운전자 보조 기술의 미래와 안전 문제

### 운전자 보조 기술의 진화와 과제
* 🚗 차량에 탑승할 때마다 **컴퓨터, 카메라, 센서로 구성된 복잡한 네트워크**가 작동하여 주행 경험과 안전을 향상시킵니다.
* 👀 사각지대 모니터와 같은 기능은 운전자에게 잠재적 위험을 경고하고, **운전자 보조 시스템은 특정 상황에서 운전 작업을 대신**할 수도 있습니다.
* 🛑 미국 고속도로안전보험협회는 이러한 기술이 **추돌 사고를 약 50% 줄일 수 있다**고 보고하며 도로 안전을 개선할 수 있는 잠재력을 보여주었습니다.
* 🔌 도요타와 렉서스 본사의 엔지니어들은 **수백 개의 전선이 다양한 컴퓨터를 연결**하여 제동 및 조향과 같은 기능을 제어하는 시스템 간의 원활한 통신을 가능하게 한다고 밝혔습니다.
* 📷 차량은 여러 대의 카메라와 센서를 사용하여 주변 환경을 인식하고 **레이더와 초음파 센서를 통해 다른 물체와의 거리를 측정**합니다.
* 👁 차량 내부의 운전자 모니터링 시스템은 센서를 사용하여 **손의 위치와 시선 접촉**을 감지하여 운전자의 주의가 도로에 집중되도록 합니다.
* 🧠 렉서스의 팀메이트와 같은 고급 시스템은 **핸즈프리 운전**이 가능하지만 **고급 모델에 주로 탑재되는 추가 컴퓨팅 성능과 LIDAR 센서**에 의존합니다.
* ⚠ IIHS는 기술의 발전에도 불구하고 자동화에 대한 지나친 의존을 경고하며, **기술이 계속 발전하고 있기 때문에 운전자는 비상 상황에서 통제할 수 있도록 항상 준비**되어 있어야 한다고 강조합니다.

운전자 보조 기술은 안전을 향상시키지만 운전자의 의존도와 비상 대응 준비에 대한 우려를 불러일으킵니다.

every time you get in your car, a web of computers, cameras, sensors, and more is working to make you a better driver. That's blind spot monitor saying like, Hey, there's people to your left. This web powers driver assist, a system that lets your car think and sometimes drive on its own. It's truly like you're driving a supercomputer at all times. But as the tech advances, there are questions about whether these automated backseat drivers are actually safe, especially as more cars hit the road without hands on the wheel. This is the tech behind driver assist. I'm not going to do anything, and then it should automatically apply brakes if this car sees a pedestrian before you do. It can hit the brakes on its own. It's a pretty abrupt stop. The Insurance Institute for Highway Safety says that technology like this can reduce rear-end crashes by about 50%, and it's just one of a suite of features that make up driver assist. The vehicle may automatically apply the brakes to prevent a crash, or it may, you know, tug the steering wheel to get you back in

your lane, and so anything that's actively helping the driver is sort of a driver assist feature. To get a closer look at the technology that allows a car to do this, the journal visited the Toyota and Lexus headquarters in Texas. If I were to take out this entire Dash, you would see hundreds, if not thousands, of wires. These wires connect dozens, sometimes hundreds, of computers throughout the car. If we just had radar, but your foot was physically attached to the throttle and it wasn't electronically controlled, we wouldn't be able to stop that as a car. So, just by having one system alone without being able to have the conversation, it's really not. It's just an empty data set. Some control specific functions like the temperature or the throttle on your engine. Others have complex chips that process data from the core Tech that makes driver assist possible, an array of cameras and sensors around the car that allow it to see. The main camera is here. It's going to be a single lens camera up there that's doing multiple fields of view, so it's going to

be scanning out far as well as short. Some cars also have cameras at the back and under the mirrors. These cameras connect to chips that are trained to quickly recognize objects like lanes, cones, and cars. But many cars don't operate on cameras alone. Obviously, there are limitations. Weather can cause problems, low light. That's why many cars also have a variety of sensors. These sonar or ultrasonic and radar sensors send out sound or radio waves. So, as those bounce off, those return to them, and it's able to tell you distance from objects. Then all of that data needs to be communicated back to the car. That radar sensor sends its signal through its wires all the way up underneath the fender into the dash, starts talking to its computers. So, then from inside the vehicle under the dash goes back out here. The request to slow down goes to the engine ECU. So, if you were using cruise control, for example, it sees that we're kind of approaching this vehicle. The radar system knows that that is a very large object in front of us.

and it is constantly calculating the distance between us and that. But while all of that is watching what's happening outside of your car, there's tech watching the inside too. Most cars made in the last decade use a sensor in the steering wheel to make sure drivers have their hands on it. But it knows that my hands are here. If I just touch the sides of it, it understands that there's being a torque applied to it. I am not steering it at all, but it's going to once again start to see how it's pulling me back into my lane. There's no audible alert, and now it's saying, Hey, you're not touching the steering wheel. So, it says Lane Departure Alert. But many cars are also starting to incorporate this. This is called a driver monitor camera. It's a couple of little laser arrays that are constantly looking for the attentiveness of my eyeballs. These systems work together to make sure a driver has their attention on the road. If not, and it's letting me know, and then it's going to

be pulling me back into the lane, but this monitoring can be one of the hardest pieces of driver assist tech to get right. Many of the vehicles will let you drive, um, you know, much longer than we think is comfortable without knowing if the driver is engaged, and one that many automakers are tackling as they move to the next step of driver assistance tech, true hands-free driving. Advanced Drive activated. I have my hands kind of near the steering wheel on my knees. I have both my feet flat on the floor, and it's handling all of this Dallas traffic. This is one of Lexus's higher-end vehicles that does hands-free driving, a system the company calls Teammate. It only has a couple of additions, one being more computing power to process additional cameras and analyze map data. Another big difference is this, an additional sensor called LIDAR. It's essentially a laser array, so it's kind of shooting out lasers. Having a laser array like a LIDAR sensor, uh, being able to judge distance, uh, will be exponentially faster than waiting for a

radar signal to come bouncing back, but LAR is expensive and is usually only available on higher-end models. But even though these systems are on the market now, they have a long way to go. The IHS recently conducted its first test of partially automated systems. Teammate was the only hands-free system tested that earned an acceptable rating; none of them earned a good. Some of the manufacturers are not monitoring the driver well. There are a couple that you don't have to be buckled up to go into partial automation. So, depending on the manufacturer, there are pieces missing in order for them to get our good rating. While the IIHS recommends cars with automated safety features right now, the agency doesn't see any benefit from automation focused on convenience, like hands-free driving. We're concerned that drivers will become over-reliant on them. Even if a system works well most of the time, the driver may not be prepared to take over in an emergency situation. Regulators have investigated hands-free systems from Tesla and Ford after fatal crashes, and while driverless cars are

Hitting the road in some cities, your car is probably still a long ways off from being able to drive itself everywhere. We're making lots of decisions instantly as we drive. Even the way different states mark the road is very different. All of those have to go into a software version to figure out a self-driving car, and it's very challenging.

How Cruise Control, Crash Prevention and Other Driver Assist Tech Work | WSJ Tech Behind

The Wall Street Journal takes you inside carefully selected stories and events in a visually captivating way so you can dig deeper into the news and insights that matter to you. 

The Wall Street Journal

자동차의 운전자 보조 시스템은 컴퓨터, 카메라 및 센서 네트워크를 사용하여 주행을 개선하고 때로는 운전대를 잡기도 합니다.

이 기술은 추돌 사고를 50%까지 줄일 수 있는 잠재력을 가지고 있으며 자동 제동 및 스티어링 휠 조정과 같은 기능을 포함합니다.

차량 내 전선과 컴퓨터의 복잡한 설정으로 인해 다양한 기능을 원활하게 통신하고 제어할 수 있어 운전자 보조 시스템이 효과적으로 작동합니다.

현대 자동차 안전의 기반이 되는 기술

최신 자동차는 카메라와 센서의 조합을 사용하여 물체를 인식하고 거리를 판단합니다.

레이더 센서는 차량의 컴퓨터와 통신하여 다른 차량에 접근하면 속도를 늦추는 등 실시간으로 조정합니다.

또한 실내 센서와 운전자 모니터 카메라는 함께 작동하여 운전자의 주의력을 확인하고 필요할 때 경고를 제공합니다.

핸즈프리 주행 기술의 도전 과제

핸즈프리 주행 시스템은 운전자의 참여를 효과적으로 모니터링하는 데 어려움을 겪고 있어 잠재적인 안전 문제로 이어질 수 있습니다.

렉서스의 팀메이트 시스템은 추가 컴퓨팅 성능과 LIDAR 센서를 사용하여 테스트에서 만족스러운 평가를 받은 몇 안 되는 시스템 중 하나입니다.

그러나 IIHS는 편의성을 위해 자동화에 지나치게 의존하는 것에 대해 경고하며, 운전자가 비상 상황에서 운전대를 잡을 준비가 되어 있어야 한다고 강조합니다.

자율주행차는 아직 갈 길이 멀기 때문에 인간 운전자가 도로 위에서 즉각적인 결정을 내려야 합니다.

주마다 도로 표시가 다르기 때문에 자율주행을 위한 소프트웨어 버전을 만드는 작업은 더욱 복잡해집니다.

이러한 요소들의 조합으로 인해 완전한 자율 주행 자동차를 개발하는 것은 어려운 일입니다.