Tesla's Autopilot sled driver in a state of "inattention" in 2018 highway crash

The driver of a Tesla Model S who crashed on a fire truck on a California highway last year did not pay attention because of "over-reliability" on Autopilot, the National Transportation Safety Board said Wednesday. In addition, researchers found that another likely cause of the crash was the design of Autopilot, "allowing the driver to disable the driving task."


No one was injured in the crash, but the use of Autopilot by the driver has brought regulator and media to the attention of the incident. There have been a number of reports from Tesla owners who used Autopilot at the time of a crash, as well as a handful of people who died. Tesla has consistently said that drivers who use Autopilot are safer than those who don't.

The crash occurred on January 22, 2018, when a Tesla Model S P85 collided with a Culver City Fire Department truck parked diagonally across the southern heavy-duty lane of the Interstate 405. The fire truck responded to a collision in the northern lane . The driver of the Model S is not identified.

The Tesla was driving in the HOV lane behind another vehicle, but when that vehicle changed lanes to the right, the Tesla accelerated and hit the rear of the fire truck at a registered speed of around 31 km / h. The autopilot was switched on for 29 minutes and four seconds prior to the crash, but the driver's hands were detected on the steering wheel for only 78 seconds. Autopilot issued "multiple" hands-off warnings in the last 13 minutes before the crash.

"For most of the time, the system has not detected the steering torque of the driver (hands on the wheel)," says NTSB.

After the front vehicle changed lanes, the Model S began to accelerate back to its adaptive cruise control speed of 80 mph. The autopilot gave a forward collision warning 0.49 seconds prior to the collision, but "the automatic emergency braking system was not activated," concludes NTSB.

Autopilot is one Level 2 semi-autonomous system that combines adaptive cruise control, lane assistance, self-parking and, recently, the ability to change lanes automatically. It uses a range of sensors, including eight cameras, radar and ultrasonic.


Car safety experts note that adaptive cruise control systems such as Autopilot rely primarily on radar to prevent other vehicles from being hit on the road. Radar is good at detecting moving objects, but not still objects. It is also difficult to detect objects such as a vehicle that crosses the road and does not move in the direction of travel of the car.

In the past, Tesla CEO has blamed Elon Musk for accidents with Autopilot for driver recklessness. "When there is a serious accident, it is almost always, in fact, perhaps always the case that it is an experienced user, and the problem is more a matter of complacency," Musk said last year. This appears to support the NTSB conclusion that & # 39; inattention and over-reliability & # 39; of the driver played in the January 2018 crash.

In a statement, a Tesla spokesperson notes that while Autopilot turns off when a driver repeatedly ignores warnings to stay involved in driving, the automaker continues to roll out updates to make its advanced driver assistance system “smarter, safer and more effective. "

"Since this incident occurred," the spokesperson added, "we have updated our system, including adjusting the time intervals between practical alerts and the circumstances under which they were triggered."