Study suggests AI cruise control could kill traffic jams by cutting out the 'intuition' factor
- Reference: 1669226593
- News link: https://www.theregister.co.uk/2022/11/23/ai_cruise_control_commutes/
- Source link:
That's the initial suggestion from a five-day trial that took place in Nashville last week that saw researchers from the [1]CIRCLES Consortium deploy 100 human-piloted vehicles with AI-powered cruise control systems in morning highway traffic on I-24.
The CIRCLES Consortium's goal with the experiment, and its overall mission, is to use deep reinforcement learning to improve traffic flow and reduce fuel consumption caused by what it calls "phantom jams" or traffic slowdowns that have no apparent cause other than how humans tend to drive.
[2]
"Driving is very intuitive. If there's a gap in front of you, you accelerate. If someone brakes, you slow down. But it turns out that this very normal reaction can lead to stop-and-go traffic and energy inefficiency," said CIRCLES Consortium Principal Investigator and UC Berkeley professor Alexandre Bayen.
[3]
[4]
The vehicles used in the experiment were equipped with AI algorithms the CIRCLES team calls "speed planners" and "controllers." Both of them use information about overall traffic conditions and immediate surroundings to determine the best speed for the vehicle to take to improve traffic flow.
"Our preliminary results suggest that, even with a small proportion of these vehicles on the road, we can effectively change the overall behavior of traffic," Bayen said.
A little AI traffic can go a long way
Because of the sheer volume of data collected over the course of the experiment, Bayen believes it may take months to get a more precise result. Still, the initial findings appear to support a [5]smaller experiment performed by UC Berkeley researchers in 2016.
In that test six years ago, 20 cars on a closed circular track were driven by human drivers, and researchers noted the appearance of similar patterns to those on highways and busy roads. Adding a single AI-equipped vehicle to the test reduced congestion and led to a 40% reduction in fuel use.
[6]Waymo turns its driverless cars into roving weather stations
[7]Tesla recalls 40k cars over patch that broke power steering
[8]Self-driving truck startup TuSimple ousts CEO over ties to Chinese rival
[9]US orders safety recall of Tesla Cyberquad-for-kids ATV
Last week's test added some new technology that made it into what Bayen described as a game changer: the vehicles coordinated actions between themselves, allowing them to react to conditions further ahead and coordinate their traffic influence network accordingly.
The AI-powered vehicles also incorporate information about local traffic conditions from the I-24 MOTION corridor where the test was performed, which is a section of highway equipped with 300 4K sensors for traffic monitoring.
[10]
Armed with data from both I-24 and vehicle sensors, the CIRCLES team plans to update their computer simulations to help them better reflect the real world. As part of that, they want their onboard AI to learn to not only better control traffic, but also learn to be a socially acceptable driver on public roadways.
"We want to train our vehicles to drive in a specific way that is not human-like, but also not completely socially unacceptable. A big focus for us during the test week was to make daily tweaks to our controllers based on feedback from our drivers," said CIRCLES chief engineer and co-principal investigator Jonathan Lee.
Eventually the team wants to see similar technology deployed in many, "if not all, vehicles," Lee said. The CIRCLES team is working on scaling its technology, but we weren't able to determine if, or when, such technology could make it to a highway near you. ®
Get our [11]Tech Resources
[1] https://circles-consortium.github.io/index.html
[2] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=2&c=2Y36mCyJoWJTyytjKXAUZwgAAAAU&t=ct%3Dns%26unitnum%3D2%26raptor%3Dcondor%26pos%3Dtop%26test%3D0
[3] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44Y36mCyJoWJTyytjKXAUZwgAAAAU&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0
[4] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=3&c=33Y36mCyJoWJTyytjKXAUZwgAAAAU&t=ct%3Dns%26unitnum%3D3%26raptor%3Deagle%26pos%3Dmid%26test%3D0
[5] https://its.berkeley.edu/node/12828
[6] https://www.theregister.com/2022/11/15/waymo_weather/
[7] https://www.theregister.com/2022/11/08/40k_teslas_recalled_due_to/
[8] https://www.theregister.com/2022/11/01/tusimple_selfdriving_ceo_fired/
[9] https://www.theregister.com/2022/10/27/safety_recall_issued_for_teslas/
[10] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44Y36mCyJoWJTyytjKXAUZwgAAAAU&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0
[11] https://whitepapers.theregister.com/
Duh. It always entertains me when the lane twitching squirrels end up just 5 feet ahead 30 minutes later and 1 mile down the road. Then there are the compulsive lane bombers who attempt to dive into an exit queue at the last millisecond.
You can't drive around compulsive lane bombers, let them drive past you and into the rear of the car in front of you. Is this AI (Adult Idiot) driving?
I think that Google sucks lots of ways but I open Google maps on my phone attached to the dashboard every time I drive and I never see any traffic problems that can't be easily driven around without even breaking the speed limits (LOL).
I once did a little experiment with lane switching. Whenever my lane was about to reach the pinnacle of its speed, I started to search for a gap to switch to the slower/stationary lane. The drivers in the slower lane didn't mind as they were not moving any way. Shortly after switch, the lane started to move and the other one slowed down. And I did indeed progress considerably quicker than the average.
But it's much less stressful to just stick to whichever lane I'm in.
So what are the AI drivers doing differently that reduces congestion?
FTA: "Last week's test added some new technology that made it into what Bayen described as a game changer: the vehicles coordinated actions between themselves, allowing them to react to conditions further ahead and coordinate their traffic influence network accordingly.
The AI-powered vehicles also incorporate information about local traffic conditions from the I-24 MOTION corridor where the test was performed, which is a section of highway equipped with 300 4K sensors for traffic monitoring."
In short, the autonomous vehicles are acting cooperatively and with greater information about traffic conditions, as opposed to humans, who generally act selfishly and with more limited information. Identifying similarities between this situation and other parts of the human condition is an exercise left to the reader.
Drivers aren't paying attention. They brake late and too hard, followed by accelerating late and too fast. It progressively amplifies the slightest disturbance into hard stop and go ripples.
Also, a significant number of drivers don't look up from their cellphones quickly enough to avoid a collision. That blocks a lane for a while.
Stick with the lane the trucks are in works for me
Much less stop/start (it's kind of harder to do in a truck so they don't unless forced to), less stress and pretty sure less fuel use. Every time you brake you convert expensive fuel into heat, so if you can avoid braking it saves fuel. In the end you won't lose much time if any, the truck drivers are as keen to get to where they're going as anyone.
Re: Stick with the lane the trucks are in works for me
Are the speed limits the same for trucks and cars in the USA?
On the right-side of the pond, the large vehicles have a speed limit roughly 10 MPH slower than cars - enforced by an on-board limiter.
So while it works, and you do use less fuel, there are two problems:
1) You get there noticeably later
2) You are regularly terrified, as a truck crawls past going 1 MPH faster than the truck you're following, boxing you in and leaving you absolutely no escape should it randomly decide to squish you.
Re: Stick with the lane the trucks are in works for me
Getting into a trucks slipstream really does improve mileage - it also tends to be slower. However, if you get TOO close then inability to see what is in front of the truck, human reaction time, and a trucks excellent brakes, you could be decapitated.
Sigh ..
The most dangerous manoeuvre on a motorway is changing lanes. The less you can do the safer you - and those around you - are. I prioritise staying in lane over pinging in and out of the LH lane every 5 seconds where I can. That means noticing there are such things as "junctions" and moving appropriately.
add some intelligence
"If there's a gap in front of you, you accelerate. If someone brakes, you slow down."
Pesky humans lol. Much more fuel efficient to do the opposite - as long as you end your data collection before the explosions burn 100% of the available fuel.
Re: add some intelligence
Yeah, that sentence missed out the all-important detail:
Humans are slow to react , so they overcompensate .
When the car in front slows down, you don't notice until you've got a few metres closer. So you have to slow down more than the car in front to maintain a safe separation.
So the car behind you slows down even more. If the traffic is fairly heavy, then eventually someone has to stop because nobody has left a large enough gap to absorb the delta.
When the car in front speeds up, the same happens but in reverse.
Adaptive cruise control already solves this pretty well.
"AI" brings nothing to the table, and in fact is simply dangerous. Knowledge of the average speed of several vehicles in front might help an adaptive cruise control algorithm pick a better target speed, but we're already into rapidly diminishing returns.
Reactive speed limits
These cars are controlling traffic. The same can be done by reactive speed limits, in jurisdictions where drivers obey speed limits (so, not most of the USA). More overhead speed controls, slowing traffic approaching a wave trough, so that cars travel at average speed instead of high speed until the see the next traffic jam.
Did they try putting public transport in too?