February 2018 – Tesla is the beta-testing phase of its updated Autopilot 2.0 hardware and software, as of late February 2018, and a full roll-out might be coming soon (as part of the version 9 release). The only completely new feature is the ability of Autopilot 2.0 to detect and show vehicles driving in lanes adjacent to the lanes in which the Tesla vehicle is driving. For example, if a Tesla is in the right lane of a three-lane road, the system detects anything on the shoulder and the left (passing) lane. If the vehicle is in the middle lane, it detects traffic in the slow (right) lane and the second passing lane to the left. These vehicles appear the instrument cluster so drivers have full knowledge of the situation without having to turn their heads to look.
Other features of the Tesla Autopilot update are improved over the previous version of Autopilot. The neural net that collects data for Tesla shows a marked improvement in lane detection alongside a reduction in lane “ping ponging” with the Autosteer function. Drivers who are part of the beta test report a driving experience at least as good as or better than the older version of Autopilot. Tesla still has to collect more data from its neural net before making fleet-wide changes. Once the new update happens, it should correct any of the problems that owners reported from Autopilot 1, such as improved highway driving.
With all of the delays plaguing the automaker’s self-driving technology since 2016, this Tesla Autopilot update is somewhat welcome news for frustrated owners waiting for more autonomy. Although this update doesn’t get Tesla to level 3 autonomy, beta testing with trained drivers is a step in the right direction. Once the beta test is complete and the neural net has enough data, the software and hardware updates go out to every Tesla Model 3, S and X currently on the road. When more drivers use Autopilot 2.0, the neural net collects even more information to make more advancements in the steady march to full self-driving autonomy.