My Thoughts on Autopilot Efficiency

[Jedi2155]

After having done a moderate drive with and without autopilot on a long stretch of highway (I've put about 330 miles in the first day I got my car ), I definitely noticed that autopilot is not that efficient compared to my normal driving habits. I noticed about a 5-15 Wh/mile increase using autopilot (don't quote me on this) compared to my personal driving habits based on 1 particularly trait: don't regen if you don't have to.

• Autopilot appears to be overly zealous on maintaining follow distance and/or speed where it regens aggressive in situations it doens't have to either keep safety or maintain speed .
• This results in excessive regen or slowdown / acceleration, when a slower more gradual speed shift would've avoided the regen entirely.

Just my thoughts on the topic because a lot of people seem to assume that autopilot is more efficient than manual driving.

Attached is my P3D+ efficiency for the first day of ownership! (32 hours to be exact). I did about 4-5 launches in that time, multiple high speed accelerations and did a lot of 70-90 mph stretches and one short burst at 100 mph.

Interested in your guys thoughts on the subject. The 569.5 miles is lifetime of the vehicle and the Wh/mi dropped from 360 Wh/mi since I got it lol.

 

[ewoodrick]

Come on, you can't talk about efficiency when doing 70-90 mph stretches.

Tesla doesn't seem to approach driving from an efficiency viewpoint, it seems to be a driving viewpoint, which is probably good. It's an electric car, it's already a lot more efficient.

Efficiency is pretty easy, turn the speed down

 

[Need]

Wow P3D uses a lot more juice than 3D. I have 2500+ miles and about 2000 on autopilot. My lifetime is 241 Wh/mi.

 

[Bays]

Yeah, I do wish the autopilot was a bit smoother. It's pretty aggressive in acceleration, and when it needs to slow down it reacts later than I'd like, and thus needs to slow down harder when it does.

 

[MXWing]

Someday effeciency is going to go too far the other way.

Algorithm detects the red light ahead, lets off accelerator to regenerate and times it so you arrive exactly when the light is green.

Unfortunately you really pissed off the person behind you as your speed dropped from 60 to 35, a quarter mile out.

Unless I’m 10 beers drunk or every window is blacked out, or I drive like my wife, I drive better than AP.

Can’t see your expectations too high on its driving skill for the time being, especially with managing regen.

I bet chill mode increases effeciency too but argggg

 

[chinnam3]

After having done a moderate drive with and without autopilot on a long stretch of highway (I've put about 330 miles in the first day I got my car ), I definitely noticed that autopilot is not that efficient compared to my normal driving habits. I noticed about a 5-15 Wh/mile increase using autopilot (don't quote me on this) compared to my personal driving habits based on 1 particularly trait: don't regen if you don't have to.

• Autopilot appears to be overly zealous on maintaining follow distance and/or speed where it regens aggressive in situations it doens't have to either keep safety or maintain speed .
• This results in excessive regen or slowdown / acceleration, when a slower more gradual speed shift would've avoided the regen entirely.

Just my thoughts on the topic because a lot of people seem to assume that autopilot is more efficient than manual driving.

Attached is my P3D+ efficiency for the first day of ownership! (32 hours to be exact). I did about 4-5 launches in that time, multiple high speed accelerations and did a lot of 70-90 mph stretches and one short burst at 100 mph.

Interested in your guys thoughts on the subject. The 569.5 miles is lifetime of the vehicle and the Wh/mi dropped from 360 Wh/mi since I got it lol.

I have similar experience with mine. Over 2000 miles trip, I experimented with Autopilot vs my ECO mode driving. I was much more efficient than Auto Pilot.

I think one of the main reason is, Auto pilot is not looking far ahead. It seems range on how far it can process cars is limited. That may be the reason it approaches slowed traffic at much faster pace and apply brakes unnecessary. I think 36.2 seems to improve little bit in that area, but still long way to go.

 

[rounderjd]

The solution to the problem the OP describes is changing the acceleration setting from Standard to Chill.

 

[outdoors]

After several multi thousand mile road trips in both modes I would say it is highly traffic volume dependent. Lots of cars you can do better on your own. Wide open spaces with signs like 38 miles till next service I find AP to be right up there with me. Personally I will always let TACC and AP do a little background work to ensure me not slamming in the back of someone at rush hour in the name of efficiency.

 

[Jedi2155]

After several multi thousand mile road trips in both modes I would say it is highly traffic volume dependent. Lots of cars you can do better on your own. Wide open spaces with signs like 38 miles till next service I find AP to be right up there with me. Personally I will always let TACC and AP do a little background work to ensure me not slamming in the back of someone at rush hour in the name of efficiency.

I was debating this as well, and I think I'll probably leave it on in the long run. I've only had the car for 3 days, so still getting use to the systems. I didn't spend $8000 on the self driving systems not to use it, but was just commenting on the idea that it is a little less efficient.

I got a nice efficiency to work this morning....stop and go traffic to work as usual.

174 Wh/mile over 28.7 Miles
It was downhill though, 804 foot elevation change, and given my vehicle (4072 lbs) + my weight (150 lbs), it equated to an extra 1278 Wh of energy. So about 219 Wh/mile is probably closer to the limit of what a P3D+ could do (maybe 200-210) but it probably won't break 200 Wh/mile like Aero equipped tires and/or RWD.

 

[Silicon Desert]

While I will generally agree that AP efficiency is most often not as good as an energy conscience driver, a 5-15 Wh/mile increase using autopilot as you mention is just too small of a percentage difference because there are so many other variables to consider even if you are trying to drive the same exact route under the same conditions. IE, as you know, variables like wind direction and velocity, road conditions, ambient temperature, speed, firmware version, etc.

 

[outdoors]

The newness had me just like you. Now it has become mindless. More fun to drive than look at Wh/mi. To each his/her own.

 

[sperkin]

I find AP is more efficient than me. I'm on AP 98% of the time. AP regens hard so it captures more energy. When I try to ride it slow I get less energy back.

RWD is at 191 lifetime but usually 180.
AWD is now at 200ish and 1st commute was 187.

 

[rob968]

I find AP is more efficient than me. I'm on AP 98% of the time. AP regens hard so it captures more energy. When I try to ride it slow I get less energy back.

RWD is at 191 lifetime but usually 180.
AWD is now at 200ish and 1st commute was 187.

View attachment 339967 View attachment 339968

This is totally off-topic, but do you find the acceleration and/or responsiveness much different between the RWD and AWD? I've read many comparisons on the forum, and some say it's a noticeable difference while others say it's somewhat negligible. What's your experience, owning both?

 

[sperkin]

This is totally off-topic, but do you find the acceleration and/or responsiveness much different between the RWD and AWD? I've read many comparisons on the forum, and some say it's a noticeable difference while others say it's somewhat negligible. What's your experience, owning both?

I'm still taking it easy but I think the AWD is a bit faster from 60mph to 80mph. I only floored it once. Also, the response from dead stop is quicker. A light tap on the pedal will get the car to jump forward. I feel RWD traction control is too tight at a dead stop so it dulls the throttle response causing a slight lag.

I would be interested in seeing Laguna Seca lap times on RWD vs non-P AWD from the same driver.

 

[omgwtfbyobbq]

I only use EAP with chill mode, and even then it's too annoying for me unless I can find someone with their cruise control on to follow. I submitted a feature request to use a distance interval instead of a static value for follow distance and to allow a driver to enable/disable chill mode for EAP and the driver separately. Chill mode with a distance interval (say 4-7 cars or w/e) should go a long way towards making everything smoother and more efficient.

 

[favo]

The solution to the problem the OP describes is changing the acceleration setting from Standard to Chill.

Yeah, I found AP too herky jerky when not on Chill mode. (FW 36.2)

 

[Silver Ghost]

Chill mode does help in my P3D+.
Amazing to see numbers below 200wh/mile.
Even with chill mode I am in the 320’s wh/mile. Without chill it can go to 340+
Speed limit is 75 mph and I drive speed limit mostly (or 5 over)

 

[Jedi2155]

I saw very little difference between chill mode on and off. Its a little smoother in Autopilot for sure, but if you have a good control of your foot, you can achieve the same results with and without chill lol.

 

[Jedi2155]

I've had V9 since Friday and I've noticed a huge improvement in autopilot TACC since then. I'm wondering if the huge difference in performance is related to the idea that they didn't optimize it for the dual motor versions of the Model 3.

Its far smoother and allows for a bigger buffer on acceleration/deceleration compared to the jumpiness of V8 (I only had 36.2). It's still not as good as a human driver especially when you see a car about lane change into your lane, but I'd say its about as good as it can get with the current camera system and AI without making huge leaps and bounds on reading adjacent vehicle behavior.

 

[Saghost]

In my experience, following distance setting makes a big difference in Autopilot's aggressiveness and therefore efficiency. If you want Autopilot to be more efficient, you have to give it longer following distances (and accept more idiots cutting in.)