On Thursday (February 17), local time, the National Highway Traffic Safety Administration (NHTSA) announced that it would launch a formal investigation into Tesla Motors. The reason is that the agency received a complaint that Tesla's autopilot assist system Autopilot had an accidental activation of the brakes.
This phenomenon is often referred to as "Phantom Braking," in which a user turns on the Autopilot Assist system of a Tesla vehicle when the vehicle brakes suddenly and without warning without warning. Because "ghost brake" events often occur on highways, vehicles slow down sharply, increasing the risk of colliding with trailing vehicles.
In its notice of investigation, NHTSA noted that it had received 354 complaints related to ghost brakes in the past 9 months. Complainants report that while driving a Tesla, rapid deceleration can occur randomly and often repeatedly over a single driving cycle.
It is reported that the survey involved Tesla's 2021 and 2022 Model 3 and Model Y models, involving a total of about 416,000 cars, which are equipped with the company's autopilot assistance system Autopilot.
NHTSA revealed that the incident will be preliminarily assessed by the NHTSA Defect Investigation Office (ODI). The initial assessment mainly involves a review of service announcements and owner complaints. Based on the results of an engineering analysis by the Defect Investigation Office, the investigation may be closed or escalated to a recalled vehicle.
This is not the first time NHTSA has launched an investigation into Tesla. In August last year, NHTSA announced that due to multiple accidents in which Tesla vehicles collided with parked vehicles, the agency has begun a formal investigation into Tesla's Autopilot automatic driver assistance system, which will involve 765,000 vehicles, including Model Y, X, S and 3 models.
Last October, Tesla said that the "ghost brake" failure phenomenon was triggered by an FSD software update. As a result, Tesla recalled the FSD software version that was updated at the time. However, after the recall, the complaint against the "ghost brake" did not go down but rose.
The time node of tesla's "ghost brake" complaint surge coincided with its announcement to stop using millimeter-wave radar as a sensor for its autopilot system, so some insiders judged that the frequent failure of Tesla's "ghost brake" was related to the pure visual autopilot perception system it insisted on building.
Last July, Tesla released the FSD Beta V9 system, which is considered an important step in Tesla's official move towards a truly pure visual self-driving route, as Tesla announced that it would cancel millimeter wave radar and rely solely on cameras to achieve assisted driving.
Why did Tesla cancel millimeter-wave radar? Musk did not explain in detail why.
However, there is some analysis in the industry. Some people believe that millimeter-wave radar has many obvious shortcomings, in the automatic driving will eventually be eliminated, for example, Tesla has previously had two accidents that collided with white trucks in the assisted driving state, and the final survey results showed that the vehicle identified the white truck as the sky and clouds, so it did not make a avoidance or deceleration response, and the cause of the accident has a certain relationship with the failure of millimeter-wave radar to judge the height of the obstacle in front of it from the ground. As Tesla's pure visual algorithms identified by cameras become more and more mature, millimeter-wave radar has become less meaningful to Tesla, and the announcement of the cancellation of millimeter-wave radar shows Tesla's confidence in its own pure visual algorithms. At the same time, the elimination of millimeter-wave radar is conducive to controlling costs, because mature millimeter-wave radar, although not particularly expensive, is also at the thousand-dollar level.
Although Tesla took the lead in canceling millimeter-wave radar, millimeter-wave radar is still valued in the field of automatic driving, because it also has its own advantages, such as millimeter-wave radar can work normally in harsh weather such as rain, fog, wind and sand, and is completely unaffected by environmental factors such as sunshine and alternating light and dark, the former has a great impact on lidar, while the latter will seriously interfere with the normal work of the camera. In addition, millimeter-wave radar is the only perception hardware in cars that can simultaneously detect the distance and speed of target objects, so it still has a place in autonomous driving perception hardware.