Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

Navigate on Autopilot - ATX Commute Experience

This site may earn commission on affiliate links.
I have a Model 3, and was fortunate enough to get the 42.2 download Friday evening. I did a little driving, but today was the first big test, a 25 mile commute in Austin, Texas, traffic.

I wanted to log my experience here, and give updates overtime as NOAP is further refined from Tesla.

First off, the parameters. I live in Cedar Park, and commute to downtown East Austin every day for work. It's a mix of 183, Texas 45, Mopac, 35, and backroads if traffic is a complete disaster. Its about 25 miles, and takes anywhere from 40-60 minutes depending on the day. Plenty of good test cases with interchanges, exits, etc.

My overall impression of 42.2:

For high speed driving (60 mph+, average to little traffic), its really good already. The UI clearly gives updates that a lane change is coming up, which will help when the car eventually begins to auto lane change. This will give the human several seconds to start to prepare and be ready for the lane change, even before the car attempts it. I think it's a pretty clever design from Tesla, and will most likely need an optional audible ding as well so the driver can be more prepared.

If you are doing a long trip at high speeds, its a great feature _today_.

In rush hour traffic, it's a different story. Currently, the car really struggles to "act" human in situations where aggressive driving is needed. If you are familiar with Austin traffic, a few examples of that would be the 35 North to 183 North exit ramp in PM traffic. That particular interchange is always backed up, and sometimes requires needing to come to a complete step in a main lane in order to merge in into the exit ramp. NOAP just isn't good enough currently to compete with human drivers and "assert" itself in that type of situation. I've played with some of the settings (including Mad Max mode), but currently those only appear to be related speed based lane changes (i.e., don't get stuck behind a slow poke).

In addition, I also found NOAP was not taking the most "optimal route" when possible for different lanes. For example, the far right lane is typically the slowest lane in rush hour, as cars are merging in and out constantly. NOAP wanted to stay in the far right lane for most of commute (southbound and northbound), and definitely cost me several minutes.

In the times that it did want to change lanes in heavy, slow traffic, I found the car again to not be aggressive enough. There were several spots where I had enough room, but the car just wouldn't go, and I needed to take over control. I would say I was about 25% successful with lane changes in heavy traffic.

Lastly, on my southbound commute, I exit Caesar Chavez off 35 South, which is very short and tricky exit ramp. NOAP knew ahead of time that it was not going to be able to execute the exit, and gave me a countdown BEFORE I got to the beginning of the exit transition. It gave an audible tone, and had a message that said something like "Unable to execute maneuver ahead. Please take control". That will be a difficult exit to master, and I look forward to progress here over time.

I also went through a construction zone while on NOAP, and it gave a warning message that said "Construction Zone Detected. Navigate on Autopilot is unavailable" until I exited the constructed zone.

TLDR: Probably counter-intuitive, but NOAP is solid for high-speed travel right now. It's not aggressive enough in slow, heavy commute traffic, and still has a ways to go for optimal route management in a commute environment.