.Kucera@hs-regensburg.de
Keywords:
Large Scale Software Integration, LSSI, Automotive Real Time, Multi-core, Many-core, Embedded Automo-
tive Software Architecture.
Abstract:
Electric and electronic functionalities increase exponentially in every mobility domain. The automotive in-
dustry is confronted with a rising system complexity and several restricting requirements and standards (like
AUTOSAR), in particular to design embedded software for electronic control units. To stand against rampant
functionalities software units could be restructured according to their affiliation and should not be attached to
a certain place. This can be effected by integration on single controllers. On the one hand the system wide
amount of hardware controllers could such be limited. On the other hand the workload for integration CPUs
will rise. To support this paradigm, multi-core systems can provide enough processing power in an efficient
way. This paper shows a first approach to combine automotive functionality on such a single controller.
1 INTRODUCTION
E/E (Electric/Electronic) systems represent an impor-
tant part in premium vehicles and are the major con-
tributor in creating added value for OEMs (Original
Equipment Manufacturer). The amount of vehicle
functions increases exponentially and functionalities
get more and more complex. This trend is tightened
by additional requirements like lightweight design
and construction, energy efficiency, functional safety
driven by the new ISO standard for road vehicles, and
last but not least by the development standard (AU-
TOSAR Administration, 2012) for automotive soft-
ware architecture AUTOSAR (AUTomotive Open Sys-
tem ARchitecture).
Facing increasing software workload and rising
need for SW/HW robustness, there is a need for more
powerful hardware resources. Furthermore, the num-
ber of ECU (Electronic Control Unit) devices must
be reduced (Gut et al., 2012). To save fuel and extend
the crusing range of electric cars, the power consump-
tion of embedded systems and in detail of electronic
semiconductors shall be minimized (Sch
¨
ottle, 2012),
(Barthels et al., 2012). The actual state-of-the-art
in science and technology (Arbeitskreis-Multicore,
2011) shows that multi-core technology can solve the
problem in a more efficient way. Ten years ago Gor-
don Moore’s Law (Moore, 1965) was contested in the
area of consumer electronics and multi-core proces-
sors were published for general public on the market.
High performance computing systems or even mobile
devices like smart phones cannot get along without
this technology to supply the requests of processing
power. This paradigm has reached the automotive in-
dustry eventually (Schneider et al., 2010) and (Monot
et al., 2010).
Dual- or quad-core systems are already available
on the market for vehicle manufacturers. In the fu-
ture this trend will be continued and the amount of
cores will rise. Many-core controllers could revolu-
tionize current development standards and strategies.
For automotive systems it is not yet completely solved
how to obtain more and more speedup with a rising
number of cores and in comparison with small but
computationally intensive, high interconnected appli-
cations. For every single vehicle function built for
domains like chassis or powertrain it is to clarify,
how to segregate them from each other and how to
achieve given constraints in time or space. According
to Amdahl’s (Amdahl, 1967) and Gustafson’s Law
(Gustafson, 1988), to optimize the systems speedup,
software modules with less dependencies to units lo-
cated on other cores can be parallelized more effi-
ciently on probably more than 16 cores. This fact
will challenge automotive software engineers to re-
view former development techniques and design rules
221
资源评论
yxiaolian
- 粉丝: 5
- 资源: 6
最新资源
- 基于HTML、CSS、JavaScript的easy云盘前端设计源码
- 基于Java、Vue等技术的优加任务管理系统设计源码
- matlab simulink半车主动悬架建模：基于ADRC(自抗扰控制)的主动悬架控制 主体模型为半车主动悬架，采取ADRC控制 输出为车身加速度，悬架动挠度，轮胎动变形 默认输入为正弦路面输
- 基于PHP和Vue的河马跑腿私域配送团队小程序设计源码
- Linux RTL8761b蓝牙驱动 Ubuntu 20.04可用
- 移动磁铁在线圈中产生感应电压分析与仿真 COMSOL 6.0案例还原及 此模型模拟磁铁在线圈中的运动，并计算感应电压，磁铁的位移很明显，因此使用动网格和滑移网格
- 基于TypeScript和JavaScript的核桃健康App设计源码
- 永磁同步电机全阶自适应观测器 自适应全阶观测器MATLAB仿真，高速电机，基础版15.9，改进版49（改进版波形精美，易于出图）下面图为改进版，低速高速都可以，最高5W转每分
- 基于Python生态的第三方库管理器设计源码
- 基于three.js和Vue3的简易智慧城市设计源码
- simulink永磁同步风机风光储VSG一次调频，风机为PMSG，风光储并网系统，频率波形和风机VSG出力如图 网侧VSG同步机控制
- 基于Vue框架的汽修门店SaaS系统设计源码
- 基于Kotlin语言的Android作业设计源码
- mmc分布式储能 恒功率控制 恒电压控制 无缝切
- 基于微信小程序的PowerLib图书馆门户小程序设计源码
- 前端分析-2023071100789
资源上传下载、课程学习等过程中有任何疑问或建议，欢迎提出宝贵意见哦~我们会及时处理！
点击此处反馈
