交通基础设施专家 Kirk Steudle 反思了网联化自动驾驶汽车快速发展的现状及未来面临的挑战。

“汽车机械和电气工程师的工作领域是车轮上方，而土木工程师则是车轮下方。”Kirk Steudle 开玩笑说，“尽管双方都得花点时间才能理解对方的行话，但由于网联自动驾驶汽车的出现，这两个行业必须团结起来，共同解决社会大众的交通出行问题。”

Kirk Steudle 是一名长期耕耘在“车对路”领域的注册工程师，去年 10 月从密歇根州运输部长一职光荣退休，职业生涯中曾取得了无数成就。在 Steudle 的领导下，密歇根州运输部大力推动网联汽车和自动驾驶技术的发展，该州也迅速在相关技术的开发、测试和基础设施标准制定方面走在了全美前列。此外，Steudle 还在退休前一年兼任美国交通运输中心（American Center for Mobility，下简称 ACM 中心）的临时首席执行官。ACM 中心占地 500 英亩，位于安阿伯市附近，是全美自动驾驶汽车试验和测试的主要场地之一。

Steudle 在退休后加入交通网络工程与集成方案提供商 Econolite 公司，担任公司交通系统集团高级副总裁。近期，Steudle 与《自动驾驶汽车工程杂志》探讨了交通行业目前的发展状况及未来可能面临的挑战。

“回首 2006 年，当时智能手机还没有诞生。美国国防高级研究计划局（DARPA）的确已经推出了自动驾驶汽车挑战赛，但各州的交通运输主管部门所关注的并不是自动驾驶或自主驾驶，而是网联汽车及相关基础设施。”Steudle 回忆道，“此外，当时还有大量有关‘国家为什么也要参与其中’的讨论，有人问道：为什么汽车行业不能自己玩自己的？毕竟交通运输部门也管不到安全带或稳定性控制。”

事实证明，政府与行业的合作对加快技术推广至关重要。福特（Ford）和通用汽车（General Motors）合作开展的“防撞度量标准伙伴关系（CAMP）”就曾携手美国高速公路安全管理局（NHTSA），推动了防撞措施在乘用车中的部署以提升交通安全，并证明了 V2I（车对基础设施）技术的重要作用。

2011 年，美国运输部在安娜堡推出了安全试点模型部署项目（Safety Pilot Model Deployment），“真正将合作的主题暴露在所有人的目光之下，”Steudle 表示，“事实证明，我们可以让不同的汽车制造商开始交流，并且取得不错的结果。”在智能交通系统世界大会（ITS World Congress）有关“安全试点”项目的演示环节中，美国运输部长 Ray LaHood 从演示车中款款走出，并宣布：“这将挽救无数民众的生命。”Steudle 介绍说，“运输部长明白，这可能会改变游戏规则。”

交通信号和通信

Steudle 断言，“技术演示（SAE 也曾安排过此类自动驾驶汽车演示活动）有助于揭开自动驾驶技术的神秘面纱，从而推动自动驾驶技术的发展。因此，这类演示多多益善。”不难想象，让政策制定者真正先坐到网联汽车和自动驾驶汽车中非常重要。他补充说，在这些自动驾驶汽车试驾中，“为车辆配备一名测试驾驶员可以在一定程度上让乘客更放心”。

Steudle 观察到，很多州市都呈现出斥巨资升级交通信号系统以使其适用于互联自动驾驶汽车的趋势。“目前，蜂窝网络 V2X/5G 和DSRC 通信技术之间尚未决出胜负。坦率地说，任何基础设施提供商或哪个州的运输部均无法独立作出决定，到底如何选择要看汽车公司和电信公司的共同决定。”Steudle 补充说，美国政府与中国不同，“不会直接选定一种技术，然后统一推行。”

但 Steudle 同时指出，由于蜂窝网络 V2X/5G 与 DSRC 通信技术之间的争论尚未尘埃落定，公共机构在进行交通信号设施投资时纷纷不约而同地采用了“对冲”的思路。“如果确定未来到底会用哪种技术，实施起来会更快。但大家必须慎之又慎，不能花了大价钱，却买了一堆很快将被淘汰的技术，就比如早些年间的‘Betamax’卡带。”在此背景下，一些机构正在同时投资双种通信技术，“以确保旗下产品可以为未来的各种可能性做好准备。”Steudle 认为，蜂窝网络 V2X/5G 通信技术将是终极解决方案，但现在还为时尚早。

在 Steudle 担任部长期间，自动驾驶汽车的公路测试是密歇根交通部“最主要的辩论话题之一”。他解释说，“现在大家的讨论重点是大家能接受多少起事故？但无论如何，如果什么都不做，每年将继续有大约 37,000 人因交通事故身亡。从长远来看，我坚信自动驾驶汽车将有助于减少道路交通事故。”

“到 2025 年，一些新技术的出现和现有技术的大幅提升将给交通出行行业带来巨变，比如车辆导航系统精度提升等。”Steudle 预测，“到时候会有很多可以在高速公路工况下提供脱手驾驶功能的车型，可以替代驾驶员完成绝大部分工作，但它们还并不是真正的自动驾驶汽车。”这也是 Steudle 认为业界必须区分的两个重要概念。

建立公众对自动驾驶技术的信心

最近，《自动驾驶汽车工程》与 ACM 中心首席技术官兼首席安全官 Jeff Rupp 探讨了 ACM 中心在制定互联和自动驾驶车辆标准方面的作用。

“我认为 ACM 在制定新标准方面可以发挥的关键优势在于，帮助我们的行业客户在标准尚未真正落地之前就能开始探索与测试。”Rupp 说，“随着新流程逐渐成型，我们可能成为首个有能力进行测试，并提供公开可用数据的单位。我们可以为行业提供一些信息，这有助于互联和自动驾驶汽车行业赢得公众的信心。"

Rupp 指出，新的自动驾驶汽车相关标准仅适用于整个车辆系统，也包括相关通信技术，比如 DSRC、V2V 或 C-V2X 等。“目前，行业在选择通信技术方面尚未达成统一，也没有哪种技术占据绝对优势。各家公司都马不停蹄地不断探索。”Rupp 观察到，“行业认识到，最终大家必须达成共识，也就是对安全有共同的定义。公司可能很难判断何时该从相互竞争、各自保密的传统，开始转变为数据共享和共同协作的新模式。NHTSA 已经开始向这一目标努力。大家都想这么做，也意识到必须这样做。”

Transportation-infrastructure expert Kirk Steudle reflects on the rapid progress toward the connected-AV future and the challenges ahead.

“In the auto industry, mechanical and electrical engineers work from the tire up. Civil engineers work from the tire down,” quips Kirk Steudle. “And although it’s taken a while for both to understand each other’s jargon, the connected-and-autonomous vehicle is forcing them all together—to solve problems for society’s mobility.”

Steudle, a registered professional engineer, has been at the vehicle-to-road nexus his entire career. Last October he retired as director of the Michigan Dept. of Transportation where, among many achievements, his leadership helped put the state and its core industry in the vanguard of connected-car and AV-related developments, testing, and infrastructure standards. During his final year at MDoT, he also served as interim CEO of the American Center for Mobility, the new 500-acre AV proving ground and test center near Ann Arbor.

Now senior VP of the Transportation Systems Group at Econolite, a traffic-network engineering and integration company, Steudle spoke with SAE’s Autonomous Vehicle Engineering about transportation’s evolution and the challenges ahead.

“Looking back to 2006, ‘smart’ phones hadn’t yet been invented. The DARPA Challenge for self-driving vehicles was happening, but at the state DoT level we weren’t talking about automated or autonomous driving; we were talking about connected vehicles and the infrastructure,” he recalled. “And there was significant conversation asking why should states be involved: Why can’t the auto industry work this out themselves? After all, DoT’s aren’t involved with seat belts or stability control, went the argument.”

Government-industry collaboration has since proved to be vital in accelerating progress. The Crash Avoidance Metrics Partnership (CAMP) between Ford and General Motors, working with NHTSA, helped implement crash-avoidance countermeasures in passenger cars to improve traffic safety, and proved V2I (vehicle-to-infrastructure) technology could work. The U.S. DoT’s Safety Pilot Model Deployment, launched in 2011 in Ann Arbor, “put this subject on the front stage,” Steudle said. “It proved that we could have different auto- makers talk to each other and that it could be successful.” 

In a demonstration at ITS World Congress that sprung from the Safety Pilot program, U.S. Transportation Secretary Ray LaHood stepped out of a demo vehicle to declare, “This is going to save people’s lives,” Steudle explained. “He understood that this could be game-changing.”

Traffic signals and communications 

Getting policymakers into the seats of connected and autonomous vehicles is important, because the demos (such as those conducted by SAE) “help demystify the technology. The more demos we can do, the better,” Steudle asserted. “Having a test driver behind the wheel provides a level of comfort to the passenger” in these activities.

Steudle sees a trend in states and cities investing heavily in their traffic-signal systems—updating the technology and adapting it for the increased use of connected AVs. “The competing communications protocols—cellular V2X/5G and DSRC—still have to be worked out. That’s frankly not going to get solved by an infrastructure provider or a state DoT,” he explained. “It’s going to get decided by the auto companies and the telecoms.” Unlike in China, government in the U.S. “is not going to select and mandate a technology.”

But he noted that public agencies are hedging on their traffic-signal investments because of the C-V2X vs. DSRC battle. “If we knew which technology we’re going to use, implementation could accelerate. But the agencies have to be careful they’re not buying a bunch of Betamax.” As a result, some are specifying dual communication modes—”to make sure their stuff is ready for the future.” Steudle believes that cellular/5G is the ultimate solution—when it arrives.

The subject of on-road testing of AVs was “a huge debate” at Michigan DoT during Steudle’s tenure as director. “The big public discussion is, how many crashes are too many?” he observed. “But if we do nothing, some 37,000 people will continue to be killed each year. In the long run, I believe autonomous vehicles will help reduce road fatalities.”

Looking ahead six years, Steudle forecasts “technologies not invented yet, and far greater precision in those currently employed such as vehicle guidance, will trans- form a lot of things related to mobility. By 2025,” he said, “I think we’ll have lots of models offering true hands-free driving for freeway use. They’ll be heavily driver-assisted, not autonomous”—an important distinction which he believes industry needs to better define. 

Building public confidence in AV tech 

During a recent visit to the American Center for Mobility, AVE spoke with Jeff Rupp, the CTO and chief safety officer, about ACM’s role in creating standards for connected and self-driving vehicles.

“I think ACM’s niche related to new standards under development is in helping our industry clients explore and try out new test methods or procedures before they become a standard,” he said. “As new procedures develop, we may be the first to test them out and show publicly-available data. We would provide some transparency to the industry and in doing so help the public develop confidence in what the industry’s doing.”

New AV-related standards will only apply to the whole vehicle system—which Rupp noted might include off-board communication—DSRC, V2V or C-V2X. “There’s no standardization yet, no winning technology. Everyone’s still exploring at breakneck pace,” he observed. “The industry recognizes that eventually there needs to be a point where collectively we say, ‘How safe is safe enough?’ It’s tough for these companies to decide when it’s okay to move from a competitive, secretive mode, to a data-sharing and collaborative mode. NHTSA’s trying to encourage that now. Everyone wants to do it and know they need to do it.”

Author: Lindsay Brooke

Source: Autonomous Vehicle Engineering