Brain vs computer


#1

can a computer driven car ever equal the rapid decision making and situational awareness of a human, which is necessary when driving?


#2

Response time is far faster than a human’s and situational awareness FOR THOSE THINGS THAT IT’S DESIGNED TO MONITOR is better than human because it’s totally objective, data-based, and operates in millliseconds. The system in the Mercedes S class can even anticipate terrain variations, road curves, and anomalies by using a combination of satellite tracking, database systems loaded with data from the same purveyors that provide the data for you own GPS and MapQuest-type systems, and other high-tech sensor systems similar to those that guided missiles use, which no human can do. They can detect, interpret, and instantly respond to things even in the dark.

And computer systems don’t drink or do drugs.

But the road is far more challenging than can be foreseen by the high-tech systems designers, and I personally believe that a human must ultimately be in charge. I think these new systems will evolve into systems that supplement human operators and provide sensing and response beyond the abilities of humans. They already are. Sensing systems that interpret the data and operate the car’s controls have actually become commonplace with Stability Control, ABS, and in many other areas.

I strongly believe in these systems as supplements. I get spooked when reading about them as the car’s primary operator. I’m too old fashioned to have confidence in that.


#3

The A in front of AI stands for artificial for the reason that it isn’t real intelligence. Computer programs can recognise patterns and learn to adapt, but they can’t improvise when there’s an emergency. I would have plowed into someone who ran a stop sign in my last car. As I slammed on the brakes with the abs clicking, it took half a second to recognise that the tires don’t have the grip to stop. Where technology came in was optimizing grip as I steered to the wrong side of the road. I don’t think we will see software that can make human decisions and violate rules anytime soon.


#4

yes mountain. I won t repeat them but in the nightmare coming thread, there were examples given of situations where “captaincy” was necessary, and as I said there, the human brain is superior in certain situations where rapid decision making was needed and there were numerous variables that had to be taken into account.

granted, computers can sense and calculate certain things better and faster, but there are certain things that the human can do better and faster too


#5

Computer are programmed by people so I’d rather do my own driving, but hook the scanner up sometime and see how fast the computer is reacting to things like fuel management.


#6

They can take off, fly and land planes. They park cars when some drivers can’t . It will be a replacement for the worst of us…and a supplement for those who want lunch.


#7

Agree,but remember the crash of the Airbus.
Had a horrible nightmare this morning and it was vivid(about trying to get away from the Nazis).so much goes on in the subconcious mind,no wonder our Mark 1000s get confused in a traffic situation from time to time,but for the most part a properly programmed computer will be more consistent and can respond a lot quicker,I suppose speed limits should be set by reaction times.(I see the art major thread got a new life).
So my vote is for the computer(Ken Jennings may be smart,but He cant out calculate a computer)-Kevin


#8

Recalculating…recalculating…Recalculating…

Hopefully you will have a good map.


#9

At least in our life time, I feel these things might be limited to on road travel. The traditional off road capable SUV, even with things like “down hill assist etc. ) must still be driven by people who prefer to go the worse possible route instead of the best. It just seems more fun to drive through a mud hole instead of around. The sign " road constriction ahead, detour” is an invite to all of us who like these vehicles. After too many miles of driving through paved roads, you look for curbs and traffic islands to drive over just to get your “off road fix”. Unless that becomes a program option on one of these buggies, I’ll keep an old Tacoma around.


#10

Thankfully the human mind does not require an external power source that’s subject to failure. They are also immune to the effects of EMP (ElectroMagnetic Pulse). Computers are great devices but they can never replace the human mind…not at this day and age anyway.


#11

For the foreseeable future, computers will always react to the situations around the car, and will do it better than we can. But they will not anticipate issues as well as people because they can’t think. We can look ahead in traffic and see situations that a computer can’t. I can see traffic stopped ahead and exit the highway to avoid it, but a computer relies on traffic data that in many cases has not registered the problem yet. This is frequently the case in my rush hour drives. Rush hour would certainly be more relaxing if I let the car do the driving, but it will take longer since it won’t pass cars but follow them when traffic slows.


#12

@jtsanders: We can look ahead in traffic and see situations that a computer can’t.

I don’t own a GPS unit, but the GPS that came with my cell phone routes me around traffic jams more efficiently than anyone could with mere sight and a map. If you wait until you can see the traffic jam to react, you often can’t avoid it, and even if you can get off the highway, you’re likely going to be stuck with a large group of people who did the same thing, jamming up side roads that weren’t designed to handle that much capacity. My GPS, on the other hand, has routed me around traffic jams that I were miles away, far too far to see in my line of sight, even if I was sitting up high in a truck.

If I wanted a human to route me around traffic as well as my GPS does, it would would require a dedicated operator in some far away location or in the passenger seat to have access to an amount of intelligence no driver could absorb while also driving.

I’ll admit my GPS does this more effectively on long trips than it does when I’m commuting, but I think that is because commutes are shorter, with fewer alternatives that would actually shorten drive time. I often take the most scenic route when commuting rather than the fastest route, and currently, my GPS only has three choices: the fastest route, the most direct route (fewest miles), and the most fuel efficient route. Maybe some day there will be an option for the route with the nicest view, or the most relaxing route.

Sometimes, driving around a traffic jam actually takes longer than going through the congested area. When that is the case on a long trip, I want my GPS to direct me through the traffic jam, especially if the jam is only a minor one. That is something you can almost never discern for yourself just by seeing traffic congestion ahead.


#13

I must have the wrong GPS and listen to the wrong traffic stations. Sometimes they are right, but many times they are not aware of the problem because it is too new to show up. I do use XM traffic and local radio traffic broadcasts to adjust my route, but they often miss traffic jams on I-95 between DC and Baltimore. If I had to depend on a computer driven car to negotiate 4 lanes of parking all the way home.


#14

What’ll happen is that a computer-driven car will avoid, say, 90% of the mistakes people make, but make some that people wouldn’t. Will we accept, say, an 80% reduction in accidents, if some of the remaining ones resulted from the computer’s bad ‘choices’?

And don’t underestimate computers’ ability to make decisions. They integration of cameras with the ‘web’ information will let them anticipate lots of stuff.


#15

Also, computer are only as smart as the people who program them. Paris airshow, Airbus does a fly over in landing configuration. Computer expects the plane to land and lowers thrust. Pilots try to increase thrust and being fly by wire nothing happens, computer controls engines. Plane crashes in the woods at the edge of the airport in a big fireball and pilots die.


#16

@jtsanders, it’s very difficult to know if you have the right GPS, because sometimes it takes longer to drive around a traffic jam than it takes to drive through it. In cases like that, I want the GPS to direct me through the traffic jam.


#17

If this was the incident:

The cause was ruled “pilot error” and the pilot and first officer were convicted of involuntary manslaughter.


#18

A very big problem with any software development is testing. The best programming teams can create software to do specific tasks and solve specific problems. Another team needs to create the “what-ifs” that can cause the software to fail. Those 'what-ifs" that are first level failures are tested, then second level failures, this fails AND that fails, what happens? This is the Design Failure Mode Analysis (DFMA or FMA) method to find errors. Humans have the ability to create abstract solutions to problems as they happen, computers can’t handle the abstract, yet.


#19

I noticed that Google was supplying test cars with no pedals and no steering wheel to test driverless cars. Seems to me this is just silly. Why not at least leave the controls there for backup use. So the computer is going to back it out of the garage to wash it or pull up to the right gas pump or parking place. I dunno, I’m not there. And the whole evolution thing of losing something you don’t use bothers me-like brains, legs, arms, etc.


#20

It’s probably impossible to verify with 100% certainty the computer program is doing what it is designed to do. You math whizzes out there, isn’t there a mathematical theorem (Gödel or Turing?) that establishes this as a fact? If so, then there is always a possibility of failure, and so liability becomes an important issue. Who is liable if someone is hurt? Google? The person who wrote the computer program? The owner of the car?