标题：自动驾驶汽车技术：是什么让它慢下来？Self-Driving Car Tech: What's Slowing It Down?

翻译难免有错误，欢迎在评论区指出，我会及时修改的！感谢大家~

(资料图片仅供参考)

正文：

W:This is your Tech news briefing for Friday, April 14th. I'm Zoe Thomas for The Wall Street Journal. Self-driving cars have been a long-awaited promise in the automotive and tech sectors, but we're still waiting. More self-driving cars are being tested in cities from San Francisco to Pittsburgh, but the technology,according to experts, is far from ready. So what's the hold-up? Bart Ziegler wrote about the future of self-driving cars for the WSJ's Journal Reports, and he's with me now to discuss what experts say is stalling its development. Hi, Bart, thanks for joining me.

W：这是您的科技新闻简报，日期为2021年4月14日。我是《华尔街日报》的Zoe Thomas。自动驾驶汽车一直是汽车和科技领域长期期待的承诺，但我们还在等待。越来越多的自动驾驶汽车正在从旧金山到匹兹堡的城市中进行测试，但根据专家的说法，这项技术还远未成熟。那么问题出在哪里呢？Bart Ziegler为《华尔街日报》的Journal Reports撰写了有关自动驾驶汽车未来的文章，他现在与我讨论专家认为什么因素拖延了其发展。您好，Bart，感谢您加入我的节目。

M: Hi!,Great to talk to you.

M：嗨！很高兴和您交谈。

W:You know in San Francisco you do see a fair number of self-driving cars on the road,but they haven't taken over the streets yet. Can you kind of sum up where this technology stands at the moment.

W:您知道，在旧金山的道路上确实可以看到相当数量的自动驾驶汽车，但它们还没有完全占领街道。您能否总结一下当前这项技术的发展状况？

M:Well,it's still in a very early developmental phase.Those robotic taxis you may see in San Francisco are operated under a very special group of rules,they're very limited and they're closely monitored by the companies behind them but so far there's nothing that the ordinary consumer could buy that would be considered a fully self-driving car.

M：嗯，目前仍处于早期开发阶段。你在旧金山看到的那些机器人出租车是根据特定的规则运作的，它们的使用范围非常有限，并且由它们的公司密切监测，但到目前为止，普通消费者无法购买被认为是完全自动驾驶汽车的产品。

W: We've been in the early stages though for a while it feels like there's been this promise that the next stage of self-driving is coming soon.So let's talk a little bit about the hold up you spoke to some experts about this. When did they say fully self-driving cars are coming our way.

W：尽管我们已经处于早期阶段，但似乎一直存在着下一个自动驾驶阶段即将到来的承诺。让我们谈一下这个拖延的问题，您曾与一些专家讨论过此事。他们说完全自动驾驶汽车何时会到来？

M:Well, there seems to be a lot of skepticism at this point whether they ever will come our way. The industry, as you say, has invested tens of billions, if not more, in autonomous driving over the past six, eight, ten years, and there's not too much to show for it. They've run into some obstacles. There are various early levels of what they call driver assist technology. Like many cars today, they have adaptive cruise control, which regulates the car you're driving so it won't get too close to the car in front of it. They have what they call lane-keeping assist, which can center your car in the lane you're traveling in, and if you steer out of it, it'll steer you back in. But that's about the extent of it. There are some higher-end cars that do allow you on certain highways that have been mapped and sometimes photographed to allow you to use hands-free driving. But these cars typically monitor you to make sure you're ready to take control if your eyes wander from the road. They will notice that and set off an alarm so you start looking at the road again. So, this sort of promised level of fully self-driving autonomous cars, where the driver literally could take a nap or watch a movie or sit in the back seat, it's just not there.

M：嗯，目前似乎存在许多怀疑，即它们是否将来真的能够实现。正如你所说，过去六、八、十年中，该行业已经投资了数百亿美元甚至更多的资金进行自动驾驶方面的研发，但这方面的成果并不十分明显。他们遇到了一些障碍。有各种早期的所谓驾驶员辅助技术。就像今天的许多汽车一样，它们具备适应性巡航控制功能，可以调节您驾驶的汽车，使其与前面的汽车保持足够的距离。它们还具有所谓的车道保持辅助功能，可以将您的汽车置于您所行驶的车道中心，如果您驾驶出了车道，它会将您带回车道中心。但那就是全部。有一些更高端的汽车可以在某些已测绘并有时拍摄的高速公路上允许您使用无需手动驾驶。但这些汽车通常会监测您，以确保您随时准备好控制汽车，如果您的视线从道路上漫移开来，它们会注意到并触发警报，以便您再次关注道路。所以，完全自动驾驶的汽车，司机可以躺下睡觉、看电影或坐在后排的那个承诺属于不存在。

W:All right.Well,let's talk about some of the technology that might help us get there.What needs to change then within the self-driving cars to make them do more than just assist us in driving.

W：好吧。那么，让我们谈一下一些可能帮助我们实现这一目标的技术。需要改变自动驾驶汽车内部的什么因素，以使它们能够为我们提供更多的帮助？

M:Well, today they have the cars that do have a bit of assistance. They rely typically on radar and cameras and they do a fair amount of telling the car where it is, where the relationship is with a highway and other vehicles, but to be fully self-driving, they need some other systems. Typically, this one called lidar, which is a laser-based system that basically shoots out laser beams you can't see but they're shooting out from the car every direction and can map an entire representation to where the car is, where other vehicles are, if pedestrians are crossing the street. It creates a map the car can use to safely drive. The other component behind all that is artificial intelligence, and that's where the experts I spoke to said a fair amount of work still needs to be done to really capture what the human brain does. AI would need to be vastly more sophisticated than it is today. Typically, today's AI systems rely on a database of driving situations that they've captured and can work from them to figure out what's happening in real-time. But one of the problems is they cannot map every possible scenario of a car and driver. They're what they're called edge cases that the systems haven't mapped, such as you're driving and suddenly a dog runs in front of the car or suddenly a road worker stands in front of you and puts up a handheld stop sign. These cars rely on past scenarios and may not know how to handle it and could abruptly jam on the brakes, which could cause a rear-end collision with a car behind it, or they could try to steer around it. So that's one of the main concerns - how you deal with these edge cases.

M：今天他们的汽车确实具备一定程度的辅助功能。它们通常依赖雷达和摄像头，并通过告诉汽车自身位置、与公路和其他车辆的关系等方面进行调节。但要完全实现自动驾驶，它们需要一些其他系统。这其中通常使用激光雷达，这是一种基于激光的系统，基本上会从车辆的每个方向发射出您无法看见的激光束，但它们从车辆的各个方向发射出去并可以绘制整个地图，以确定汽车、其他车辆以及行人是否正在穿过街道。它创建一张地图，汽车可以安全地使用这张地图进行驾驶。在所有这些背后的另一个组成部分就是人工智能，这就是我与之交谈的专家所说的需要大量工作才能真正了解人脑的运作方式。人工智能需要比今天更为复杂。通常，今天的人工智能系统依赖于它们捕捉到的驾驶情况数据库，并从中找出发生的实时情况。但其中一个问题是它们无法映射每种可能的车辆和驾驶员情况。例如，当您驾驶时，突然有一只狗跑到汽车前面，或者忽然间一名道路工人站在您面前，竖起手持停止标志。这些汽车依靠以往的情况进行判断，可能不知道如何处理它，并可能突然刹车，导致后面的车追尾，或者它们会试图绕过它。因此，如何处理这些边缘案例是主要关注点之一。

W: We're seeing more sophisticated though. You know, certainly there are concerns that AI is going to do lots of other roles that humans do at the moment, and driving feels pretty simple. You know, when most of us get behind the wheel, it's pretty intuitive, so why couldn't AI just take that on?

W：我们正在看到更加复杂的技术。当然，有人担心人工智能将承担目前人类所做的许多其他角色，而驾驶似乎相当简单。当我们大多数人上路时，驾驶是相当直观的，为什么人工智能不能胜任呢？

M:Well, it turns out, the human brain does an amazing number of things when you're driving, and experienced drivers don't even think about it. You're making split-second decisions - should I speed up, should I steer around something, is it safe to make a left turn at this intersection or not? Your brain has an amazing amount of processing capacity and inputs from your eyes, your ears, and your experience that are awfully difficult to mimic with artificial intelligence. The experts I talked to have said there is ongoing work in what's called detecting human intent, which would be ways to read human expressions, body language, etc., but it's very early on. And one expert said these techniques are far from foolproof, so that is a very important area and one that faces a number of challenges.

M：实际上，当您驾驶时，人脑会完成许多惊人的事情，而有经验的驾驶员甚至不会想太多。您需要做出瞬间决策——我应该加速、绕过某些东西、在这个十字路口左转是否安全？您的大脑具有惊人的处理能力和来自您的眼睛、耳朵和经验的输入，这些很难用人工智能模拟。我谈话的专家们表示，目前正在进行所谓的人类意图检测的工作，这将是读取人类表情、身体语言等方式，但这还处于早期阶段。其中一位专家表示，这些技术远非万无一失，因此这是一个非常重要的领域，面临着许多挑战。

W:We talked a little bit about the technology that needed in the cars.Are there any changes that needed on our roads or Highways to make self-driving more a reality.

W：我们谈到了汽车所需的技术，但是我们的道路或高速公路是否需要做出任何改变，以使自动驾驶更加现实呢？

M: The experts I spoke to said that is an important component of future autonomous driving, and it's in the very early stages. There's a technology they're working on that would connect a car to say traffic lights, weather reports, and police accident reports, and eventually, car-to-car communication where one intelligent car could communicate with another, and it could help the car you're driving in have a better sense of what the cars around it will be doing. But there are only very early versions of that, and one issue is some of them aren't compatible with each other. The experts told me it could be quite expensive to rewire our highways to have all these signals being sent out by things along the roadside.

M：我所采访的专家表示，这是未来自主驾驶的重要组成部分，目前还处于早期阶段。他们正在研究一项技术，将汽车连接到交通信号灯、天气预报和警方事故报告等方面，并最终实现车对车的通信，即智能汽车之间可以互相通信，可以帮助您所驾驶的汽车更好地了解周围汽车的行动。但是，这些技术仍处于早期版本，其中一个问题是它们并不完全兼容。专家告诉我说，重新布线我们的高速公路以使路边设施发送出所有这些信号可能会非常昂贵。

W:All right,that's Bart Ziegler who wrote about this for the WSJ’s journal reports. Thanks for joining us Bart.

W：好的，那是Bart Ziegler为WSJ的《Journal Reports》撰写的。感谢你加入我们，Bart。

M: You're welcome.

M：不客气。

W:And that's it for Tech news briefing this week. TNB's producer is Julie Chang, we had production assistants from Zoe Culkin,our supervising producer is Melanie Roy and our executive producer is Chris sinsley.I'm your host Zoe Thomas.Thanks for listening and have a great weekend.

W：这就是本周的科技新闻简报。 TNB的制片人是朱莉·张，我们有来自Zoe Culkin的制作助手，我们的监制是Melanie Roy，执行制片人是Chris sinsley。我是您的主持人Zoe Thomas。感谢收听，祝您周末愉快。

关键词：