Automotive

"Bilateral" adaptive cruise control could help reduce traffic jams

An algorithm developed by an MIT professor could be applied to a modified Adaptive Cruise Control (ACC) system to help eliminate such traffic jams (Photo: Shutterstock)
An algorithm developed by an MIT professor could be applied to a modified Adaptive Cruise Control (ACC) system to help eliminate such traffic jams (Photo: Shutterstock)

In 2007, mathematicians from the University of Exeter showed that the freeway traffic jams that appear to occur for no reason are actually the result of a "backward traveling wave" initiated when a driver slows below a critical speed. This sets off a chain reaction that ultimately results in traffic further down the line coming to a complete standstill. An MIT professor has now developed an algorithm that could be applied to a modified Adaptive Cruise Control (ACC) system to help eliminate such traffic jams.

Last year, Honda announced plans to conduct public-road testing of technology that detects whether a person's driving style is likely to create traffic jams and encourages them to adopt a driving style that would avoid this. At the time, Honda said it would be possible to further improve this system by connecting it to cloud servers that would allow a vehicle's ACC system to automatically sync with the driving patterns of vehicles further up the road.

Berthold Horn, a professor in MIT’s Department of Electrical Engineering and Computer Science, has come up with a somewhat similar approach that would also rely on a vehicle's ACC system, but without the need for the system to connect to the cloud. However, it would require current ACC systems, which only monitor the speed and distance of vehicles in front, to be modified to also take into account the speed and distance of the vehicle traveling directly behind.

Horn describes this system as "bilateral control," as it looks both forward and backward at the same time. By gathering this information, the ACC system is able to keep the car at roughly the midpoint between the vehicle in front and the vehicle behind. In this way, the system is able to avoid slamming on the brakes too hard if the car in front brakes, thereby avoiding causing a large disruption to the car behind that is then amplified as it is passed onto vehicles following behind.

Being a computer scientist, Horn was able to construct a computer simulation to put his algorithm to the test – a sample of which can be seen in the animated GIF below that even includes brake lights and has the bilateral-control algorithm switching on at the one-minute mark.

Although the simulation appeared to verify his hunch, he sought further mathematical proof that came through modeling the bilateral control using the "damped-wave equation," which describes how waves propagating through a heavy fluid die out over distance.

Horn found that his algorithm worked very efficiently when taking into account various values for a range of variables, such as driver reaction times, their desired speeds and how rapidly they accelerate to reach those speeds when a gap opens up in front of them. The only thing that changes as these variables change is the time it takes for the algorithm to smooth out the disruptions.

The major problem facing the implementation of the algorithm is the technology required. Currently, ACC systems are generally only available as an option on high-end vehicles as they rely on relatively expensive sensors such as radar or laser rangefinders. Additionally, these only monitor the speed and distance of the vehicle in front, so Horn's system would require a second system to be installed to monitor vehicles traveling behind the car.

However, Horn does suggest that alternative cheaper technologies, such as digital cameras, could be employed to bring his algorithm to the roads. But these have their own drawbacks.

“There are several techniques,” Horn says. “One is using binocular stereo, where you have two cameras, and that allows you to get distance as well as relative velocity. The disadvantage of that is, well, two cameras, plus alignment. If they ever get out of alignment, you have to recalibrate them.”

Horn's system would also only be effective if a large percentage of vehicles on the road were using it, meaning it probably won't be helping to cut traffic jams anytime soon.

Horn presented his algorithm at the IEEE Conference on Intelligent Transport Systems held earlier this month.

Sources: MIT, University of Exeter

  • Facebook
  • Twitter
  • Flipboard
  • LinkedIn
14 comments
Jeronimo
Why not use a traffic sensor system (camera's/pressure sensors) plus highway lights/signs to monitor traffic and visibly relay the 'suggested driving speed' to cars further back in traffic? (i.e. take out the ACC automation and rely on people's acceptance of a more efficient system).
Denis Klanac
Eliminate the day dreamers and problem solved.
Suzanne Bradley
One day you won't have to drive. One day the car (if it's still known as such) will take you to your destination after you tell it where that is. You won't drive, it will be GPS guided all the way. You will just sit back and relax.
One day...
Rocky Stefano
Much of the bunching up in section or what we in Toronto call the "accordion" effect is caused by ineffective civil engineering planning of the traffic light systems. You will have a very busy avenue that is pockmarked with slower side streets. For inexplicable reasons they will allow these side streets to dominate the ingress/egress traffic light times vs the heavily trafficked avenue. So instead of having a busy avenue with moving traffic, you have small side streets with traffic lights lasting 3 minutes dictating the flow of traffic on a busier street. Stupid
Alien
This is valuable work and eventually will be useful. I love the animation which does clearly prove the concept.
Sadly, until absolutely every vehicle on the road is fitted with ACC and the necessary modification, there would seem little chance of it being implementable.
I'd love to think this could happen sooner but it seems more likely to be twenty years away from utilisation on the road network.
Perhaps there are other potential applications ...perhaps the concept could be adapted for air traffic control to avoid 'stacking' while waiting for landing slots.
Synchro
Rocky - that's nothing, you should try France. They do something similar, but without the traffic lights called "Priorité a droite", so small side roads have priority over the main roads they join. Except they repealed this law. Or maybe they didn't. Or maybe only in some places. Or on some days. Or when you see this sign (which might vary), or where you don't see this sign. Net result: total confusion; nobody has a clue who has right of way, and road deaths in France are double the UK, despite having less than half the traffic density. And how does the French presidency deal with this insanity? By giving an amnesty to all outstanding speeding offences on winning an election. Maybe they just want fewer voters?
Roger Garrett
I would wonder how this behaves when you include the possibility of cars changing lanes when they see a tie-up up ahead.
fancyflyer
Much of the slowdown in highway traffic is due to turbulence caused by drivers who repeatedly pass vehicles moving at posted speed and return to the same lane. They advance to the next regular-speed vehicle and repeat the maneuver until there is no more room in the next lane to pass again. Those vehicles must then slow down to posted speed with the rest of the traffic ahead, creating a backup (bunching up) of cars where there used to be normal spacing. Vehicles traveling behind must now slow down, as the vehicles that were behind them are now crowded in front, using up all the safe gap space that used to be there. It takes only one jumper on a busy highway to have an effect; with many of them you can get a real slowdown that affects everyone, jumpers alike. When all traffic moves together at the same speed everyone gets down the road quicker.
Gregg Eshelman
An aftermarket adaptive cruise control addon would be very nice.
One more thing that causes these traffic snarls is when people do "pulse and glide". Slow acceleration to just above or right at the speed limit then take their foot off the accelerator pedal until they slow down to a certain speed under the limit, then they slowly accelerate again. There's even little microcontroller based computers that can do this automatically.
But some of the real snobby ones doing that have no care for anyone else on the road and will P&G on a busy road instead of just choosing a speed and sticking to it. They'd rather be a jerk and make it impossible for anyone behind them to use their cruise control.
Josh Kahan
no offense to the University of Exeter folks, but we knew about the backward travelling waves (2nd order differential wave equations) at MIT back in 1987.
Still really cool though