Page 36

EETE NOV 2013

AUTOMOTIVE ELECTrONICS Adaptive vehicle software architecture to correct malfunctions In modern cars, software controls many functions - including safety-critical ones. For drivers (and for the industry, too) it would be a horror vision if a software bug would trigger a potentially fatal accident. Researchers now are developing a software architecture that compensates for such malfunctions. Increasingly mechanical components in vehicles are replaced and displaced by electronic controls. The deployment of these x-by-wire systems does not stop short of safety-critical functions. A standard approach is to safeguard such functions by redundancies, at the expense of costs and energy consumption. For this reason, Fraunhofer Institute for Embedded Systems (Munich) has joined with a group of companies including Delphi Germany, Duracar, Fico Mirrors, Tecnalia, Pinifarina, Siemens and TTTech and launched the SafeAdapt research project that aims at developing an adaptive software architecture for vehicles which avoids these shortfalls. This adaptive system is based on the standardized automotive software environment Autosar and supports safeguard processes according to ISO 26262. To enable developers to immediately utilize this architecture, the consortium develops the corresponding design methodology. This includes the early definition of an abstraction which eases and streamlines the vehicle design process. The approach to create the necessary safety measures in software instead of hardware redundancies enables designers to omit additional control units in the vehicles and thus reduce complexity, cost and not least weight. With SafeAdapt, developers only define the degree of adaptivity instead of describing each single potential scenario. It facilitates reducing hardware redundancies since it eliminates the need to have a second, redundant control unit in standby mode. Instead, any other ECU which is not busy at the moment can execute the respective software function. Delphi www.delphi.com Cell monitoring chip simplifies balancing of li-ion battery cells In battery electric and hybrid vehicles, it is necessary to implement a robust cell balancing mechanism to optimize power output of lithium-ion batteries. Typically, this balancing requires a powerful microprocessor running a complex software. Now chipmaker ams has introduced a cell monitoring IC that significantly re-duces the hardware requirements in battery cell management systems. The AS8506 performs distributed cell monitoring and balancing operations for stacked cell modules, including Safe Operating Area (SOA) checks and passive or active cell balancing. It suits for all lithium-based cell chemistries, such as those found in hybrid and fully electric vehicles, as well as for supercaps (EDLCs). In conventional systems, a complicated algorithm running remotely on a high-end microcontroller decides which cells have to be balanced. The architecture implemented in the AS8506 can control balancing locally at the cells, enabling designers to create a more streamlined cell management system which eliminates the powerful host controller, complex software and vulnerable serial communication links normally used today. The AS8506 can implement both passive and active cell balancing autonomously, or it can support a microcontroller-based system via its Serial Peripheral Interface. An advanced analog circuit in the AS8506 compares up to seven cell voltages against an internal or external reference with an accuracy of 1mV, to support cell-balancing and cell-monitoring functions. Cell voltage measurements can also be digitized with an accuracy of 5mV and reported to a host controller. Active and passive cell balancing use a similar circuit design, but active balancing requires an additional flyback transformer. The control circuit is integrated in the AS8506. The device also features internally or external adjustable upper and lower cell voltage limits. Temperature measurement is carried out through two external NTC sensors. AMS www.ams.com. TI rolls heterogeneous number crunching SoC for automotive video processing Cameras and image processing functions along with downstream sensor fusion algorithms increasingly become the technology of choice when it comes to developing intelligent driver assistance systems which detect and intelligently recognize the surroundings of cars in the flowing traffic. Image processing however demands huge amounts of computing power. Texas Instruments has developed a family of System-on-Chip devices combine the number crunching abilities of a DSP with the properties of a general-purpose microprocessor. With the TDA2x SoCs, Texas Instruments (TI) follows the approach of a heterogeneous multicore computing resource. With two C66x DSP cores of the latest generation, up to four embedded vision accelerator cores (EVEs) with a capacity of 10.4 GMACs each, two ARM15 cores and two dual ARM M4 cores (for “housekeeping” tasks inside the chip only), the device should offer enough power to process the video input of up to six cameras in real-time and forward it to multiple displays. In addition, this computing environment at chipscale level contains a graphics engine, a display subsystem, a vision acceleration unit and several more functional units and memories. The device is manufactured in 28 nanometre process technology which enables the high degree of integration we see in these chips. The entire chip consumes 2 to 4 Watt of electric power, where the lion’s share certainly is allocated to the ARM15 cores. The three versions of the TDA2x family are optimized for different flavours of image-processing tasks in the car - front camera applications such as lane assist, park assist and sensor fusion to blend signals from cameras, radar and other sources. “Sensor Fusion computers will probably be the most powerful systems in future car generations”, believes Frank Forster, ADAS Europe Marketing and System Applications Manager at Texas Instruments. Texas Instruments www.ti.com 28 Electronic Engineering Times Europe November 2013 www.electronics-eetimes.com


EETE NOV 2013
To see the actual publication please follow the link above