Great article and discussion Junko...how ia air travel industry dealing with these issues? planes are flown on auto-pilot mots of the time...but pilots take over on take-offs and landings (and in critical situations)...perhaps cars can follow that model
Of course no company will be directly responsible - at least you will get paid by their insurance company and I am sure they will have some limits... and I would generally not label these accidents as "Manufacturing Defects", but as Software Bugs.
This is all based on the idea that "people" are better at driving then robotics which I find highly dubious - especially when they will be able to look at reality at 50GHz in 3D and could plan out 2 million variable and probabilities, could individually brake or speed up all tires individually, and trigger air bags individually at different rates - not now, but in the not too distant future.
I don't think much will be able to be changed as I am sure there will be real time data sent securely over the network and any tampering/usage would be logged and monitored - not to say it couldn't happen, but in the future there will be cameras from every car as well as who knows where else..
I read Earth by David Brin 25 years ago before half of all our advances were even here and he scared the crap out of me reg monitoring networks, data and the future.
I believe they will only do this if it becomes a law as "purists" they will try to "be free" - many of them could get much better performance with computerized ignition and such, but many stick with manual. Also, if you look at the amazing recent changes to cars regarding electronics, and the trend to integrate - everything - there just might not be enough market, or legal support to allow for such modifications.
I agree, but all the cars in a road are a bit more tightly packed, but in reality I believe this can help as they will all form a kind of mesh network which can actually give you quite a bit more data to work with to prevent accidents or other issues better then most people will be able to do.
There are so many other benifits such as traffic flow regulation, reduced lane changes and a biggie - monitoring of drivers so if you are drunk (the leading cause of accidents), tired (another large chunk of problems), or just confused or even breaking traffic laws you will be warned at the least and arrested (and monitored) at the most.
Not that many cars now have live interactive traffic services, or at least our phones can help the problem is much less about "the network" and much more about software and politics.
@Duane, I love your story about your grandfather with a glider! What a great piece of history! I can easily imagine how proud you are.
It turns out I am a daughter of an ultimate tinkerer in Japan -- my late father used to build not only his own radio (which was quite common), but also stereo systems, even a crude PC. I used to tag along with my father's Sunday outing to Akahabara in Tokyo.
I agree. How far people go out of their way to hack things often amazes me but that hacker mentality is the backbone of new developments...
I think another important missing component is the notion of driver leadership. In the current situation, a figurative (or collective) leadership exists in human driven driving. In a world of fully autonomous driving, who or what sets that leadership?
The human element "leadership," in driving or just about any repetitious task, gradually gives way to automation. And we don't even question this, most of the time. Like I said previously, existing systems such as ABS and stability control have already taken some of that "leadership" away.
Autopilot in airplanes is similar to cruise control in cars. It is a system that uses few inputs, and requires the constant attention of the human operator. As we get beyond that level of dependency on a human operator, as we have in plenty of cases, the control algorithms have to become more sophisticated.
In that Air France flight, for example, reliance on the human operator proved not very good, right? Humans are easily deceived, in determining attitude, when visual queues are either non-existent, or disagree with our ears.
In a fully automated system, the fact that the air speed sensor values did not match the GPS speed, or the fuel burn rate for level flight, or the onboard gyros, would have been taken into consideration. So, this would be an example of not enough automation, in cases where the human is really hard pressed to make the right decision.
Imagine a sofa falling out of a moving van; how would the fully autonomous vehicle set respond to it?
Not sure why you find this one difficult. In all of these examples, you ask yourself first: how would a human react? A human would hopefully, if not distracted as they too often are, see the sofa with his eyes. An autonomous vehicle would do this far more effectively. Because the falling sofa would be noticed immediately (as opposed to after the idiot gets his eyes up from his iPhone), and the car would have a WAY BETTER idea where to swerve to avoid it. A human rarely has the situational awareness needed for this. Two eyes in front of your face only. The car can have visual spectrum and radar sensors front, left, right, and rear, which are always active, and are more effective than human eyes in a lot of cases (e.g dark, fog, rain, human in drunken stupor).
Junko - My paternal grandfather grew up in a small town in North Dakota in the early 20th century. The family has a photo of him, at the age of 16, with a car and a person-carrying glider. Neither of those are particularly remarkable, except that he and his friends built both from the ground up with parts and junk they scavenged. They'd take turns riding in the glider, towed behind the car, or launched off the hill near town. (I like to think some of that hacker blood runs through my veins).
When I grew up, no one that I knew was building cars from scratch, but a lot of guys would make some serious modifications. They'd put in high performance cam shafts, carburetors, shift kits, headers, and other parts that weren't legal to have on the street.
Even knowing that, I'm thoroughly amazed at what some people have hacked together mechanically and electronically. Between 3D printers, laser cutters, home CNC machines, and all of the tech stuff we write and read about here, I just can't believe that there aren't people, right now, trying to put these systems together.
There are a lot of problems to overcome, but the hacker mentality if a very powerful force.
What are the engineering and design challenges in creating successful IoT devices? These devices are usually small, resource-constrained electronics designed to sense, collect, send, and/or interpret data. Some of the devices need to be smart enough to act upon data in real time, 24/7. Are the design challenges the same as with embedded systems, but with a little developer- and IT-skills added in? What do engineers need to know? Rick Merritt talks with two experts about the tools and best options for designing IoT devices in 2016. Specifically the guests will discuss sensors, security, and lessons from IoT deployments.