Hello, I constantly communicate with Tesla technicians and when it comes to working with NOA I can give you some recommendations to avoid such situations in the future
NOA and AP are both Level 2 Assist-Autonomy which expects you to follow the following basic rules
- A. keep your hands on the wheel all the time (The actions of the user override the actions of the machine 100% of the time)
- B. you have read the entire manual and understand the behavior and limitations of the autonomous system. (i.e. what objects it recognizes and what it ignores, how it reacts to which lanes, signs, types of pavements, traffic patterns)
- C. What sensors do you have their range and limitations, where they are placed).
If you haven't done A,B,C please please take the time to look those up in the manual or online that's CRUCIAL knowledge
### NOA Feedback Program
NOA is weight-based human-feedback AI autonomy. when you find a constant wrong pattern, please report it as defect by using voice command: Report. or holding the vehicle icon for 3 seconds. (his sets of a checkpoint timestamp so when you go in for service you can mentioned you reported it and they will find issues easily).
#### Configuration for reliability
- Confirmation on lane change: YES, alert = Sound
- Traffic avoidance: MILD
- Speed (Don't driver over 5 mph above speed limit)
#### Human Behaviours:
- Always merge in and out of High Way with full manual
- Reach to the middle lane (if applicable) and then Turn on Autonomy while foot is on Accel (Take it away no sooner than 1 second after switching NOA/AP on).
##### Teaching AP2.0 Hive Mind:
- Steer tech looks primarily at lanes, make sure you hold and direct the steering wheel at the path you intend. as AP wrestles with you it learns the human desire and over time adjusts, it's okay if you break out of AP and then set it back, this is just you telling the machine that it did so much wrong you took over, data like this helps it understand that the vector path it decided is wrong and it should try another one, it also tells Tesla to review your actions and the machine actions. data is not tied back to the driver / owner.
#### Difference between Geofencing autonomy and independent Autonomy
There are many Autonomous Driving machines in at least 3 other nations than the US, what separates Tesla apart is it skipping 10 years worth of evolution and jumping into AI based independent Autonomy which was not expected until after our life time, (we're skipping Lidar driving, Geofencing, Cloud based 5G remote navigation, and so much more).
Tesla may work something with 5G in the future for report support during Project Robotaxi) but that's in Autonomy Level 5 2030.
Currently once Level 3 is out NOA will be nearing Level 4 confidence and you can confidently be distracted because machine warning will be based of projection and as early as 30 minutes away, it will also have better corrective navigation maneuvers than the current (go back to your lane) solution.
#### Updates preference (New features sooner come at grater bug risks)
It goes without saying to engineers but to non-tech people here is the short version
Stick to normal update cycle rather than advanced keeps you in the stable non buggy experience.
Advanced is basically stable-beta release Where even if most rough kinks are ironed out there are defects to be expected some can be really troublesome like software lag, connectivity, or infotainment crash, but you'll never have driving interruptions ever. recently Tesla has been working better with Stable-beta than last year. driving core software is untouched in minor releases.
If you want to enjoy your experience avoid testing new stuff. I know it sounds mega lame. but I enjoy stability over cutting-edge especially with kids around yikes.
Beta cycle content is deployed within 30 day window unless a major hot fix is required.
Disclaimer: I don't work for Tesla. some of my intel especially around numbers can be off.