一批汽车厂商已经承诺将在2022年前使自动紧急制动（AEB）系统成为一项车辆标配，此举极有可能改变车辆的电子架构，并促进不同模块供应商之间的合作。

最近，20家汽车生产商与美国国家公路交通安全局（NHTSA）和美国公路安全保险协会（IIHS）达成自愿性协议，将在未来几年内推动紧急制动系统从豪华车型向大众车型的普及。 

为了实现这一点，很多汽车平台将利用已经上路检验过的硬件和软件。然而，这些系统还需经过一些调整才能在成百上千款车型上发挥高效性能。

“未来我们的精力将主要集中在对系统在不同平台上的表现进行验证与测试，这并不是无关紧要的工作。”大陆汽车系统项目管理负责人Dean McConnell 表示，“另一个问题在于，真正的功能性要求是什么？车辆变速范围、制动距离及一些其他性能，都将成为不同系统之间拉开差距的因素。”

根据IIHS的估计，美国每年发生车祸的数量大约在500万起，而AEB技术最高可以将车祸数量降低20%。最近，在位于瑞典阿尔维斯尧尔市附近的采埃孚天合(ZF TRW) 测试场举行的冬季测试期间，《汽车工程》的总编有机会体验了自动紧急制动系统的优势。

虽然已在先前的干路面测试环节中试驾了多款来自不同供应商的AEB原型车，但总编还是第一次在光滑的冰冻湖面上驾驶配备AEB系统的车型。试驾结束后，这位总编相信，这项技术可以在事故发生前激活限制系统，在预防车祸以及降低车祸严重程度方面具有很高价值。见以下视频。

繁重的工作量

对于汽车行业而言，AEB系统的推广涉及多个不同模块。传感器必须要能保证车辆能够注意到潜在威胁。当即将发生碰撞时，车辆的多个系统将在不让车辆打滑的前提下协同工作，尽量降低车速。

“一般来说，这需要摄像头、雷达等传感器，配合控制器、软件和算法协同工作。”博世集团创新及底盘系统控制总监 Frank Sgambati 表示，“这些元件都必须与车辆的制动系统和电子稳定系统 (ESC) 协同工作，所有操作都必须保持协调一致。”

在短短6年内将AEB系统推广至所有车型并不容易，这需要进行大量的设计和研发工作。AEB系统中的模块通常来自多家供应商。未来在不同供应商之间、供应商与汽车厂商之间，都极有可能展开更加紧密的合作。

“实现目标的主要挑战在于繁重的工程量，也就是为了将AEB技术推广至数百种车型所需进行的各项工作。”采埃孚天合全球电子部产品规划总监Aaron Jefferson指出，“这也进一步对不同供应商间的高效合作提出了更高要求。”

需要更强劲的控制器吗？

从一种升级配置到一种车辆标配，自动紧急制动系统地位的变化也有可能改变车辆的电子架构。目前，分布式计算在汽车行业中相当常见，车辆的传感器和控制器内均配备了微型处理器。当所有汽车都采用更加通用的紧急制动系统时，这种后加式模块化AEB功能可能会逐渐消失，转而走向系统整合的道路。

“作为车辆下线后加装的独立升级配置，这些功能模块在设计时必须拥有自己的独立控制器。” McConnell 表示，“而一旦AEB系统成为一种标准配置，工程师就可以省去不同模块的冗余设置。当然了，出于安全考虑，车辆仍将保留一部分冗余。”
 

这将改变汽车行业对半导体的需求。目前，高级驾驶员协助系统在数据到达控制模块前，主要是依赖智能传感器做出决定。AEB系统可以减少此类分布式计算的应用，从而将精力更多放在功能更强大的控制器上。

“如今的高级驾驶员协助系统是分散的，每个传感器都拥有自己的大脑。”瑞萨电子美国分公司高级驾驶员协助系统解决方案细分营销经理 John Buszek 表示，“厂商也许会选择减少一些分散式“大脑”的部署，转而将汽车智能设备集中起来，而这或许意味着必须使用功能更强大的控制器。”

提升计算能力和增加内存容量可能并不是芯片制造商需要面对的唯一变化。过去几年中，半导体制造商一直在提供更多软件。AEB系统的普及极有可能拓宽 AUTOSAR（汽车开放系统架构）的适用范围，这种标准可以帮助厂商兼容来自不同供应商的模块。

“对很多厂商而言，抢占产品上市时间非常重要。” Buszek 解释说，“提高上市速度的一个方法就是利用标准，因为标准可以提供共同的基础，加速各项工作的推进。目前已经存在一些支持 AUTOSAR 的呼声，我认为未来将出现更多AUTOSAR的推动者，因此硅谷公司未来可能会面对开发此类软件的需求。”

虽然AEB协议正在从很多方面影响着车辆及车辆次级系统的工程研发，但大多数观察者均认为，这与汽车厂商为了应对美国不断收紧的强制性燃料经济性规定而进行的努力是不同的，并不会引起大规模的颠覆性改变。AEB系统主要是借助现有技术提高车辆安全性。

“AEB系统将大幅降低车祸发生的概率。” Jefferson 表示，“这种系统已经进入市场，运转良好，且能带来安全效益，但目前仅能造福一小部分汽车，这就是汽车厂商签订协议，促进AEB推广的原因。”

作者：Terry Costlow
来源：SAE《汽车工程》杂志
翻译：SAE上海办公室

Automatic braking pact for 2022 will alter system development

Vehicle OEMs have pledged to make automatic emergency braking (AEB) a standard feature by 2022, a move that is likely to alter electronic architectures and increase collaboration between module suppliers.

Twenty automakers recently teamed up with the U.S. National Highway Traffic Safety Admin. and the Insurance Institute for Highway Safety to set voluntary programs to move emergency braking from a luxury option to mainstream vehicles over the next few years. 

For that to happen, many vehicle platforms will leverage hardware and software that’s already being proven on roadways. However, these system will all have to be tweaked to provide efficient performance on hundreds of different vehicle models.

“The main effort will be to validate and test systems on new platforms, which is not an insignificant job,” said Dean McConnell, Project Management Leader atContinental Automotive Systems. “The other question mark is what the actual functional requirements are, the speed range, stopping distance and other functions that will be important differentiators.”

IIHS estimates that as many as 20% of the 5 million vehicle crashes that occur annually in the U.S. could be prevented by the technology. The advantages of AEB were recently demonstrated to Automotive Engineering's Editor-in-Chief during a winter-test session at ZF TRW's proving ground near Arvidsjaur, Sweden.

Having driven AEB-equipped prototypes from various suppliers in previous dry-surface tests, this was the editor's first experience with AEB on a low-mµ-surface frozen lake. He came away convinced of the technology's high value in avoiding crashes as well as reducing the severity of crashes that may not be avoidable, by prepping the restraint systems before impact (seehttps://www.youtube.com/watch?v=8RxifQ_UDoA&feature=youtu.be).

Taxing engineering capacity

For the industry, the engineering task to implement AEB will encompass a number of different modules. Sensors will have to guarantee that they’re seeing something that’s a potential threat. When a collision is imminent, several systems will have to work harmoniously to slow the vehicle without sending it into a skid.

“There’s generally a combination of sensors like cameras and radar, along with controllers, software and algorithms,” said Frank Sgambati, Director of Innovation, Chassis Systems Control, at Robert Bosch. “They all have to work with vehicle braking systems and electronic stability control, all operating in concert.”

Outfitting all these vehicles with AEB in just six years will require a lot of design and development effort. Modules will often come from multiple suppliers, which is likely to foster closer working agreements between suppliers as well as OEMs.

“The main challenge to meeting this goal is engineering capacity — the amount of work needed to add AEB technology to several hundred vehicle models,” noted Aaron Jefferson, Director, Product Planning for ZF TRW Global Electronics. “This further enhances the need for multiple suppliers to work together efficiently.”

More powerful controllers needed?

Shifting from option to standard function may also alter electronic architectures. Distributed computing is currently quite common, with microprocessors deployed in both sensors and controllers. When all cars have more generic types of emergency braking system, the need to use modules that can be added as options can disappear, opening the door for integration.

“The need to design for options is part of why the strategy has gone towards separate controllers for different functions,” McConnell said. “The redundancy of separate modules may go away once it’s a standard feature. Some redundancy will still be required because it’s a safety function.”

That could alter requirements for semiconductors. Advanced driver assistance programs now rely on intelligent sensors that make some decisions before data is transferred to control modules. AEB could reduce the usage of this type of distributed computing, putting more focus on more powerful controllers.

“ADAS today is decentralized; the sensors all have brains,” said John Buszek, Segment Marketing Manager for ADAS Solution at Renesas Electronics America. “OEMs may decide to remove some of those brains and centralize intelligence, which will probably mean using more powerful controllers.”

Boosting computing power and memory size might not be the only change for chipmakers. Over the past several years, semiconductor makers have been providing more software. The standardization of AEB is likely to broaden the role of AUTOSAR, a standard that helps OEMs employ modules from multiple suppliers.

“The big thing for many OEMs will be time to market,” Buszek explained. “One way to move faster is to use standards — they provide a common base so work can proceed quickly. AUTOSAR already has momentum, I think there will be more of push behind it, so silicon vendors will probably be asked to create that software.”

Though the AEB agreement may impact a number of aspects of vehicle and subsystems engineering development, most observers feel that it won’t cause the type of scrambling seen in some efforts to meet tighter fuel economy regulations mandated by the U.S. The voluntary pact relies on existing technology to improve safety.

“AEB will greatly help reduce the number of accidents,” Jefferson said. “The AEB agreement is based upon the foundation that these systems are in the market today, performing well and providing safety benefits, but on a relatively small scale."

Author: Terry Costlow
Source: SAE Automotive Engineering Magazine