You have to applaud Ford's efforts to make this kind of cutting-edge technology available in cars that the average consumer can afford. I'm assuming that the lane-keeping technology can be applicable not just to keep drowsy drivers more alert, but also to serve as a warning for the countless folks distracted by cell phones, texting, and eating during the course of their rides.
Hear, hear!! I had exactly the same thought reading the article. Here in Atlanta, despite a fairly strict "no cellphone or texting" law, based on my observations, at least 40% of all drivers I see are breaking those laws, plus another 15-20% distracted in some other way, brushing hair, shaving, reaching down to do whatever, etc. The only ones who do NOT cross over lane markings while doing so are stopped in traffic or at traffic signals! (the GA law doesn't even allow use of cell phones, etc. even then, but it's rarely enforced at all) The above applies on city streets and interstates alike. Maybe in this one case it MAY be possible for artificial intelligence to overcome natural stupidity!
A pessimist would say that this is just one more system which could potentially manfunction, thereby prompting future editions of Sherlock Ohms or Made by Monkeys. I'm not sure I like the idea of a system which automatically applies a steering torque to my vehicle. It seems like a simpler solution would simply be to not drive while drunk, drowsy, or distracted.
@Susan: I'm sure you could steer the car yourself, but if the microcontroller is sending a faulty message to the electronic power steering assist because it thinks you're going out of your lane, you will presumably have to overcome whatever torque it is applying. Or it might apply the "rumble strip" vibration to the steering column while you are wide awake. I'm not sure how likely either of these two scenarios is, but they are nuisances I'd rather not have to worry about.
I think you missed the part of the article that strongly implied that each of the 3 steps were user-configurable as to capability; thus, if you didn't want the "auto-steering" mode, you could turn that off while keeping the warning, etc. Or you could just buy a car that would let you suffer the consequences of driving sleepy!
I understand what you mean about not driving drunk, drowsey or distracted. But life intervenes sometimes, especially for parents of young children. During those years, you're always drowsey and distracted. I remember filling a tippy cup with milk while driving. I held the tippy cup between my knees and opened the milk carton while steering. Coulda used this technology then. Give me the Google car that drives itself.
I agree it would be nice if everyone didn't drive distracted but the reality of the data shows that distracted driving, I believe, is the largest cause of accidents. So I think it's great that Ford is looking to introduce something that will save lives.
I am a little leary that this product will be the best thing in the world because often the first solution on the market is the one that has the most problems and everyone learns from. However, you would also expect a safety device like this would require significant testing.
I, too, would love to see the algorythm. To me there seems to be a lot of variables and extenuating circumstances.
Having worked on aproject intended to provide a similar benefit, I can appreciate what Ford is accomplishing. The system will undoubtedly provide a warning, not a correction. My suggestion has been for the signal to be a recorded dog barking, and Ford is welcome to use that idea for free. The challenge is always to develope an algorithm that does not create false alarms. My employer had purchased the rights to some software and hardware that did not really work, except for the one sample that came as the demo. The most difficult part of the development is the code that determines the baseline that the drowsy driver deviates from. The problem with some systems is that they do the calibration right after the trip starts, which is usually before the driver enters the boring part of the trip. We verified that, at least for "big trucks", it was far better to do the calibration about 20 minutes after the trip began.
But the Ford plan of watching for lane departures is not so very new and original, so either they have licensed the technology or found a fundamentally different approach, which the two camera description would imply.
The one remaining problem now is liability, since some fool will fall asleep and claim that "they said it would keep him safe". It will be very interesting to see the instructions and disclaimers associated with their new system.
I'd be very interested to learn how the algorithm was verified, and what the legal liability might be in cases of accidents caused. --- proveably or not -- by false positives or false negatives. Granted, this is a warning, not a hard correct, but it's all the more interesting because that's where this technology is headed.
I'm wondering what sort of verification testing Ford did as far as assessing false positive and negative warnings and what their potential legal liability might be for accidents (proveable or not) caused by same. Relevant because one day this will be a control, rather than advisory, technology.
Living and driving in various snowy, sleety, slushy road conditions where the lines on the road are partially/completely covered, my first thought was how this would work during those conditions. I think my vehicle unexpectedly "assisting" me by applying torque to the wheel in one direction or another while I'm on a snowy/slushy road would be a danger of its own. Even if the lines are only partially covered with reflective white snow, how would it decipher in/out of lane position? This is one of those technologies that I would wait until a number of critical revisions have been tried out successfully before I would be comfortable adopting it myself.
Lauren: Ford engineers spent years developing software algorithms to recognize various road conditions, including rain, snow, and ice. Ford engineer Mike Kane told us, "We've worked with our suppliers to make sure that in different environmental conditions and different situations, we can still see the lane lines." One interesting note: Historically, this software was an offshoot of facial recognition software. Ford and its suppliers took that technology and re-engineered for use on highways.
Thanks for clearing that up, Chuck. Good to know all the enviromental factors were taken into consideration. I agree, it is very interetsing that the software is an offshoot of facial recognition software.
That facial recognition heritage is very interesting. I wonder if they could replace/supplement the key fob with either facial recognition or another biometric (eg., fingerprint) tool to verify the owner of the car before turning it on. Great way to prevent theft. Personally, I would also like a similar control to prevent towing, based on a recent, unpleasant personal experience where I parked in front of an unmarked driveway.
sdoyle's point is that the lines can be totally obscured. Lines are also sometimes missing or wrong. I'm sure Ford worked hard to maximize the capabilities of the system, but there will always be circumstances that it cannot overcome. Your comment from the Ford engineer doesn't adress what the system does if it cannot discern a line, which I be interested in reading.
The algorithm that my employer puyrchased was never released because the project was cancelled because the technology did not work as claimed. So they are safe on that one. But there are quite a few systems around presently, since the trucking companies find this effort very useful. And I can tell you that there are very few things more boring than watching a video recording of a truck driver getting drowsy.
There is a system that monitors eye blinks, and at least one that watches the lines on the road, another that tracks the drivers focus, and then there was the one we had that tracked the drivers head movements. The failure is that some people can fall asleep without ever nodding their heads.
I would think it will takes years to perfect all the different nuances and variables that are necessary to create such an algorithm. As much as I applaud the effort and know eventually that the technology will evolve to a point that it's highly reliable and effective, it's going to take a long while until people feel confident, especially for setting the sensors so it actually initiates a correction, not just an alert. To me, it's like those rear-view cameras designed to help you back up. Have one on my car. Totally don't trust it.
When this idea was first proposed, the general reaction was "nice idea, but even if you could get it to work, the liability issues would keep the big automakers from releasing such a thing". But for a couple of years now, it has been offered as an option on high-end vehicles. Even the Prius at its top trim level has a package that includes
- automatic parking
- lane keeping assistance and
- adaptive cruise control (which in crowded traffic maintains a set distance to the car ahead instead of a set speed).
I didn't take it. (I preferred the sun-roof/solar roof offered at the trim level below, and in April 2011, one also had to worry about availability of configurations that were not actually on the dealer lot.)
Now I can safely text-and-drive!! Other than more complacency while driving, I wonder about the auto-correct function (if enabled). If you're fully alert, but have to suddenly swerve into adjacent lane to miss a rock or a kid, how hard would you have to fight the steering?
@David: You bring up my first thought. I think the first consequence of this technology will be a less attentative driver who will expect the electronics/mechanics to keep them safe. Let's see, we have an automatic transmission, speed control, GPS, and lane control. Drop the screen and lets watch a video on the way.
I think this could be a great feature. A few years ago or more I remember seeing an offroad vehical using a camera to scan the terrain, determine where the dirt road was and control speed, steering, and braking to stay on the roadway. Amazing if they got the computer to fit in a rearview mirror cause there computer was fairly large. Funny.. the steering system wasn't developed for your safety, but for increasing MPG, which equates to a fine/tax for the automaker. Oh, well still an interesting idea. Hope it works out great I always liked the idea of a semiautomated car.
EPS.. didn't replace older hydraulic power steering because of fuel effiency per se.
It's primary advantage: it is a easier / more flexible system to incorporate into new car designs.(easier to route wires than hydraulic lines) This leads to faster design turns (less engineering). The load is now on the alternator - not a hydraulic pump.
And it translates to hybrid/electric tech easier. (no need to support multiple systems in the future).
It is a strategic design move...
Yea, it is easier to incorporate new advanced features like auto parking/and warning systems.
As to warning someone falling a sleep.. good intentions (but I remain a bit skeptical because of "law of unintended consequences" )
Sorta like...making a road "safer" for it's posted speed, only to find people drive faster because of the improvement.
or allowing people driving licenses.. that can't parallel park.
This will really change the face of automobile safety. If this can be linked with GPS and road profiles , this can be really enchanced and can reduce accidents considerably. With the availabilbity of powerful low cost, low power MCUs technology one day will make it even possible for the blind to drive. This is one of the best examples of technlogy improvign quality of life..
I know exactly how to test a drowsy driver detection system, since we did it for almost a year. The fact is that simulations produce simulated data, and only an actual test with real truck drivers in a regular truck will produce valid data, which is what we did. BUt since this blog would not accept my first entry on the topic, it will not be repeated. BUt the conclusion is that watching a professional driver become drowsy enough to make a mistake is both tedious and quite boring. But it does produce valid data.
Here are some interesting metrics from NHTSA, in their Traffic Safety Facts newsletter (pdf download) from March, 2011. It says that "drowsy driving was reportedly involved in 2.2% to 2.6% of total fatal crashes annually during the period 2005 to 2009." In 2009, this equated to 832 deaths. From this I think we can qualitatively see that Ford's lane-keeping technology will have some real-world positive impact. (Though I guess impact is not the best choice of word.)
Great information, Alex. If you add those 832 deaths that you've mentioned here to all of the annual deaths that could be prevented by vehicle-to-vehicle communications (30,000 per year) and electronic stability control (10,000 per year for this feature, which is now mandated by law), it becomes obvious that automakers believe safety will be enhanced by progressively allowing the electronics to take control of steering and braking. Over the next ten years, we'll see an increasing number of vehicles incorporate the three main building blocks of autonomous safety: adaptive cruise control; lanekeeping; and collison avoidance. I've said elsewhere on this site that I have no intention of letting an autonomous vehicle cart me around, but over the next 20-30 years, I may not have a lot of choice.
I should have added that the data picture regarding drowsy driving is even worse if you look beyond the deaths figure I mentioned in the previous comment and add in overall crash data. According to NHTSA, the annual average over the 5-year period of 2005-2009 includes the afforementioned 886 deaths, but also 37,000 injuries per year caused by drowsy driving and another 45,000 involving property damage only.
I simply do not believe the reported number of deaths that will be avoided by means of vehicle stability control systems. There is no given source, and it is clearly somebodie's number pulled out of the air. What we will get when we can no longer purchase a vehicle without these features, is a huge dumbing down of driving skills. It is one thing to offer a system that would protect beginning drivers, it is an entirely different thing to force all drivers to buy a system that will not make the best choice most of the time. That is where I see the problem. As with our present ABS package that has no clue as to the best way to stop on gravel or sand, or, worse yet, on a road covered with wet leaves. So please get big-brother off of my back!
As a worth while alternative, how about making the drivers test a lot more comprehensive, and including a bit of driving with a simulator. Not only would it make our roads safer, but it could reduce the number of drivers as a whole, and reduce roadway crowding.
Once again, safety does not sell, we all know that the only way to get people to buy safety systems is to force them to buy them. If everybody wanted to be so safe, they would all be driving Volvos, however, as you probably have noticed, not everybody is driving them.
I agree, William, safety is not a strong selling point. The safety features that are common on our vehicles now were mandated -- seat belts, air bags. I can't imagine the drowsy driver function will ever become mandated. Still, it could be useful on long trips.
Rob, I did not think that anti-lock brakes would become mandatory either, but now we are stuck with them. And they are hoping to mandate stability controll as well. The problem with the stability control system is that it is predicated on the driver making all of the wrong moves. It will undoubtedly take the loss of many lives to convince the safety people that the control algorithm is not, and never will be, able to make the correct decision much more than half the time. That is commonly called the un-intended results of some mandate that "seemed like a good idea at the time".
But just waking a drowsing driver before they fall too soundly to sleep seems quite unable to cause any serious problems.
I'm glad to see that (so far) they have only added a warning and not tried to take control of the car. Since it does have options built in, I wonder if there is an option to shut it off (even if the feature becomes mandatory). Since it is being based soley on the road condition, I can see some situations where it could be an annoyance.
I do hope that it doesn't get to be another one of those safety features that people rely on when they would have stayed off the road without it.
There are several products that don't require a view of the roadway to monitor driver drowsyness. One of them watches the drivers eyes and monitors blinking, another one monitors the drivers head motions and position. while a third system pays attention to how the driver moves the steering wheel. I would favor a system that kept track of the drivers brain wave patterns, however, the logistics and technology of such a system would be a daunting challenge.
My point is that it is not nesessary to utilize the roadway to detect the changes in a driver becoming drowsy.
In a bid to boost the viability of lithium-based electric car batteries, a team at Lawrence Berkeley National Laboratory has developed a chemistry that could possibly double an EV’s driving range while cutting its battery cost in half.
Using Siemens NX software, a team of engineering students from the University of Michigan built an electric vehicle and raced in the 2013 Bridgestone World Solar Challenge. One of those students blogged for Design News throughout the race.
Robots that walk have come a long way from simple barebones walking machines or pairs of legs without an upper body and head. Much of the research these days focuses on making more humanoid robots. But they are not all created equal.
For industrial control applications, or even a simple assembly line, that machine can go almost 24/7 without a break. But what happens when the task is a little more complex? That’s where the “smart” machine would come in. The smart machine is one that has some simple (or complex in some cases) processing capability to be able to adapt to changing conditions. Such machines are suited for a host of applications, including automotive, aerospace, defense, medical, computers and electronics, telecommunications, consumer goods, and so on. This discussion will examine what’s possible with smart machines, and what tradeoffs need to be made to implement such a solution.