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

Tesla Semi Aims to Put an End to Dangerous Jackknifing

This site may earn commission on affiliate links.
If a truck pulling a trailer skids, the trailer can push the truck from behind until it spins the vehicle around and faces backwards. Caused by equipment failure, improper braking, or adverse road conditions, this precarious situation is known as a jackknife.

The Tesla Semi aims to end this dangerous, decades old problem on roadways. The truck uses independent motors for each of the wheels and can sense distribution of weight across those wheels. This allows the truck to engage the motors or brake as needed to maintain control of the vehicle. The system can also correct an oversteer in the event of a significant weight shift.

semifront34.jpg
All of this happens automatically. It requires no advanced driving skills from the person in the cab. And, maybe that’s the point – the company does not hide its intentions to make the Tesla Semi fully autonomous.

But, while humans are still behind the wheel, Tesla is being thoughtful about creating a safer experience for drivers.

“Jackknifing is impossible,” Musk said at the Tesla Semi unveiling.

 
Last edited by a moderator:
Then Tesla does not understand jackknifing. It is caused by the trailer, not the tractor. It happens on icy roads, windy conditions, or under emergency braking. The trailer simply gets past the tractor, due to length and weight of the trailers. The trailer only has eight wheels near the back and a higher COG. The tractor has 10 evenly spaced at all four corners. The trailer cannot stop as fast as the tractor so any evasive maneuver can swing the trailer around the tractor.
 
Last edited:
I provided some physical explanation on how this would work in my post of last May on this topic:

Tesla Semi

There is actually nothing useful you can change to the trailer to prevent jackknifing. So you might as well leave it standard and let the semi handle it.

This is not true. Adding ABS to trailer axles (which is simple and becoming very common) pretty much eliminates jackknifing in most conditions, including low mu and split mu situations.

Performing actions on the tractor via the motors is not sufficient (or necessarily even useful) for stopping a jackknife. It may in fact be counterproductive, since one must also consider tractor stability. If you're pulling two trailers, you also have to consider the connecting dolly, which itself can jackknife independently of the trailers. I have no idea how doing anything on the tractor would address dolly or second trailer jackknifing, much less the front trailer.

Modern trailers have ABS. Modern tractors have ABS and ESB. ESB is mandated on tractors as of 2017. ESB *can* intervene on the tractor to reduce speed before the truck approaches a dynamic region where jackknifing or rollover might occur. This would be no different on the Tesla Semi, but wouldn't be unique to the Semi.

The Tesla tractor has one advantage in the stability realm, that the COG is likely to be much lower than a traditional tractor. A traditional tractor will tend to roll in a stability loss event, while a car tends to yaw. Perhaps the Tesla semi will behave more like a car and yaw. Not that this has anything to do with the jackknifing question.

If the rear wheels could steer, thus shifting the tail of the tractor laterally, then *maybe* I could see a way to "get in front of" a jackknifing trailer. But I don't see how just redirecting torque or braking does it. Trailers jackknife when the trailer tires start sliding. Stop the sliding and you stop the jackknife. Thus ABS on the trailer and you've gotten in front of the problem.

I'd be happy to proven wrong on this, but it reflects my current understanding of combination vehicle dynamics.
 
  • Like
Reactions: Ulmo
ABS on the trailer certainly helps but it is not as efficient as regen torque vectoring because the braking forces of the tractor are usually much higher than those of the trailer, especially if the trailer CG is high. It is better to focus on the wheels that apply the highest braking forces on the road.

Regen vectoring (or negative torque vectoring) on the tractor can be very powerful in stabilising it, even with a passive trailer.

Maybe my explanation in the link was not very clear. Let me try differently (I do airplane dynamics not trucks so I'll try to be not too confusing):

Let's suppose a jackknifing event is beginning with the trailer nudging the tractor aft wheels to the left during while braking on a slippery road. The tractor body then starts to rotate more and more to the right if nothing is done, and the more it rotates the worse it gets - jackknifing is a rapidly divergent event. So before things get worse and the jackknifing angle is still small, the computer applies regen braking primarily to the left wheels. Since the trailer now pushes the tractor body on the right side of the tractor braking force vector, this rotates the tractor body back to the left and away from jackknifing. With a proper set of gyros and accelerometers, and proper system reaction time, you can set up braking torque vectoring so that the tractor is always applying left side or right side braking forces fast enough to provide appropriate stability during braking on slippery roads or in small turns.

In short, the computer will continuously balance left/right side regen so that the braking vector of the tractor remains always centered on the decelerating force vector of the trailer, and the whole system becomes completely stable.

I disagree with EM when he says that jackknifing would be impossible because emergency braking in a tight turn would still lead to it, there is a point where the force vector of the trailer would be outboard of the wheels braking vector and then jackknifing becomes inevitable.

But in most situations he is right, the tractor-trailer system will now be as magically stable as a Segway.
 
  • Helpful
  • Informative
Reactions: Pkmmte and TaoJones
ABS on the trailer certainly helps but it is not as efficient as regen torque vectoring because the braking forces of the tractor are usually much higher than those of the trailer, especially if the trailer CG is high. It is better to focus on the wheels that apply the highest braking forces on the road.

Regen vectoring (or negative torque vectoring) on the tractor can be very powerful in stabilising it, even with a passive trailer.

Maybe my explanation in the link was not very clear. Let me try differently (I do airplane dynamics not trucks so I'll try to be not too confusing):

Let's suppose a jackknifing event is beginning with the trailer nudging the tractor aft wheels to the left during while braking on a slippery road. The tractor body then starts to rotate more and more to the right if nothing is done, and the more it rotates the worse it gets - jackknifing is a rapidly divergent event. So before things get worse and the jackknifing angle is still small, the computer applies regen braking primarily to the left wheels. Since the trailer now pushes the tractor body on the right side of the tractor braking force vector, this rotates the tractor body back to the left and away from jackknifing. With a proper set of gyros and accelerometers, and proper system reaction time, you can set up braking torque vectoring so that the tractor is always applying left side or right side braking forces fast enough to provide appropriate stability during braking on slippery roads or in small turns.

In short, the computer will continuously balance left/right side regen so that the braking vector of the tractor remains always centered on the decelerating force vector of the trailer, and the whole system becomes completely stable.

I disagree with EM when he says that jackknifing would be impossible because emergency braking in a tight turn would still lead to it, there is a point where the force vector of the trailer would be outboard of the wheels braking vector and then jackknifing becomes inevitable.

But in most situations he is right, the tractor-trailer system will now be as magically stable as a Segway.

Actually, there are two sorts of things that happen. Jackknifing as you describe, where the tractor is forced off its path from a trailer shove, can be solved at the tractor with ABS. In this case, Elon hasn’t invented anything new, except that he can replace an off the shelf ABS unit with his motor control. That’s a nice benefit of having motors at the wheel, and this contradicts what I’ve been saying here and in other posts, because I’ve been talking about the other case: trailer slew coming from the trailer itself becoming unstable and sliding sideways. The end result is essentially the same dangerous loss of stability, but is solved in a different way. The tractor can’t stop this, but trailer ABS can.

So fundamentally we agree on jackknifing as you describe it, but a tractor based system can’t stop the loss of trailer control that come from instability that starts from the trailer itself. And in the case you describe, as I’ve said, this is basically nothing new. Tractors have this today, they just do it with an ABS unit instead of per-wheel motor control.
 
  • Like
Reactions: Etna
Actually, there are two sorts of things that happen. Jackknifing as you describe, where the tractor is forced off its path from a trailer shove, can be solved at the tractor with ABS. In this case, Elon hasn’t invented anything new, except that he can replace an off the shelf ABS unit with his motor control. That’s a nice benefit of having motors at the wheel, and this contradicts what I’ve been saying here and in other posts, because I’ve been talking about the other case: trailer slew coming from the trailer itself becoming unstable and sliding sideways. The end result is essentially the same dangerous loss of stability, but is solved in a different way. The tractor can’t stop this, but trailer ABS can.

So fundamentally we agree on jackknifing as you describe it, but a tractor based system can’t stop the loss of trailer control that come from instability that starts from the trailer itself. And in the case you describe, as I’ve said, this is basically nothing new. Tractors have this today, they just do it with an ABS unit instead of per-wheel motor control.
On standard setups are the trailer brakes controlled separately from the tractor brakes? Can regen instead of braking make up for a lack of trailer ABS?
 
Thanks for the explanations above.

Just to ensure I've got this right, does that mean that in this case ---> [ = ] - [ 1 ] - [ 2 ] - [ 3 ]

which is my rudimentary attempt to depict a tractor with 3 trailers (such as what one sees in, say, Utah, wobbling down the highway with occasional high crosswinds),

is it correct that jackknifing starts from the 3rd trailer and would be stopped if the trailer braking included ABS (independent of the tractor)?
 
Thanks for the explanations above.

Just to ensure I've got this right, does that mean that in this case ---> [ = ] - [ 1 ] - [ 2 ] - [ 3 ]

which is my rudimentary attempt to depict a tractor with 3 trailers (such as what one sees in, say, Utah, wobbling down the highway with occasional high crosswinds),

is it correct that jackknifing starts from the 3rd trailer and would be stopped if the trailer braking included ABS (independent of the tractor)?

It can start with any of the trailers or the connecting dollies, actually. In your example, there are 5 towed entities, 3 trailers and two dollies. I’ve been in doubles with ABS on the tractor and trailers but disabled on the dolly connecting them, and the dolly jackknifed by itself (on the track). Weird result, and not as bad as the trailer swinging, but it demonstrates just how complicated a problem it is to solve, and having ABS everywhere is the best defense against all forms of jackknifing/trailer slew.
 
  • Helpful
  • Like
Reactions: Ulmo and TaoJones
[...]this is basically nothing new. Tractors have this today, they just do it with an ABS unit instead of per-wheel motor control.

An ABS unit does not do torque vectoring, which provides the ability to displace the braking force vector laterally - a little bit like the ability to hold a stick vertically on your finger.

While agree that ABS greatly reduces jackknife, what Elon was talking about and I was attempting to describe goes beyond ABS in terms of stabilization capabilities.
 
ABS on the trailer certainly helps but it is not as efficient as regen torque vectoring because the braking forces of the tractor are usually much higher than those of the trailer, especially if the trailer CG is high. It is better to focus on the wheels that apply the highest braking forces on the road.

Regen vectoring (or negative torque vectoring) on the tractor can be very powerful in stabilising it, even with a passive trailer.

Maybe my explanation in the link was not very clear. Let me try differently (I do airplane dynamics not trucks so I'll try to be not too confusing):

Let's suppose a jackknifing event is beginning with the trailer nudging the tractor aft wheels to the left during while braking on a slippery road. The tractor body then starts to rotate more and more to the right if nothing is done, and the more it rotates the worse it gets - jackknifing is a rapidly divergent event. So before things get worse and the jackknifing angle is still small, the computer applies regen braking primarily to the left wheels. Since the trailer now pushes the tractor body on the right side of the tractor braking force vector, this rotates the tractor body back to the left and away from jackknifing. With a proper set of gyros and accelerometers, and proper system reaction time, you can set up braking torque vectoring so that the tractor is always applying left side or right side braking forces fast enough to provide appropriate stability during braking on slippery roads or in small turns.

In short, the computer will continuously balance left/right side regen so that the braking vector of the tractor remains always centered on the decelerating force vector of the trailer, and the whole system becomes completely stable.

I disagree with EM when he says that jackknifing would be impossible because emergency braking in a tight turn would still lead to it, there is a point where the force vector of the trailer would be outboard of the wheels braking vector and then jackknifing becomes inevitable.

But in most situations he is right, the tractor-trailer system will now be as magically stable as a Segway.
I agree that huge improvement can be made through many methods such as you described, and also agree that it is not 100% eliminated.
 
I believe BAMF tractor can have negative and positive torques simultaneously.
If BAMF is used with modern trailer (non-Tesla made) and prime mover can measure
trailer angle (two ultrasonic sensors on BAMF facing trailer for example)
it should be possible to totally eliminate jackknifing. But that doesn't mean vehicle
will not stay on road after loss of traction due to whatever reason (ice on road).
AFAIK main reason it happens is because driver did not act fast enough.
With computers, it should be possible to react in time no matter driver's behavior.
It's like ESP on Teslas. Can't turn it off. Hard to spin out whatever you do.

Visualized:
hi_jack.png
 
I believe BAMF tractor can have negative and positive torques simultaneously.
If BAMF is used with modern trailer (non-Tesla made) and prime mover can measure
trailer angle (two ultrasonic sensors on BAMF facing trailer for example)
it should be possible to totally eliminate jackknifing. But that doesn't mean vehicle
will not stay on road after loss of traction due to whatever reason (ice on road).
AFAIK main reason it happens is because driver did not act fast enough.
With computers, it should be possible to react in time no matter driver's behavior.
It's like ESP on Teslas. Can't turn it off. Hard to spin out whatever you do.

Visualized:
hi_jack.png

spit balling:
I think part of the issue is the trailer CG force vector. If the CG inertia vector moves outside the king pin, there is no way to get in front of it again. You may be able to get the tractor to point the same direction as the trailer is pointing, but that doesn't change where it is going. (rotating around pin). The tractor can get pointed in the direction of the trailer vector, that will keep things from getting worse as long as the tractor does not need to slow faster than the trailer can (truck drifting!).
 
Could trailers designed by Tesla for the Semi have additional regen only motors? (in addition to ABS brakes?) That would help with both the jackknifing issue and also add more regeneration to the battery on downhill grades, further reducing overall brake wear?

Obviously normal trailers would not have this.
 
Although jackknifing does happen, it's much more prevalent in the movies ;).

Elon's technology may be able to reduce the chances of a jackknife event happening but nobody should claim they can eliminate it, there are far too many variables that could contribute to the incident. Driver turning the wrong way and all drive wheels skidding sideways will contribute, acceleration may pull it out.

I've never seen or heard of wind causing a jackknife, it just blows the whole unit over. Intentionally jackknifing a truck would actually prevent it from blowing over.

ABS on air brakes are way to slow to be very effective. Good driving skills will always out pull air brakes and traction control, ABS on trailers does save tires to some degree and may help during a jackknife event. There was a rumor engineers were working on electrically actuated brakes for trucks and trailers but I've seen nothing yet.
 
Could trailers designed by Tesla for the Semi have additional regen only motors? (in addition to ABS brakes?) That would help with both the jackknifing issue and also add more regeneration to the battery on downhill grades, further reducing overall brake wear?

Obviously normal trailers would not have this.

If the trailer wheels had motors I think the overall unit could be more efficient. I think this because
 
Oops, somehow posted incomplete thoughts and there is no edit button...

Anyway wheels that are not driven produce drag and although there is still parasitic drag in driving components I wonder if overall the unit would become more efficient.

Another possible big benefit would be the power unit could be sized considerably lighter and cheaper as the towed units could makeup the power as per their designed capacities. An example would be multi trailer units vs single trailer or as trailers are added performance would be affected very little as required energy would be increased with the load. It would make regenerative braking more effective as well, and traction would be greatly improved.