Menu Sign In Contact FAQ
Banner
Welcome to our forums

Removing the human factor from potential aircraft accidents?

Airborne_Again wrote:

That doesn’t follow at all. You assume both that the autonomous aircraft will at every moment be able to calculate the optimum action and that it has complete knowledge of the winds at every point in its projected flight path both in this and future instants. That is clearly not true

It doesn’t need to know “everything”, just enough to make a decision. It wouldn’t decide to return if it wasn’t 100% sure it would make it. Hence it would make it in 15 of 15 attempts.

ENVA ENOP ENMO, Norway

Not quite. A fully autonomous plane would either have been successful in 15 out of 15 attempts, or in zero out of 15 attempts. This would depend on the programming. If the programming was done in a way for the plane to make that “decision” to be made and to execute it, it will be more reliable executing this plan than humans.

The major problem with aviation is not the environment (less predictable than railway, more predictable than the roads) or that things fail more or less often than in other applications (there are hundreds of thousands of car break-downs, signalling failures, traffic outages etc. every day), but one simple, unchangeable fact:

All other traffic system have one fail-safe option that aviation does not have – stop. And stopping is really, really easy.

So basically – when stuff goes wrong, you work around it it. When stuff goes so wrong you can’t work around it, you fall back to doing the easiest thing possible. Just stop.

In aviation – when stuff goes wrong, you work around it. When stuff goes so wrong you can’t work around it, you die.

The only question is – when will the number of times people die be so low, that we accept the additional deaths that at least in theory could be prevented by humans.

The currently completely irrational reaction to deaths where car “autopilots” are involved indicates that this may be further away than we think.

Last Edited by Cobalt at 14 Sep 16:38
Biggin Hill

Peter wrote:

There are no autonomous ships, surely.

2greens1red wrote:

They are becoming autonomous as we speak……

Have any of you ever been on an oceangoing cargo vessel? The ongoing maintenance required to keep these things running, while not overwhelming, is constant. You simply cannot automate that. Might work on short routes, but ocean-crossings? Not so much. It’s also quite intriguing how difficult it can be to get a large vessel into port. Simply put – I doubt the RR claims.

Somewhat counter-intuitively, aviation and rail are probably the best candidates for automated mass transport, as they operate within their own – relatively closed – systems. More and more trains (especially urban mass transit systems) run autonomously, and drones can be pre-programmed to fly very intricate courses. However, as the Sully flight demonstrates (at least to me), they cannot make quick decisions that are outside the normal operating envelope, e.g. landing in the Hudson. The human brain is still pretty damn good at coming up with solutions to unforseen problems.

There’s also the basic question you always have to ask: ‘cui bono’? What’s the point in installing hugely expensive systems which need to be operated by skilled personnel, when you already have that personnel on board? I doubt very much that the cost savings are as massive in real life as some automation evangelists are claiming.

Patrick Smith is a well-known pilot / author / journalist, who sent me these links today, giving his opinions on pilotless aircraft. Well worth a read.

Pilotless Planes – Not so Fast
https://goo.gl/fgMgAH

NY Times article – Why Pilots Still Matter
https://goo.gl/wpfEkW

Cockpit Automation
https://goo.gl/2V2bEq

Swanborough Farm (UK), Shoreham EGKA, Soysambu (Kenya), Kenya

what_next wrote:

And somewhere I read that flight crews who tried the Sullenberger scenario in the simulator were able to glide into Teterboro. But only because they knew what was about to happen and began their diversion immediately.

I think did read the same, and if I recall correctly:

4/4 crews made it gliding to Teterboro when initiated the manoeuvre as soon as there was a sign of engines out.
0/4 made it when given a 30s delay in between engine failure and being able to initiate a course of action.

I think a computer could have identified the failure pretty easily. Regarding the ditching option, I think it’s likely the human might have done a better job. mostly because the human would be able to avoid small boats etc (and possibly know that it’s a good place to ditch as loads of ferry). But that’s one we’ll hopefully never know!.

I could see autonomy coming to Freight flights first. they could even at first operate entire different airports to minimise interference with human flights.
I think pilotless flight is not a question of if, but when. For passengers, given the time allowed for certification, I highly doubt that in 20years, but I wouldn’t be too surprised if some flights (maybe not the large commercial) were autonomous in 40

Martin wrote:

Because you want coordination

ATC is not the only way, and far from the most efficient way to organize a flock. Birds and fish do this much better than any ATC is capable of, and has done so for hundreds of mullions of years. Just study the activity at a bird mountain, and you will see just how inefficient and stone aged the ATC really is. A non towered, well organized VFR airport is also muxh more efficient than an ATC. The only reason for ATC, it it’s the only one with knowledge of everybody, their position, speed, heading, alt and their intentions. This can easily be communicated between autonomous vehicles, and each vehicle has their own algorithm to “follow the flow”, not unlike a busy non towered VFR airport.

I think it is interesting to look at the Munin project that investigated possibilities for autonomous ships. The overall strategy they came up with is like this:

MUNIN would normally rely on automatic and fully deterministic control functions to run the ship. However, various sensor systems will be needed to detect problematic situations such as unexpected objects in the sea, dangerous weather conditions or danger of collision. If an unexpected situation occurs, an autonomous control module will be invoked trying to remedy the situation within its given constraints. If the system cannot achieve this, it will request support from a remote operator or start a fail-to-safe procedure if the operator is not available. Properly implemented, this type of autonomy will reduce the need for human supervision while maintaining a high and well defined level of safety. However, a major challenge will be to device sensor systems so that all relevant dangerous situations are reliably detected and appropriately acted upon.

For an aircraft, it is hard to envision some sort of “fail safe” action. Also a remote operator jumping in and saving the day doesn’t seem very probable (unless he is on the plane being up to date all the time, which defeats the purpose). An aircraft would have to rely on the deterministic control functions and the autonomous control module only. At first thought this looks like a much more complex task for the autonomous control module, since it has no further fall back options. But at the same time, that is the nature of flying in any case. No matter what kind of system is used (manual or automatic), there are no fail safe actions. I plane cannot stop in thin air, waiting for things to get solved. The only question left is if the autonomous system, on average, produce better results than manual intervention (from deterministic control functions). I think it would, by orders of magnitudes, because such a system is able to “investigate” 100s or thousands of options with multitude of factors in no time.

ENVA ENOP ENMO, Norway

LeSving wrote:

because such a system is able to “investigate” 100s or thousands of options with multitude of factors in no time.

Has it ever occurred to you that that is exactly what our brain does – all the time?

Properly implemented, this type of autonomy will reduce the need for human supervision while maintaining a high and well defined level of safety. However, a major challenge will be to device sensor systems so that all relevant dangerous situations are reliably detected and appropriately acted upon.

Yeah, good MBA language there

Administrator
Shoreham EGKA, United Kingdom

I hope this isn’t too demeaning to the professional pilots here, but wouldn’t it save almost exactly the same amount of money to get rid of a couple the cabin crew and have the pilots serving drinks during cruise?

I mean it’s a bit unconventional, but actually doesn’t require much tech to implement, infact the copilot of a Ryanair could do it tomorrow if it wasn’t for the political implications.

And you still have two warm bodies up front during the critical phases of flight.

Last Edited by LondonMike at 15 Sep 20:55

LondonMike wrote:

I mean it’s a bit unconventional, but actually doesn’t require much tech to implement, infact the copilot of a Ryanair could do it tomorrow if it wasn’t for the political implications.

I believe that there is a minimum Cabin Crew head count that needs to be observed mostly for safety reasons in the unlikely event of having to exit the acft during an emergency.

FAA A&P/IA
LFPN
Sign in to add your message

Back to Top