Did the first fatal accident sound the death knell for Tesla's autonomous driving technology?

The U.S. government announced on June 30 that the National Highway Traffic Safety Administration (NHTSA) is conducting a preliminary evaluation of the performance of the Model S's autopilot function. The reason is a traffic accident in Florida on May 7 in which 15 Tesla Model S cars collided with a tractor-trailer.

In this accident, the Tesla owner had activated the autopilot function, but the car's recognition system failed to recognize the trailer in front of it, causing the roof of the Tesla to be completely lifted off when it passed under the trailer, and the driver, Joshua Brown, died on the spot. This is also the first fatal accident caused by Tesla's Autopilot autopilot function since the OTA upgrade last year.

The deceased owner, Joshua Brown, was a die-hard fan of Tesla's autopilot function. He had said publicly many times that Tesla was the best car he had ever owned and that he wanted to use the autopilot function to the fullest. The driver of the trailer said that the deceased was watching Harry Potter on a portable DVD in the car when the accident happened.

Tesla: The accident was caused by recognition error, implying user error

Tesla later said in a statement that the accident occurred because the camera failed to recognize the white trailer against the bright sky, so no braking measures were taken. It also reiterated that the autopilot function is still in the public beta stage, and Tesla's cars always remind drivers to put their hands on the steering wheel and be ready to take over the car at any time when the driver enables the autopilot function.

This is actually Tesla's subtle way of saying that the deceased Joshua Brown mistakenly handed over control of the Model S, ignoring Tesla's warnings and the fact that Autopilot is only an assisted driving system rather than a true autonomous driving technology.

Tesla also stated in the statement that this was the first fatal car accident after the autopilot had safely driven more than 130 million miles, and that on average, a fatal car accident occurs every 60 million miles. In other words, although Tesla's autopilot caused this accident, its driving ability is still much better than manual driving.

Unlike Google, Audi, and Mercedes-Benz, which have been experimenting and debugging their autonomous driving systems privately, Tesla is the first manufacturer to apply autonomous driving to commercial products. Unlike Google, which has been investing for several years but has only millions of kilometers of experience on the road, Tesla's autonomous driving function has accumulated hundreds of millions of miles of practical data even though it is only in public beta. Tesla said that this will also enable its autonomous driving to recognize more uncommon scene experiences and avoid tragedies like this accident from happening again.

Tesla's autopilot has become a safety hazard on the road

Since Tesla's v7.0 system upgrade brought about autonomous driving last year, car lovers and technology geeks have collectively been excited, shouting that the future is now. Google, which has been developing autonomous driving systems for a long time but is still far from being commercialized, has also been slapped in the face. However, Google, which has always been willing to spend money and time on scientific research, is very calm and is still slowly conducting tests. Facts have proved that Google may be right, and it is still uncertain who will have the last laugh.

Tesla has been promoting assisted steering, traffic-aware cruise control, automatic parking, automatic lane change and other functions. Although they are called Autopilot, they are still far from true autonomous driving and can only be called assisted driving systems. Every time the driver switches from manual driving to autonomous driving mode, Tesla will tirelessly inform the driver that this assistance system is still in the testing phase, and remind the driver to keep both hands on the steering wheel and be ready to take over control of the vehicle at any time.

Tesla claims that this practice is to allow Tesla's autopilot system to collect data on the road in advance, so that the software logic algorithm can continuously improve the safety of autopilot by accumulating operating experience. However, in fact, almost no one will strictly follow Tesla's requirements. There are countless videos on YouTube of drivers driving dangerously without their hands on the steering wheel. Joshua Brown, the victim of this accident, also uploaded a video to the Internet without holding the steering wheel.

Tesla's self-driving car is over-hyped, but it's not as good as it seems

Tesla's problem is not only that users find it difficult to comply with or choose to ignore the rules it sets. More seriously, compared with the little-known Mercedes-Benz autonomous driving technology, which has not yet launched a civilian product on the market, Tesla's autonomous driving technology, which has entered the public beta stage, can hardly be said to be advanced.

Tesla's self-driving technology recognition mechanism is actually very simple. Its main recognition targets are lanes and the position of the vehicle in front. The recognition device mainly uses a camera located on the windshield using the MobileEye solution, and a radar that can measure 150 to 200 meters below the front of the car. There is a parking radar on the side of the car that traditional manufacturers use for parking. Tesla engineers use it as an auxiliary side radar with a very narrow recognition angle. However, these devices are very unstable and are easily affected by factors such as rain.

The so-called automatic driving mechanism is even simpler, and the main strategy is to identify the lane and follow the car in front. When encountering some unfavorable factors that make it impossible to identify the lane, such as rainy days or when the car is driving on a country dirt road, Tesla can only choose to follow the car in front. In other words, Tesla's automatic driving is completely based on the belief that the driver of the car in front will not make mistakes. If the driver of the car in front falls into a pit, you can only follow. Or like this accident, the car in front was not identified, and the trailer with a higher chassis was not identified because the front radar was too low. Although it is relatively rare, the driver can only blame himself for his bad luck.

Were they the first to enter the civilian market simply because they were brave?

Why was this immature autonomous driving solution put into the civilian market so early? Is it Tesla's marketing trick, or is Tesla overly confident that users will follow the "friendly tips"? We have no way of knowing, but what we do know is that it is definitely inaccurate to say that Tesla is wrongly confident in this technology.

The US government has issued 29 self-driving car licenses, allowing their holders to legally test self-driving cars on public roads. Google received 25 of these 29 licenses, while Audi and Mercedes-Benz received two each. In other words, Tesla, which did not get a license, should be well aware that its technology is only at the primary stage of assisted driving, far from the concept of true driverless driving.

Compared with Tesla's so-called autonomous driving, which requires drivers to keep their hands on the steering wheel, Google's new driverless car has no steering wheel or accelerator or brake pedal. The Mercedes-Benz S500's autonomous driving successfully passed the test of nearly 100 kilometers of complex road conditions from Mannheim to Pforzheim in Germany three years ago, during which the driver completely handed over the control to the S500.

Tesla is fully aware that its autonomous driving technology lags far behind several industry pioneers, and Elon Musk has repeatedly stated that he is willing to work with Google to develop driverless technology.

In pursuit of technological progress, we should also protect the rights of consumers

The emergence of Tesla has indeed driven the advancement of electric vehicle technology, and therefore Elon Musk has always been favored by investors. The popularity of the concept of autonomous driving does show that it represents the direction of technology, but it does not mean that blindly pursuing the development of technology is correct.

The idea of ​​using users to collect data is not a bad idea, but Tesla should have anticipated that putting such high-tech in the hands of users would be abused, not to mention that it does have its own flaws.

As the saying goes, a car accident is more ferocious than a tiger. If Tesla does not take this car accident seriously, it will cause users to resist and fear the concept of autonomous driving, which will only backfire and make it more difficult to popularize autonomous driving technology. The arrival of true autonomous driving is still a long way off, and even Google cannot give a specific timetable. Tesla's so-called "autonomous driving" is just a prototype of assisted driving. I hope there will be no more tragedies like Joshua Brown.

As a winner of Toutiao's Qingyun Plan and Baijiahao's Bai+ Plan, the 2019 Baidu Digital Author of the Year, the Baijiahao's Most Popular Author in the Technology Field, the 2019 Sogou Technology and Culture Author, and the 2021 Baijiahao Quarterly Influential Creator, he has won many awards, including the 2013 Sohu Best Industry Media Person, the 2015 China New Media Entrepreneurship Competition Beijing Third Place, the 2015 Guangmang Experience Award, the 2015 China New Media Entrepreneurship Competition Finals Third Place, and the 2018 Baidu Dynamic Annual Powerful Celebrity.