You are not logged in.
Seeing how far driverless cars have come, I don't think there will be much call for drivers on Mars...
https://www.youtube.com/watch?v=yjztvddhZmI
Apart from some possible requirement for human driving in an exploration setting (not certain but possible) we can be assured driverless vehicles will have no problem dealing with cleared road trails on Mars. This will be very helpful when it comes to bringing in raw materials such as ores from mining operations at a great distance from the industral processing centres. A steady stream of driverless rover trucks will bring in the material, possibly in convoy. Will be a great sight watching a convoy travelling through the night.
Let's Go to Mars...Google on: Fast Track to Mars blogspot.com
Offline
Louis,
Based upon everything I've read about what you envision people doing on Mars:
1. They're not going to drive anywhere
2. They're not going to explore anything
3. They're not going to be involved any construction tasks
4. They're not going to be involved in growing and caring for their food crops
5. They're not going to be involved any mining or prospecting tasks
6. They're not going to use any energy to accomplish those tasks
That sounds like a luxury resort for the obscenely rich, not a frontier colony environment. If we're not gainfully employing people, then why bother sending people to Mars at all? You can do all of those things right here on Earth if that's what you really want to do. The rich still use humans rather than robots, but that's pretty much what they already do. I'm not personally interested in moving to another planet tens of millions of miles from Earth so that I can have my own personal army of mechanical slaves waiting on me hand-and-foot. I want the personal freedom to live dangerously without constantly requiring permission from anyone else, to use my ingenuity and hard work to build a life for my family while expanding humanity's presence in space, and to use my free time to better understand Earth within the much greater context of our solar system and the universe by exploring another planet that was once remarkably similar to Earth.
Offline
Remote driving is technologically more practical than driverless. No AI needed, just remote control. Remote control of drones and equipment is mature technology now. It needs to have a high reliability obviously, before it is trusted with human passenger requirements. A remote control taxi, like a Johnny cab could play a big part in human mobility. A driver is extra space and extra weight. If you are calling a cab for a journey of ten miles say, then you don't care if it has range of only 20 miles. And the driver doesn't need to get home. I can see this opening a lot of opportunities, as only a minority of vehicles would need long range. Much cheaper energy storage can be deployed for the majority of vehicles.
Last edited by Calliban (2021-07-23 17:40:31)
"Plan and prepare for every possibility, and you will never act. It is nobler to have courage as we stumble into half the things we fear than to analyse every possible obstacle and begin nothing. Great things are achieved by embracing great dangers."
Offline
Louis,
The only way to truly learn and grow is to do things for yourself and to be permitted to make mistakes or even fail, however painful or costly they may be. If we're effectively putting people inside a very technologically sophisticated rubber room, how do you propose we accomplish that? We need people to boldly go where no one has gone before, you know? It's that kind of mission. There are times when it's not fun or seems like drudge work, but there are also times when it is incredibly rewarding. I can think of few other tasks as rewarding as building a city from scratch on another planet and exploring places that no human ever has. If we take that away from people, then what do we have left?
Offline
No luxury self piloting or AI cars on mars as the GPS and roads have not been mapped or created for a cars computer to read where it is let alone whether its on the right road.
Offline
Calliban,
I'm sure it'll be great from that aspect, but how many of those taxi drivers are going to "learn to code"?
We're creating all of this automated technology that mechanizes huge swaths of industry, but I don't get the sense that the people doing the mechanizing have given any thought about what the growing number of unemployed people are going to do with their time.
Offline
Calliban,
I'm sure it'll be great from that aspect, but how many of those taxi drivers are going to "learn to code"?
We're creating all of this automated technology that mechanizes huge swaths of industry, but I don't get the sense that the people doing the mechanizing have given any thought about what the growing number of unemployed people are going to do with their time.
In this specific case, the cars would still have drivers. They just wouldn't be in the car. Similar technology could be used to that already used for military drones. The driver may not even be in the same state or even the same nation.
Genuinely automatic driving, controlled by AI, is not practical yet and may not be in the foreseeable future. But remote driving could be done right now. It isn't necessarily easy technologically, but it already has precedent in drone control.
Last edited by Calliban (2021-07-23 17:59:42)
"Plan and prepare for every possibility, and you will never act. It is nobler to have courage as we stumble into half the things we fear than to analyse every possible obstacle and begin nothing. Great things are achieved by embracing great dangers."
Offline
My experience with computer controls is very mixed. Things that are hard-wired to do a job have turned out reliable, after half a century or so. Things that are programmable with alterable software are still too unreliable, as all of our experiences with desktops and laptops has proven since about 1980, and still today.
Even the decades-of-experience experts at JPL still have trouble with it, doing those probes in space and on other worlds. And look at Hubble: almost lost due to an unreliable computer they could not fix. It works again ONLY because there were some back-up units that had not also failed! And take one guess as to why the F-35 is years late and tens of billions of $ over budget. I double-dog dare you! Guess! (Answer: bad software code.)
If you have some computer-controlled self-driving car or remote-operated car squiring you around on Mars, and you have a Hubble-type computer failure because dust and extreme cold got into the works, a failure that strands you miles away from water and oxygen, just what are you going to do when you cannot get it going again? Get out and walk many miles back in one of those bulky immobilizing "space suits" they have been designing?
You're gonna die out there!
Now, if it's "just a car" with a hard-wired power supply and power controls, and all-mechanical manual driving controls (steering, brakes,
throttle, lights), the stranding is unlikely to happen in the first place! Why? Because it has way-to-hell-and-gone fewer possible failure modes! Every single line of code in a computer software controlled system is another failure mode. (So is parts count, but that ranges from a few to about a thousand or so. A LOT less!)
I repeat: EVERY SINGLE LINE OF CODE IN A COMPUTER SOFTWARE CONTROLLED SYSTEM IS ANOTHER FAILURE MODE!
Typically, computer software-controlled systems like that have hundreds of thousands to millions of lines of code. That is one whopping lot of failure modes! We already have troubles at home with items that are only thousands to at most tens of thousands of lines. Stuff like Windows 8 and 10, for example. What pieces of crap they have proven to be!
You DO NOT want that on the moon, on Mars, or even on any of the public highways right here on Earth! They push this for trucks and cars on the highways here because of the money to be made. IT IS NOT SAFE! It will NOT be safe for over half a century yet (if then)!
A car driving around sightseeing (for exploration purposes) is not the only off-world application here. There is also construction equipment: stuff like front-end loaders, backhoes, cranes, bulldozers, etc. You are simply NOT EVER going to automate the manual art (yes, I said ART, and art it is) of operating stuff like that! Unless you have operated such equipment (and I have), you will NOT understand the truth of that statement! So, all I can say is trust those of us who have, and thus know. You WILL need human drivers for all of those things, and more!
Now you might rig systems to remotely operate construction equipment like that, on the moon and on Mars. IF AND ONLY IF you have sufficient sensory information for the remote operator to see AND FEEL what has been going on! You might be successful, if you do this as a hard-wired system with as low a parts count as possible. You WILL NOT be successful if try to do this with software code, precisely because EVERY LINE OF CODE IS ANOTHER FAILURE MODE!
Obviously, I am no fan of AI stuff. I have half a century's workplace experience to back that assessment up.
GW
Last edited by GW Johnson (2021-07-24 12:18:55)
GW Johnson
McGregor, Texas
"There is nothing as expensive as a dead crew, especially one dead from a bad management decision"
Offline
They might drive for leisure. Maybe for exploration where it's rough terrain. But generally they won't need to drive. For much of the 19th and 20th century most people got around on foot or by train. They didn't own a horse and they didn't own a car. It didn't make them less human.
Robots still need oversight. If you had a mining outpost, I think you'd need a small team there to oversee robotic operations.
It's interesting to speculate where all the labour will go in a colony of say 10,000 (permanent and temporary residents)...I'd say something like this:
1. Administration, planning and design (e.g. urban planning).
2. Spaceport operations/communications
3. Agriculture, food processing and food preparation.
4. Energy sector
5. Recycling operations
5. Life support (including water and sanitation and waste disposal)
6. Scientific research/university staff/exploration missions
7. Health care and monitoring
8. Leisure, culture and sports/maintaining Earth-like-environments
9. Retail and warehousing
10. Tourism and hotels.
11. Advertising and marketing (primarily on behalf of Earth-based companies)
12. Interior maintenance.
13. Industrial processes and manufacture.
14. Transport maintenance.
15. Media work (both within the colony and in broadcasting to Earth).
16. Construction.
17. Mining operations.
It doesn't take long before you realise you've got 10,000 people being very busy.
Louis,
Based upon everything I've read about what you envision people doing on Mars:
1. They're not going to drive anywhere
2. They're not going to explore anything
3. They're not going to be involved any construction tasks
4. They're not going to be involved in growing and caring for their food crops
5. They're not going to be involved any mining or prospecting tasks
6. They're not going to use any energy to accomplish those tasksThat sounds like a luxury resort for the obscenely rich, not a frontier colony environment. If we're not gainfully employing people, then why bother sending people to Mars at all? You can do all of those things right here on Earth if that's what you really want to do. The rich still use humans rather than robots, but that's pretty much what they already do. I'm not personally interested in moving to another planet tens of millions of miles from Earth so that I can have my own personal army of mechanical slaves waiting on me hand-and-foot. I want the personal freedom to live dangerously without constantly requiring permission from anyone else, to use my ingenuity and hard work to build a life for my family while expanding humanity's presence in space, and to use my free time to better understand Earth within the much greater context of our solar system and the universe by exploring another planet that was once remarkably similar to Earth.
Let's Go to Mars...Google on: Fast Track to Mars blogspot.com
Offline
Make mistakes? Including killing someone on a road? Hmmm... I agree with the video guy that there comes a point where you have to say "Is that ethical if you can avoid it?" We didn't say "let engine drivers make mistakes on the railway", we brought in automatic signalling and automatic braking.
I realise millions enjoy driving. I quite enjoy it myself. It gives you at least an illusion of powerful autonomy, if only till the next set of traffic lights brings you to a halt. But I don't think it's really an essential part of being human.
I am a gradualist though. I think most people will adopt driverless driving because in the end the ability to rest, sleep, work, entertain yourself or properly engage with family will prove more alluring than the illusion of powerful autonomy. In somewhere like the UK where nearly the whole country is being turned into a 20 MPH zone, there isn't much fun in driving now anyway.
Humans will be required at every stage of building a city on Mars and yes I agree it would be fantastic to be part of such a process. People will need to plan, perform some of the trickier manouevres in the construction process e.g. perhaps plumbing and cabling, electrics and so on.
Louis,
The only way to truly learn and grow is to do things for yourself and to be permitted to make mistakes or even fail, however painful or costly they may be. If we're effectively putting people inside a very technologically sophisticated rubber room, how do you propose we accomplish that? We need people to boldly go where no one has gone before, you know? It's that kind of mission. There are times when it's not fun or seems like drudge work, but there are also times when it is incredibly rewarding. I can think of few other tasks as rewarding as building a city from scratch on another planet and exploring places that no human ever has. If we take that away from people, then what do we have left?
Let's Go to Mars...Google on: Fast Track to Mars blogspot.com
Offline
GW,
As someone who does computer programming for a living, I can attest to the fact that software complexity is very real complexity. I know that in theory electrical / electronic systems can be very reliable, but in reality faster chips and more sophisticated programming only means that things go wrong at the speed of light. The mere fact that a casual observer with zero knowledge of software and electrical engineering can't see the complexity of the overall solution doesn't make it any less complex, it only has the outward appearance of being less complex or is mechanically less complex. It's not the software that typically fails, though. I know how to write software that doesn't contain programming flaws with respect to what it's intended to do, but if the hardware it's running on experiences a bit flip or electrical short or other failure, then the software can behave in unintended ways.
Back in the day, control system complexities included a mess of wiring, actuator solenoids, hydraulics or pneumatics, and sensors within discrete control circuits. Now that every part of the control system is completely electrical, it outwardly appears to be simpler to the casual observer. In reality, it's absurdly more complicated than what it replaced, but there are fewer mechanical parts or discrete components. Instead of spending money on discrete parts, you're spending money on software.
Millions of lines of code is a bit over-the-top for most control systems that I'm aware of, but tens of thousands to low hundreds of thousands is quite common, and yes, evaluating the reliability of something containing hundreds of thousands of lines of code can be very complicated.
I personally have no use for AI-enabled, remotely-controlled cars. That's a bit flip / hack / bad programming disaster waiting to happen. Someone not thinking about what they're doing could accidentally create a situation where every car simply turns right or left, comes to a halt, or speeds up, and not only is that the end of the AI experiment with cars, it's also the end of every motor vehicle so-equipped. People seem to think that something malicious or intentional has to be done to make these things fail, but a simple power surge or reset or programming error is typically far more dangerous.
Offline
We've had rovers on Mars that have kept going for 12 or more years.
The idea we can't produce reliable driverless vehicles for Mars is nonsense. But of course, I accept anything can "go wrong".
However, I am sure on any road trail on Mars there will be multiple ways of attracting help. If you break down and are unable to communicate, your absence of progress will be noted. There will be in-vehicle radio communications and emergency supply points for oxygen, food and water etc. at walkable intervals. There may well be a cable laid roadside for communications (though I am not sure that will be necessary). Any rover on the road will have emergency supplies to last at least three days. Journey plans will be automatically filed (similar principle to when mountaineers go climbing or pilots go flying, filing their route beforehand), so a rescue party will come find you if you don't complete your journey within say three hours after you were supposed to complete it. On most road trails there will be other traffic (human and robot) that will stop for vehicles that are immobile and are showing a visual distress signal.
When it comes to exploring on Mars expeditions will have mutiple rovers so that people can be rescued if one breaks down.
Did you actually watch the video through? There's a lot of safety info at the end. These driveless vehicles are incredibly safe compared with human drivers.
My experience with computer controls is very mixed. Things that are hard-wired to do a job have turned out reliable, after half a century or so. Things that are programmable with alterable software are still too unreliable, as all of our experiences with desktops and laptops has proven since about 1980, and still today.
Even the decades-of-experience experts at JPL still have trouble with it, doing those probes in space and on other worlds. And look at Hubble: almost lost due to an unreliable computer they could not fix. It works again ONLY because there were some back-up units that had not also failed! And take one guess as to why the F-35 is years late and tens of billions of $ over budget. I double-dog dare you! Guess! (Answer: bad software code.)
If you have some computer-controlled self-driving car or remote-operated car squiring you around on Mars, and you have a Hubble-type computer failure because dust and extreme cold got into the works, a failure that strands you miles away from water and oxygen, just what are you going to do when you cannot get it going again? Get out and walk many miles back in one of those bulky immobilizing "space suits" they have been designing?
You're gonna die out there!
Now, if it's "just a car" with a hard-wired power supply and power controls, and all-mechanical manual driving controls (steering, brakes,
throttle, lights), the stranding is unlikely to happen in the first place! Why? Because it has way-to-hell-and-gone fewer possible failure modes! Every single line of code in a computer software controlled system is another failure mode. (So is parts count, but that ranges from a few to about a thousand or so. A LOT less!)I repeat: EVERY SINGLE LINE OF CODE IN A COMPUTER SOFTWARE CONTROLLED SYSTEM IS ANOTHER FAILURE MODE!
Typically, computer software-controlled systems like that have hundreds of thousands to millions of lines of code. That is one whopping lot of failure modes! We already have troubles at home with items that are only thousands to at most tens of thousands of lines. Stuff like Windows 8 and 10, for example. What pieces of crap they have proven to be!
You DO NOT want that on the moon, on Mars, or even on any of the public highways right here on Earth! They push this for trucks and cars on the highways here because of the money to be made. IT IS NOT SAFE! It will NOT be safe for over half a century yet (if then)!
A car driving around sightseeing (for exploration purposes) is not the only off-world application here. There is also construction equipment: stuff like front-end loaders, backhoes, cranes, bulldozers, etc. You are simply NOT EVER going to automate the manual art (yes, I said ART, and art it is) of operating stuff like that! Unless you have operated such equipment (and I have), you will NOT understand the truth of that statement! So, all I can say is trust those of us who have, and thus know. You WILL need human drivers for all of those things, and more!
Now you might rig systems to remotely operate construction equipment like that, on the moon and on Mars. IF AND ONLY IF you have sufficient sensory information for the remote operator to see AND FEEL what has been going on! You might be successful, if you do this as a hard-wired system with as low a parts count as possible. You WILL NOT be successful if try to do this with software code, precisely because EVERY LINE OF CODE IS ANOTHER FAILURE MODE!
Obviously, I am no fan of AI stuff. I have half a century's workplace experience to back that assessment up.
GW
Let's Go to Mars...Google on: Fast Track to Mars blogspot.com
Offline
Louis,
Make mistakes? Including killing someone on a road? Hmmm... I agree with the video guy that there comes a point where you have to say "Is that ethical if you can avoid it?" We didn't say "let engine drivers make mistakes on the railway", we brought in automatic signalling and automatic braking.
Would it be ethical to murder a single child if you could cure all cancers by doing so?
As some who has spent the past 20 years programming computers, admittedly mostly business systems for major corporations, I can tell you unequivocally that there is a major difference between a discrete computer control systems and one that has full authority over an entire motor vehicle or network of motor vehicles. If my sales history loading code has an issue, then the worst that can happen is that the sales history doesn't get loaded for last month. Every networked vehicle could become unusable at the exact same moment if the control network crashes or the cell towers fail.
In the quadrillions of miles that human drivers have driven motor vehicles, there has never been a single instance of an event wherein every driver suddenly decided to turn left or right at the exact same instant in time across multiple continents. That kind of incident can and does happen when networked control systems have full authority over traffic like network traffic (IP routing, for example), and it happens with boring regularity. Every computer on the network will suddenly decide to start routing its traffic to a single system that it overloads / overwhelms, a disaster that happens at the speed of light. Since the data being transmitted is simple business data, most of the time nobody dies as a result.
I don't get the sense that you have any slight clue about how dangerous this has the potential to be. Elon Musk sure does, and he's not so gung-ho on the idea.
Offline
Louis,
We've had rovers on Mars that have kept going for 12 or more years.
We have a team of engineers sitting in a room at NASA, carefully planning out where the rover will drive to, 24 hours to 1 week in advance of them sending the commands to actually drive there. I can moonwalk backwards through sand faster than those rovers actually drive.
How practical is it to have a semi-autonomous / remotely operated vehicle that drives slower than a human can walk, assuming it's not for a science mission where the tax payers can afford to keep an entire team of specialists in a computer room for 10+ years?
Offline
I have all sorts of trouble with automated systems. Most of the time, I do not understand what the questions (or the instructions) mean. I see English words, but in context, there is no common meaning. That is what I mean when I say "we share a vocabulary, but not a common dictionary". There is absolutely NOTHING about computer anything that is obvious or intuitive to me. This technology DID NOT EXIST until I was well into middle age, and my mind has since rigidified. There is no help for that.
I just went through that problem trying to get AT&T to fix a 60-Hz loud buzz on my landline phone line. This is the second repair request I have submitted; the first repair lasted less than 30 minutes. There is NO WAY to talk to a human being at AT&T. The automated voicemail system you drop through trying to create a repair ticket is more brainless, utterly brain-dead, than a squashed June bug. Right down to simulated clicking noises as if a real human being were typing something in.
I can tell this is utter BS, and I really resent these automated voicemail systems! It is so patently obvious that they care more about money than customer service, even on what is ostensibly a customer service line! That lie is why I resent it so. AT&T is not the only one I really resent for this reason, but it is one of the very worst.
GW
GW Johnson
McGregor, Texas
"There is nothing as expensive as a dead crew, especially one dead from a bad management decision"
Offline
I accept the point about scale but it does show that at such a vast distance, we can keep such vehicles alive and functioning despite the challenges of temperature shifts, dust storms and rough terrain. No reason to think it would be any different for big rovers on Mars.
Louis,
louis wrote:We've had rovers on Mars that have kept going for 12 or more years.
We have a team of engineers sitting in a room at NASA, carefully planning out where the rover will drive to, 24 hours to 1 week in advance of them sending the commands to actually drive there. I can moonwalk backwards through sand faster than those rovers actually drive.
How practical is it to have a semi-autonomous / remotely operated vehicle that drives slower than a human can walk, assuming it's not for a science mission where the tax payers can afford to keep an entire team of specialists in a computer room for 10+ years?
Let's Go to Mars...Google on: Fast Track to Mars blogspot.com
Offline
My car has operated without any problem with its various automated systems for the last 7 years. Pretty remarkable really. I managed to split a tyre on knife-sharp kerb and also smash a wing mirror - total driver error!
I accept your point about the way computers "talk" to us. That is partly I think a function of IT specialists who, in my experience, are completely incapable of setting out their thoughts in English - invariably their poor grasp of grammar and semantics means they end up producing ambiguous sentences, but they are of course quite unaware of the ambiguities themselves! I remember being blithely told by one IT guy in the early days of computing that you had to click the start button to switch off your computer as though that were obvious and why would I have been confused by that!...
So, not really a problem of the systems themselves I would say.
I've had a similar problem re Virgin media. You are dealing with people who are working from a script, won't tell you if they have any engineering experience and are clearly on a bonus for NOT sending round an expensive engineer to your home!
I have all sorts of trouble with automated systems. Most of the time, I do not understand what the questions (or the instructions) mean. I see English words, but in context, there is no common meaning. That is what I mean when I say "we share a vocabulary, but not a common dictionary". There is absolutely NOTHING about computer anything that is obvious or intuitive to me. This technology DID NOT EXIST until I was well into middle age, and my mind has since rigidified. There is no help for that.
I just went through that problem trying to get AT&T to fix a 60-Hz loud buzz on my landline phone line. This is the second repair request I have submitted; the first repair lasted less than 30 minutes. There is NO WAY to talk to a human being at AT&T. The automated voicemail system you drop through trying to create a repair ticket is more brainless, utterly brain-dead, than a squashed June bug. Right down to simulated clicking noises as if a real human being were typing something in.
I can tell this is utter BS, and I really resent these automated voicemail systems! It is so patently obvious that they care more about money than customer service, even on what is ostensibly a customer service line! That lie is why I resent it so. AT&T is not the only one I really resent for this reason, but it is one of the very worst.
GW
Let's Go to Mars...Google on: Fast Track to Mars blogspot.com
Offline
I agree that systems will have to have built in protection against terrorist attack. We are already living with these sorts of realities in relation to nuclear power stations, nuclear weapons, oil refineries, dam operations and rail systems. So it wouldn't be that new. I'm pretty sure you could design triggers that would bring traffic to a safe halt wherever any unusual pattern emerged.
Louis,
louis wrote:Make mistakes? Including killing someone on a road? Hmmm... I agree with the video guy that there comes a point where you have to say "Is that ethical if you can avoid it?" We didn't say "let engine drivers make mistakes on the railway", we brought in automatic signalling and automatic braking.
Would it be ethical to murder a single child if you could cure all cancers by doing so?
As some who has spent the past 20 years programming computers, admittedly mostly business systems for major corporations, I can tell you unequivocally that there is a major difference between a discrete computer control systems and one that has full authority over an entire motor vehicle or network of motor vehicles. If my sales history loading code has an issue, then the worst that can happen is that the sales history doesn't get loaded for last month. Every networked vehicle could become unusable at the exact same moment if the control network crashes or the cell towers fail.
In the quadrillions of miles that human drivers have driven motor vehicles, there has never been a single instance of an event wherein every driver suddenly decided to turn left or right at the exact same instant in time across multiple continents. That kind of incident can and does happen when networked control systems have full authority over traffic like network traffic (IP routing, for example), and it happens with boring regularity. Every computer on the network will suddenly decide to start routing its traffic to a single system that it overloads / overwhelms, a disaster that happens at the speed of light. Since the data being transmitted is simple business data, most of the time nobody dies as a result.
I don't get the sense that you have any slight clue about how dangerous this has the potential to be. Elon Musk sure does, and he's not so gung-ho on the idea.
Let's Go to Mars...Google on: Fast Track to Mars blogspot.com
Offline
The Apollo lunar rover was a simple electric car, driven by hand. None failed on the moon.
The X-15 was manually flown with the aid of a hard-wired (NOT software controlled) stability augmentation system. It was intended to relieve pilot workload, and that is what it did. There was 1 fatal crash out of 199 flights.
The Space Shuttle could be "manually" flown during entry and landing, but was usually flown in autopilot mode. Its control system was software based, using the original 8086 chip. It killed two crews out of about 120-some flights, not counting the drop test flights with non-space-capable Enterprise. The fleet was retired for multiple reasons, but prime among them was unavailability of 8086 chips, except used, off Ebay.
Mercury was automatically controlled by hard-wired stuff, from ground control. It could be flown manually during entry, and both Scott Carpenter and Gordon Cooper did that. Cooper flew manually because the automatic systems had failed, and I think that might have been true of Carpenter, too. 6 manned flights, no fatalities.
The control and track record with Gemini is similar, that being Mercury Mark 2 with a bigger ICBM for a booster. The closest thing to a loss was a stuck thruster on Armstrong's Gemini flight with an Agena docking target. He regained control manually (!!!!), and landed manually immediately, due to depleted thruster propellant.
Apollo flew with a flight control computer that was actually software-controlled. It was a minimal automatic control, and worked well, the biggest issue being the primitive man/machine interface. The one on Armstrong's lunar lander overloaded and crapped out twice. Armstrong landed on the moon in manual control mode.
So, what is the better track record in spacecraft? Manual control or automation?
Bear in mind, stability augmentation systems and autopilot systems are OK, if they are more hard-wired than software-controlled. The physical parts count with such things are not all that high (hundreds to thousands), and if there are only thousands of lines of code to debug, that job can be done well. But if you try to take it too far, and automate everything, you get hundreds of thousands to millions of lines of code, and the track record with that is quite poor. It gets back to every line of code being another failure mode. Too many is just too much!
GW
GW Johnson
McGregor, Texas
"There is nothing as expensive as a dead crew, especially one dead from a bad management decision"
Offline
Virtually all air miles are flown on autopilot and in bad weather crews switch to autopilot to land...I rest my case!
The Apollo lunar rover was a simple electric car, driven by hand. None failed on the moon.
The X-15 was manually flown with the aid of a hard-wired (NOT software controlled) stability augmentation system. It was intended to relieve pilot workload, and that is what it did. There was 1 fatal crash out of 199 flights.
The Space Shuttle could be "manually" flown during entry and landing, but was usually flown in autopilot mode. Its control system was software based, using the original 8086 chip. It killed two crews out of about 120-some flights, not counting the drop test flights with non-space-capable Enterprise. The fleet was retired for multiple reasons, but prime among them was unavailability of 8086 chips, except used, off Ebay.
Mercury was automatically controlled by hard-wired stuff, from ground control. It could be flown manually during entry, and both Scott Carpenter and Gordon Cooper did that. Cooper flew manually because the automatic systems had failed, and I think that might have been true of Carpenter, too. 6 manned flights, no fatalities.
The control and track record with Gemini is similar, that being Mercury Mark 2 with a bigger ICBM for a booster. The closest thing to a loss was a stuck thruster on Armstrong's Gemini flight with an Agena docking target. He regained control manually (!!!!), and landed manually immediately, due to depleted thruster propellant.
Apollo flew with a flight control computer that was actually software-controlled. It was a minimal automatic control, and worked well, the biggest issue being the primitive man/machine interface. The one on Armstrong's lunar lander overloaded and crapped out twice. Armstrong landed on the moon in manual control mode.
So, what is the better track record in spacecraft? Manual control or automation?
Bear in mind, stability augmentation systems and autopilot systems are OK, if they are more hard-wired than software-controlled. The physical parts count with such things are not all that high (hundreds to thousands), and if there are only thousands of lines of code to debug, that job can be done well. But if you try to take it too far, and automate everything, you get hundreds of thousands to millions of lines of code, and the track record with that is quite poor. It gets back to every line of code being another failure mode. Too many is just too much!
GW
Let's Go to Mars...Google on: Fast Track to Mars blogspot.com
Offline
GW,
I have all sorts of trouble with automated systems. Most of the time, I do not understand what the questions (or the instructions) mean. I see English words, but in context, there is no common meaning. That is what I mean when I say "we share a vocabulary, but not a common dictionary". There is absolutely NOTHING about computer anything that is obvious or intuitive to me. This technology DID NOT EXIST until I was well into middle age, and my mind has since rigidified. There is no help for that.
I can explain why that is.
The first thing you need to understand is that everything in software engineering is contextually-based. Secondly, everything in computer programming that I do, or that any other software engineer does, is arbitrary to a significant degree. Beyond the basic concepts required for a Turing-complete language (a general purpose programming language) and the use of boolean logic, the rest of the lexicon (the language's dictionary) and its functions, is or can be very arbitrary. Standardization only exists to the degree necessary for programmers to comprehend the language constructs. It's not like math, where an arrangement of symbols can only have one meaning. This is one of many of the problems that Stephen Wolfram's Mathematica was attempting to solve using symbolics. Both language constructs and the context of the syntax can be and are altered at will, to suit the style of the person creating the programming language or manipulating its constructs.
The C language may as well be Cantonese to a programmer who is accustomed to using a verbose English-like Visual Basic language, for example. However, once you've used enough different languages, you begin to see how very similar they all are, but you also begin to understand that it's all trivia, and that what you really need is the lexicon and actual examples of how a task can or should be performed using some specific language, as well as what tasks a language is manifestly unsuited to performing. For example, I required no special "learning phase" when I first started using C# after years of using Java, I only needed to know the lexicon, because everything else about both languages was so similar- meaning and usage was identical in many cases. It was like using "Microsoft's Java library of Microsoft-specific functionality". I use new programming languages with some regularity and I like doing it, so learning yet another language is like purchasing a new tool to put in my tool belt. I'll don't need to think twice about doing it.
Regarding mindset solidifying with age, I have the exact same problem or reaction to new motor vehicle technology. In my mind, a steering wheel / column is supposed to be directly connected to the steering gear. If it's not, then it's "wrong" or "bad" (to me). It took me almost 5 minutes to get a Mercedes-Benz in gear, because I did not understand the concept of a tap shifter. The gear shift was so soft that I couldn't feel anything, but I released the tiny little lever a bit too soon or held it too long, so it wouldn't go into gear. To my way of understanding gear shifts, that was a bad design, but someone at Mercedes obviously thought it was a good idea. With a mechanical shifter, if the shifter won't go into gear or the car doesn't start / stop rolling, that's how I immediately "know" that something is wrong. With the electronic shifter, there's no audible or kinesthetic response to indicate that the vehicle is in gear. When you put a GM vehicle in gear, you can feel it, and I like that. What I do know is that no wiring, switches, servos or solenoids are actually required to put a GM-built truck in gear, because GM's design has a mechanical linkage. If something isn't working, you can feel it immediately.
Offline
Louis:
The autopilots you cite are closer to hard-wired systems than software-controlled systems. The recent exception has been the B-737 Max disaster, which makes my case, not yours: Boeing was trying to relay on software to compensate for an otherwise fundamentally non-airworthy design. That is a matter of the record, BTW. It has to do with relocating the thrust axis vs the vertical position of the vehicle cg, forced by the larger engines they wanted to use, instead of simply lengthening the landing gear.
You just made my case by citing aircraft autopilots. I know you did not intend to, but you did!
About 4 decades ago, I experienced a rough landing in an early B-737 with a hard-wired autopilot. About 5 decades ago, I was in a non-fatal (by the skin of our teeth) air crash in an Lockheed Electra-2. That one had a hard-wired autopilot. In both cases conditions occurred that were way outside what the autopilot was expected to cope with. They were due to violent weather.
In both cases, the pilots took full manual control.
In the Lockheed case 5 decades ago, the craft stalled and fell short, reaching the concrete by about a yard (which is why I am still here to debate with you), then bounced like a rubber toy about a mile down the runway until the pilots finally regained control, striking sparks and dragging wingtips and props on the runway. My fear of heights today is PTSD from that crash then. We didn't know what "PTSD" was back when that happened.
In the early B-737 case, the pilots managed to stop the stall, and just made a rough, bouncy landing. I knew, and they knew, just how close we came to dying, but no one else on the airplane did. I could tell, from their reaction to what I said to them as I left the aircraft. It was "y'all done good", with a hand gesture showing the radical gust-induced yaw followed by leading wing stall.
Both cases were precipitated by 45+ degree yaws, induced by extreme turbulence.
In both cases, the autopilot was inadequate to the task, but the human pilots were capable (again, that is why I am still here!). This was very minimal automation, to boot. Software-controlled is actually worse, because most software programmers have little or no experience with the real-world things they are trying to control.
Automation is VERY definitely NOT what you (and so many others) crack it up to be!
GW
Last edited by GW Johnson (2021-07-25 17:04:14)
GW Johnson
McGregor, Texas
"There is nothing as expensive as a dead crew, especially one dead from a bad management decision"
Offline
Kbd512 re post 21:
Well, I can tell you really understand the computer stuff. In my day, I learned scientific programming within Fortran-2 and got really good in the workplace doing things in Fortran-4. That was in the days when jobs were submitted punch card "batch" in steel trays, and turnaround time might be hours. I also learned a little Basic, and used something called QuickBasic 4.5 on my PC's, which only ran under DOS. I never ever really understood what we called "job control language", which seemed to be machine-specific, and there was NEVER EVER a manual available. The analogue to it in PC's seems to have been DOS, but I am unsure that I am right about that. Like I said, I never understood any of that stuff.
What happens to me when I try to deal with laptops (and especially these damned smart phones!!!) is analogous to what you felt with that Mercedes shifter. Maybe worse. I see English words, but I cannot answer the questions during an install process, because I simply do not understand the meaning intended. "No common dictionary". Every question is a show stopper, when you cannot understand what the software wants.
I had troubles with electricity and electronics, too, from not having the right teacher in my first electrical engineering class. Electricity was an evil invisible monster that would sneak out of an outlet and bite you. It took decades before I got comfortable enough and good enough to do straightforward stuff like wire my farm shop (and I got all the polarities correct first time up, too).
Airflow is also invisible, but seems real enough to me. Lots of other folks have trouble with it though. I majored in it. And I worked decades in it. Air flow, thermo, heat transfer, all the way from 5 mph to hypersonic. I was pretty good at all that stuff. I did it mostly by figuring with pencil and paper. The fancy CFD codes still didn't work right when I was in engineering school. So, I never did it that way. The "exception" was automating in Fortran or Basic the same repetitive calculations I did as a fill-in matrix on a big piece of paper. This was 2 decades before there was any such thing as spreadsheet software.
I have managed over the decades to become minimally-competent with Word, Excel, and Powerpoint in the Office suite, and using Google chrome as a browser. I did get my laptop replaced, finally. It was really hard to get all its questions answered to get it set up and running, but I am using it for this posting. I still have to get all my recovered data loaded, and get the wireless mouse running again. I used to be fairly good at drawing sketches in Windows' Paintbrush (the 2-D, not the 3-D), but I have to have a mouse to do that. Net upshot: new laptop is still a work-in-progress.
GW
Last edited by GW Johnson (2021-07-26 11:58:02)
GW Johnson
McGregor, Texas
"There is nothing as expensive as a dead crew, especially one dead from a bad management decision"
Offline
Autopilot systems use computers to generate the control output. The control software reads the aircraft's current speed, pose, height and location and then issues control signal to a flight control system, which is a lower-level actuator controller, to adjust the control surfaces of the aircraft in order to maintain the aircraft's attitude, height and speed while guaranteeing the lateral, vertical and longitudinal stability.
https://www.sciencedirect.com/topics/ea … atic-pilot
Autopilots are clearly governed by software.
Louis:
The autopilots you cite are closer to hard-wired systems than software-controlled systems. The recent exception has been the B-737 Max disaster, which makes my case, not yours: Boeing was trying to relay on software to compensate for an otherwise fundamentally non-airworthy design. That is a matter of the record, BTW. It has to do with relocating the thrust axis vs the vertical position of the vehicle cg, forced by the larger engines they wanted to use, instead of simply lengthening the landing gear.
You just made my case by citing aircraft autopilots. I know you did not intend to, but you did!
About 4 decades ago, I experienced a rough landing in an early B-737 with a hard-wired autopilot. About 5 decades ago, I was in a non-fatal (by the skin of our teeth) air crash in an Lockheed Electra-2. That one had a hard-wired autopilot. In both cases conditions occurred that were way outside what the autopilot was expected to cope with. They were due to violent weather.
In both cases, the pilots took full manual control.
In the Lockheed case 5 decades ago, the craft stalled and fell short, reaching the concrete by about a yard (which is why I am still here to debate with you), then bounced like a rubber toy about a mile down the runway until the pilots finally regained control, striking sparks and dragging wingtips and props on the runway. My fear of heights today is PTSD from that crash then. We didn't know what "PTSD" was back when that happened.
In the early B-737 case, the pilots managed to stop the stall, and just made a rough, bouncy landing. I knew, and they knew, just how close we came to dying, but no one else on the airplane did. I could tell, from their reaction to what I said to them as I left the aircraft. It was "y'all done good", with a hand gesture showing the radical gust-induced yaw followed by leading wing stall.
Both cases were precipitated by 45+ degree yaws, induced by extreme turbulence.
In both cases, the autopilot was inadequate to the task, but the human pilots were capable (again, that is why I am still here!). This was very minimal automation, to boot. Software-controlled is actually worse, because most software programmers have little or no experience with the real-world things they are trying to control.
Automation is VERY definitely NOT what you (and so many others) crack it up to be!
GW
Let's Go to Mars...Google on: Fast Track to Mars blogspot.com
Offline
Also required data signal position input from a number of sources.....
Offline