Jump to content

Poll & Prediction: Autonomous Car Equipment at 5k by 2019


Mallette

Autonomous Vehicles: Good or Bad  

49 members have voted

  1. 1. Are autonomous vehicles a good witch, or a bad witch?

    • Good
      20
    • Bad
      28


Recommended Posts

8 minutes ago, Don Richard said:

The facts are that companies are working on AV solutions and have made progress

Most certain part of what you said.  The rest is opinion.  My opinion is that the issues are readily solved by large amounts of data and that the data is increasing exponentially as systems are placed into use would lead me to say the times you suggest will doubtless decrease.  From decades of designing, programming, then beta testing software I can state categorically that the timelines are mostly about the amount of beta test data you have to work with.  The work on AV software has magnitudes more data coming in right now than any other software in history and that will only increase with time.  Improvement will follow an exponential curve never seen before as this is the perfect storm for software and hardware development both in the nature of the issues as well as the rewards of success.  Rarely has any development offered both profit to the 1%  and the rest of us as well so perfectly.  In fact, the only comparable thing I can think of is Henry Ford's work from 1905 to 1915.  That was a human paradigm shift.  This is an even greater one.

Dave

Link to comment
Share on other sites

12 hours ago, Mallette said:

I can state categorically that the timelines are mostly about the amount of beta test data you have to work with

 

Which is why GM is leading the pack in AV development right now. There are not just technical challenges to consider, but also legal ramifications. Nothing will be sold until the lawyers give their blessing, and that won't happen until AV technology is bulletproof.

Link to comment
Share on other sites

3 hours ago, Don Richard said:

and that won't happen until AV technology is bulletproof.

Makes one wonder why the airlines fly.  Hardly bulletproof.  And certainly the same for railroads.  The spate of Amtrak accidents hopefully will make the need to automate them clear.  Same logic should apply to cars.  

Dave

Link to comment
Share on other sites

OK - with the successful launch of the Falcon Heavy (congrats to Musk/SpaceX) I can now admit to a Fully Autonomous Vehicle in operation, albeit in orbit around the sun for eternity. The FH carried a two-set Tesla with space suit wearing dummy as payload for the planned orbiting. Now this AV I can support - !

Link to comment
Share on other sites

On 2/2/2018 at 9:27 PM, dwilawyer said:

It's to be able to walk away from the landing. 

 

All of those autopilot settings are entered and adjusted during flight by a human.  

 

They crashed a B-2 right after takeoff because of misreadings in the sensors.

 

The problem is you have to automate for contingencies such as the instruments and gauges are supplying incorrect information.  You also have to automate for loss of power on one side, and also total loss of power.  With a total loss of power it has to pick and find a safe place to land, it has to dump fuel, and about 100 other things.  It can be done, but planes are not flown in an autonomous fashion.  Just the opposite, a Captain is there in case something goes wrong and a First Officer in the event something goes wrong with the Captain.

 

Military drones are not even automated.  When that happens you might begin to see it move towards civilian aviation. 

I don't think you realize just how much automation goes into the "manual controls" to which you speak. Everything is going through a control system because everything is drive-by-wire. Pilots can't get anywhere close to this level of control with truly manual controls. Deciding where to fly is a relatively arbitrary task....most of that real-time flight adjustment is related to picking new flight paths based on new weather information. It's amazing how much less turbulence we experience today than just 10 years ago. It's to the point that it's a rare occurrence to experience turbulence. I remember when it was rare to have an entirely smooth flight. This is due to better weather prediction / monitoring, and better automation on the planes.

 

It's actually possible for the sensors to have more immediate feedback than what the pilot is able to experience. Sometimes sensors are faster than humans - sometimes slower. I grow tired of this thread always assuming all sensors are always faster than humans. I've worked on some of these sensor systems - I can tell you that some of the sensors (especially automotive) are much slower than humans, and it's not an issue of cranking up clock rates or throwing money at it. It's fundamental to the method of sensing. The best way to solve this problem is with infrastructure exterior to the car (built into the roads).

 

As far as system failures and sensor failures, and all of that stuff.....that is precisely why I don't think automation makes sense for a personal owner of a vehicle. The liability and maintenance requirements simply don't make sense. This is why there is a huge maintenance team on aircraft with incredibly regular inspection and repair. Sure, the cost of automation screwing up on a plane is much higher.....but I still stand by the notion that it's a much simpler design problem than the automotive scenarios. When all systems are fully functional, the automated plane will have a lower failure rate. No doubt about it, and the proof is in the air today.

 

There is a reason cars refer to it as "auto-pilot" and not "auto-driver". Planes got to it first - and that's because it's an easier problem to solve.

 

It's also worth noting that planes have a much longer safe hand-off period between "sensor issues, turn off auto-pilot" and "human takeover". In a car, that could be fractions of a second. You often have minutes when in a plane.

Link to comment
Share on other sites

On 2/5/2018 at 6:28 PM, Mallette said:

This is an even greater one [pardigm shift].

I'm not sure I agree with that. The paradigm will be the same for anyone living in the city without a car. Having an uber/taxi take you around is no different than an AV taking you around. Half the time you don't even recognize the existence of the driver - could be a robot for all you know.

 

This type of lifestyle is already experienced on a rather large scale.

 

I think it's very clear from this thread that personal ownership of an AV is not the 90th percentile solution.

Link to comment
Share on other sites

  • Moderators
15 hours ago, DrWho said:

I don't think you realize just how much automation goes into the "manual controls" to which you speak. Everything is going through a control system because everything is drive-by-wire. Pilots can't get anywhere close to this level of control with truly manual controls. Deciding where to fly is a relatively arbitrary task....most of that real-time flight adjustment is related to picking new flight paths based on new weather information. It's amazing how much less turbulence we experience today than just 10 years ago. It's to the point that it's a rare occurrence to experience turbulence. I remember when it was rare to have an entirely smooth flight. This is due to better weather prediction / monitoring, and better automation on the planes.

 

It's actually possible for the sensors to have more immediate feedback than what the pilot is able to experience. Sometimes sensors are faster than humans - sometimes slower. I grow tired of this thread always assuming all sensors are always faster than humans. I've worked on some of these sensor systems - I can tell you that some of the sensors (especially automotive) are much slower than humans, and it's not an issue of cranking up clock rates or throwing money at it. It's fundamental to the method of sensing. The best way to solve this problem is with infrastructure exterior to the car (built into the roads).

 

As far as system failures and sensor failures, and all of that stuff.....that is precisely why I don't think automation makes sense for a personal owner of a vehicle. The liability and maintenance requirements simply don't make sense. This is why there is a huge maintenance team on aircraft with incredibly regular inspection and repair. Sure, the cost of automation screwing up on a plane is much higher.....but I still stand by the notion that it's a much simpler design problem than the automotive scenarios. When all systems are fully functional, the automated plane will have a lower failure rate. No doubt about it, and the proof is in the air today.

 

There is a reason cars refer to it as "auto-pilot" and not "auto-driver". Planes got to it first - and that's because it's an easier problem to solve.

 

It's also worth noting that planes have a much longer safe hand-off period between "sensor issues, turn off auto-pilot" and "human takeover". In a car, that could be fractions of a second. You often have minutes when in a plane.

I was using "autonomous" in the sense that an autonomous airplane has no pilot, either in the cockpit, or on the ground.  

 

Computer generated artificial stability is in a lot of things we don't realize, and it makes for smoother and safer flight.  However, it doesn't put us any closer to getting on a commercial airliner that has no pilot.  The Capt. and FO are there to program, adjust for conditions, and most importantly,  to be there when something designed, built, maintained and supplied by humans, goes wrong.  It it is this last part that is the most difficult to automate, and you have got to build up flight hours in something, to begin to integrate it into a passenger aircraft.  For a strategic bomber,  I think a pilot will always be on board, even if the plane could be fully automated, just from a fail-safe perspective.

Link to comment
Share on other sites

17 hours ago, DrWho said:

Sure, the cost of automation screwing up on a plane is much higher.....but I still stand by the notion that it's a much simpler design problem than the automotive scenarios.

 

The skies are not nearly as crowded as our roads so there is less for the autopilot systems on aircraft to contend with. Automotive vehicles have to stay within the two dimensional roadway, aircraft have more options in 3D space. Yep, aircraft autopilots are a piece of cake in comparison.

 

I just saw where an AV 18 wheeler just completed a coast-to-coast test trip. This is a Level 3 autonomous vehicle that self drives only on highways with the driver taking control in town. The article I read said the truck went off of auto few times, but did not give a number.

Link to comment
Share on other sites

3 hours ago, Don Richard said:

 

The skies are not nearly as crowded as our roads so there is less for the autopilot systems on aircraft to contend with. Automotive vehicles have to stay within the two dimensional roadway, aircraft have more options in 3D space. Yep, aircraft autopilots are a piece of cake in comparison.

 

I just saw where an AV 18 wheeler just completed a coast-to-coast test trip. This is a Level 3 autonomous vehicle that self drives only on highways with the driver taking control in town. The article I read said the truck went off of auto few times, but did not give a number.

 

Six thousand flights per day reduce the possibility of sunburn.

JJK

Link to comment
Share on other sites

  • Moderators
On 2/8/2018 at 2:40 PM, Don Richard said:

 

The skies are not nearly as crowded as our roads so there is less for the autopilot systems on aircraft to contend with. Automotive vehicles have to stay within the two dimensional roadway, aircraft have more options in 3D space. Yep, aircraft autopilots are a piece of cake in comparison.

 

I just saw where an AV 18 wheeler just completed a coast-to-coast test trip. This is a Level 3 autonomous vehicle that self drives only on highways with the driver taking control in town. The article I read said the truck went off of auto few times, but did not give a number.

Autopilot work great if all engines are working.  What autopilots cannot contend with is power loss.  

Link to comment
Share on other sites

But you're okay with an "automobile autopilot" dealing with a flat tire? An engine out on a plane has a much longer safe transition time back to human control than a tire on a car.

 

I'm also not convinced a damaged engine is that hard of a problem to address on a plane, but nobody here is an engineer in the avionics industry so its just speculation. I just know planes don't crash in tenths of seconds after a failure.....it takes a long time to fall out of the sky :)

Link to comment
Share on other sites

10 hours ago, DrWho said:

But you're okay with an "automobile autopilot" dealing with a flat tire?

Certainly more than myself dealing with it.  Given my age, I've had this happen too many times.  But not enough to really learn how to handle it except by instinct.  I'd rather have a computer using data specific to the behavior of the vehicle I am in and based on thousands of incidents.  Looks like, as it were, a no brainer to me.  

 

As to aircraft, as I mentioned, Capt. Haynie was probably one of a few pilots in history to be able to handle Sioux City, and Sullie the same on the Hudson.  Programming their reactions into a flight system would certainly improve the odds over the average pilot when these circumstances arise again.

 

Dave

Link to comment
Share on other sites

Actually, seems to me that whether computer or human controlled, it should be possible for ground control to take charge of an aircraft in the event of hypoxia, hijack, or whatever.  I really don't get this any more than I understand why trains are under human control.  While, for whatever reasons, they still intend to have a human in them, I read this morning that Amtrak will discontinue service on lines where they cannot get positive train control installed by the deadline until they can.  Good move, long overdue.  

 

Dave

Link to comment
Share on other sites

44 minutes ago, Mallette said:

I'd rather have a properly-working computer using data specific to the behavior of the vehicle I am in and based on thousands of incidents.

I added the underlined part to help everyone decipher what you have really been saying all along. 

 

Who wouldn't want a computer to do something if it could do it better than we could?  This is a non-sequitor.

 

In the meantime, we have reality, which seems to be everyone's counter-point - another non-sequitor.   How many thousands of posts will it take to see this obvious circle-jerk?

Link to comment
Share on other sites

1 hour ago, Mallette said:

I really don't get this any more than I understand why trains are under human control. 

 

If you think about it Dave it's really easy to understand. The answer is simply that we ain't Japan. They got 319 MPH technology/funding, we got 60 MPH technology/funding.

 

Keith

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...