Taxies BUT in the Sky - Boeing-backed Wisk Aero reveals a four-seater autonomous air taxi

[Takumidesh]

2 minutes ago, Sauron said:

A demonstration during an air show is not a good example here

Those pilots were air transport pilots, there were 136 people on board, 3 fatalities, and 50 injuries. These were pilots that were entrusted with the lives of over 130 people. regardless of them being at an air show, they behaved recklessly and disregarded the lives of themselves and the passengers. Both of the pilots were very well regarded with thousands of hours of flight time, accolades, and the FO was the training captain of the airline.

3 minutes ago, Sauron said:

You can't account for every possible unforeseen mechanical malfunction and if you could there wouldn't be any.

The problem was unforseen for the first crash yes, and you are right, any software in place most likely would not have been able to correct for it. However a second plane crashed where the pilots behaved the same way. and a third plane had the same issue only the reversal returned back to normal operation before the plane crashed. that's three incidents where the same behavior occurred and the pilots all three times failed to correct the situation.

 

 

[suicidalfranco]

how is this better than an heli?

[Brooksie359]

28 minutes ago, Takumidesh said:

I hear about a lot of situations because I worked as a licensed commercial aircraft mechanic for years. I have also been involved in FAA investigations personally (though not from my doing). I have had extensive training on human factors, as wells as Gen Fam certs for Boeing 737NGs and Airbus A320. I was engine run and taxi qualed on Boeing 737NGs.

I actually have first hand experience riding jump seat from things like mis-adjusted trim tabs causing extreme pitch changes on simulated hydraulic failures and similar situations.

 

Often times one screw up causes the other to screw up.

 

That is exactly my point. the pilot did not improve the situation by being a pilot, if the pilot did the exact same thing the machine would have done, what was the benefit of having the pilot (in this instance) and yes I am familiar with PID loops and machine controls. and not every automated control is a PID loop on an aircraft.

 

A piece of code absolutely can solve that problem. The pilot naturally rejected the feedback of other systems of the aircraft because their intuition told them it should be happening this way. It is a similar phenomena to death spirals pilots experience in IFR conditions.

 

One of the most infamous crashes is air france 296, where the pilots overconfidence and disregard for the systems present in the aircraft resulted in a crash.

The Plane was not equipped to be flown on such a small runway, requiring all VFR, and the purpose of the demonstration was to effectively bypass every restriction of the plane, deliberately disabling or otherwise maneuvering around the flight assist as much as possible, including a last minute approach change to a different runway to get closer to the crowd.

 

The plane crashed because of deliberate over confidence and defeat of the protections in place by the aircraft.

 

EDIT: I should add about the 737 hardovers. had the pilot been able to identify the actual location of the rudder ( which was deflected in the opposite direction than intended) they would have understood how to correct their actions. This is something a computer (and pid loop) would easily be able to correct for.)

 

From what I have looked at it seems like the pilots are the ones inputting alot of the data that the autopilot uses to run so its not like autopilot runs 100% independently of the pilot. Also there are cases where autopilot has malfunctioned and resulted in a crash so let's not pretend that pilots are the only ones that cause crashes. 

[Sauron]

16 minutes ago, Takumidesh said:

Those pilots were air transport pilots, there were 136 people on board, 3 fatalities, and 50 injuries. These were pilots that were entrusted with the lives of over 130 people. regardless of them being at an air show, they behaved recklessly and disregarded the lives of themselves and the passengers. Both of the pilots were very well regarded with thousands of hours of flight time, accolades, and the FO was the training captain of the airline.

Ok, and? Again, they were deliberately being unsafe. Accidents like this are rare to the point of irrelevance. There are plenty of equally disastrous software failures I could point to.

18 minutes ago, Takumidesh said:

The problem was unforseen for the first crash yes, and you are right, any software in place most likely would not have been able to correct for it. However a second plane crashed where the pilots behaved the same way. and a third plane had the same issue only the reversal returned back to normal operation before the plane crashed. that's three incidents where the same behavior occurred and the pilots all three times failed to correct the situation.

Software doesn't just magically evolve every time it crashes a plane... if the problem was identified in the first crash in a way that would have allowed implementing a software solution it should have been fixed mechanically. And the pilots would likely have been trained specifically for this situation... but of course training costs money so I wouldn't be surprised if the company just ignored the issue until more people died. You know what also costs money? Software engineering.

[Takumidesh]

38 minutes ago, Brooksie359 said:

let's not pretend that pilots are the only ones that cause crashes

Never claimed that.

38 minutes ago, Brooksie359 said:

From what I have looked at it seems like the pilots are the ones inputting alot of the data that the autopilot uses to run so its not like autopilot runs 100% independently of the pilot

Sure, they take the atis information for example and use their iPads to do all of their weight and balance calculations and type it all into the FMS and it calculates max togo information airspeed and throttle limits along with many other parameters. the weight and balance calcs are mostly automated, and the atis information is delivered via an automated system over radio (it is literally microsoft sam telling you the weather over vhf). There is no reason that stuff cannot be cut out from the pilots responsibilities. Oh and the pilots are often entering that information because they are part of disparate systems.

 

I am not saying that a current flying aircraft can be turned into a fully automated one today. We are discussing a plane that won't be production certified for at least a decade, and is being built up from the ground to be automated.

 

In my opinion we will see completely autonomous commercial aircraft before we see completely autonomous cars. My main point is that the infrastructure exists for aircraft. Everything is primed and ready for it.

 

22 minutes ago, Sauron said:

Ok, and? Again, they were deliberately being unsafe. Accidents like this are rare to the point of irrelevance. There are plenty of equally disastrous software failures I could point to.

I don't understand how it is not relevant? a human pilot made deliberate mistakes and killed people, but because its rare, that makes it irrelevant?

 

How about suicidal pilots that fly into mountains, or pilots that fight with each other and forget to fly the plane, or pilots that forget to enable the cabin pressurization and experience hypoxia so fast they can't even figure out what button to press, or pilots that ignore master caution lights. A computer does not have these problems.

 

https://en.wikipedia.org/wiki/Germanwings_Flight_9525

https://en.wikipedia.org/wiki/Helios_Airways_Flight_522

for some non demonstration examples.

 

22 minutes ago, Sauron said:

Software doesn't just magically evolve every time it crashes a plane... if the problem was identified in the first crash in a way that would have allowed implementing a software solution it should have been fixed mechanically. And the pilots would likely have been trained specifically for this situation... but of course training costs money so I wouldn't be surprised if the company just ignored the issue until more people died. You know what also costs money? Software engineering.

My point in sharing information about that crash, is not that a computer hypothetically could or couldn't have better. It is that the pilots DID NOT do better, even after being made aware of the crashes.

 

FADEC completely eliminated the flight engineer from being a member of the flight crew, they do not exist anymore. modern aircraft engines are completely throttle by wire and the I/O has changed from monitoring an entire panel of gauges, to having a list of 5-10 numbers and some throttles. Over time more and more work will be done by the aircraft and the pilots input will become less and less.

 

Like I said earlier, new airbuses under normal law WILL NOT let the pilot do dumb stuff, and even in direct law still override the pilots. The engineers of these planes seem to think that pilots can't always be trusted.

 

You are right that stuff costs money. you are correct that airlines and manufacturers will ignore issues and all of that, there is corruption in the FAA and DOT, all of those things are true and I am not arguing that. But pilots are not infallible and pilots have in the past and will in the future, make poor decisions. I understand for a lot of people it feels wrong that a computer may make the wrong decision instead of a person, but is that really any better or worse? if the computer makes the mistake or the pilot makes the mistake?

EDIT: I want to add, my intention is not to sound overly argumentative. so apologies if i come off as heated. I realize that I am typing a lot of text walls. I just really like airplanes.

Edited October 3, 2022 by Takumidesh
more info.

[Sauron]

25 minutes ago, Takumidesh said:

I don't understand how it is not relevant? a human pilot made deliberate mistakes and killed people, but because its rare, that makes it irrelevant?

Yes, this is a statistics game. One instance (or a handful of instances) of a pilot intentionally breaking the rules is not sufficient evidence of humans being less reliable pilots than autonomous machines overall. We don't really have fully autonomous aircraft just yet so it's really hard to make a direct comparison, but in the situations you described it's perfectly possible a machine would have made the same exact error if not worse - and you don't hear about all the cases where a pilot was quick witted enough to avoid a crash despite a mechanical failure or other unforeseen circumstances that would have definitely stumped a machine.

28 minutes ago, Takumidesh said:

How about suicidal pilots that fly into mountains, or pilots that fight with each other and forget to fly the plane, or pilots that forget to enable the cabin pressurization and experience hypoxia so fast they can't even figure out what button to press, or pilots that ignore master caution lights. A computer does not have these problems.

Computers have a lot of other problems. And remember it's still humans programming them.

29 minutes ago, Takumidesh said:

Like I said earlier, new airbuses under normal law WILL NOT let the pilot do dumb stuff, and even in direct law still override the pilots. The engineers of these planes seem to think that pilots can't always be trusted.

Of course! Machines can be extremely helpful and yes, there are human blind spots that machines can compensate for. You can also define boundaries of behavior for humans that more often than not make the whole thing safer.

30 minutes ago, Takumidesh said:

I understand for a lot of people it feels wrong that a computer may make the wrong decision instead of a person, but is that really any better or worse? if the computer makes the mistake or the pilot makes the mistake?

I'd just rather have as few mistakes as possible... and so far the way you get that seems to be to have a human aided, but not entirely replaced, by a machine.

31 minutes ago, Takumidesh said:

EDIT: I want to add, my intention is not to sound overly argumentative. so apologies if i come off as heated. I realize that I am typing a lot of text walls. I just really like airplanes.

It's fine I design automation software for a living, my distrust for code when it comes to human safety is immeasurable.

[Kisai]

8 hours ago, Dominik W said:

Apologies to be the fear-mongering one but...

 

WHAT IF IT JUST STARTS FALLING?

 

With autonomous cars you can just take over the wheel. With a plane though, like what am I supposed to do? Accept death?

If someone knows let me know . 

You jest, but one of the municipalities here, when it was headed by the completely-out-of-touch mayor, actually proposed having air-taxies for the rich to bypass their urban sprawl... you know instead of building reasonable transit.

 

There are "air taxi" services already, but they're between city centers like Victoria and Vancouver, not short hops that would take you an hour by car or 40 minutes by metro, and they're just conventional helicopters.

 

Making them automated, and electric solves two problems but creates a handful of other problems, like air traffic control and weather. There is not enough ATC people in the world to have flying taxis.  The entire thing would need to be automated for it to function.

 

This vehicle seems like it's designed to do a helicopter-style VTOL but you can be sure that if these "took off", what would happen is they would be used for only rural (think Alaska) land traversal where more typical, dangerous, bushplanes would have been used. 

 

[Takumidesh]

4 hours ago, Sauron said:

It's fine I design automation software for a living, my distrust for code when it comes to human safety is immeasurable.

You make a lot of good points and that's why I was saying earlier, I think we will see full autonomous planes before cars. I definitely understand it's not there as it stands.

 

Also I also work in industrial automation now, which funnily enough has increased my trust in code, yes programming for edge cases is difficult but PLC safety when implemented correctly is very good.

[StephanTW]

looking as boeings coding from the recent past i hope the code is written by wisk aero, but honestly this just looks like all the prior multi roter flying car, why don't we just call them what they are helicopters. and the timing of this isn't great, economics aren't great and I imagine they aren't gonna be cheap to run or buy.
Also hope it fairs better then their other vertical takeoff ughum starliner

[mr moose]

16 hours ago, Takumidesh said:

By relying on a pilot you introduce a host of uncontrollable variables, by relying on TWO pilots you introduce even more uncontrolled variables.

 

having more than one pilot does not increase human variables,  by adding another pilot you simply add a redundancy for the first human variable. You wouldn't say that a second computer system that only kicks in if the first one fails is a extra set of variables, it is a redundancy measure, not an additional risk in process operation.

[TOMPPIX]

Elysium is becoming reality faster every day.

[Taf the Ghost]

I'd find this more interesting if they were talking up bringing in ultra-quiet rotor tech, but this really seems like just another massive waste of funds. It's an Electric Driver VTOL craft that will have no extra utility to a helicopter. You're still going to need pre-flight checks before every take off from a ground crew.

[Shzzit]

On 10/3/2022 at 8:07 AM, Dominik W said:

Apologies to be the fear-mongering one but...

 

WHAT IF IT JUST STARTS FALLING?

 

With autonomous cars you can just take over the wheel. With a plane though, like what am I supposed to do? Accept death?

If someone knows let me know . 

Lol time to bail out and pull the shoot brother. 

[Takumidesh]

On 10/4/2022 at 6:41 AM, mr moose said:

having more than one pilot does not increase human variables,  by adding another pilot you simply add a redundancy for the first human variable. You wouldn't say that a second computer system that only kicks in if the first one fails is a extra set of variables, it is a redundancy measure, not an additional risk in process operation.

Well if both the computers were confidence engines that conflicted with each other I would, Which is what humans are. The FO is not sitting waiting for the captain to die to start flying the plane, it is common for fight or flight (no pun intended) responses to kick in for captains who then defer to less experienced FOs because they get overwhelmed. Seriously everyone, read up on Crew Resource Management and human factors.

 

https://en.wikipedia.org/wiki/Crew_resource_management

 

https://www.hf.faa.gov/role.aspx

 

Quote

The men are utterly failing to engage in an important process known as crew resource management, or CRM. They are failing, essentially, to cooperate. It is not clear to either one of them who is responsible for what, and who is doing what. - https://www.popularmechanics.com/flight/a3115/what-really-happened-aboard-air-france-447-6611877/

Quote

A malfunctioning compass gave the crew an incorrect heading, although the instrument landing system and Global Positioning System indicated they were off course. The first officer made several attempts to indicate the problem to the captain but a failure to follow airline procedures and a lack of a standardized communication protocol to indicate a problem led to the captain dismissing the first officer's warnings. Both pilots were also overburdened with making preparations to land, resulting in neither being able to pay full attention to what was happening. -From the above wiki page.

Eastern Air 401:

Quote

the failure of the flight crew to monitor the flight instruments during the final four minutes of flight, and to detect an unexpected descent soon enough to prevent impact with the ground. Preoccupation with a malfunction of the nose landing gear position indicating system distracted the crew's attention from the instruments and allowed the descent to go unnoticed. - FAA report. https://en.wikipedia.org/wiki/Eastern_Air_Lines_Flight_401

 

CRM is monumental in preventing crashes. and is 100% something that A) is a human problem and B) is exacerbated by adding additional people to the flight-deck.

[mr moose]

21 hours ago, Takumidesh said:

Well if both the computers were confidence engines that conflicted with each other I would, Which is what humans are. The FO is not sitting waiting for the captain to die to start flying the plane, it is common for fight or flight (no pun intended) responses to kick in for captains who then defer to less experienced FOs because they get overwhelmed. Seriously everyone, read up on Crew Resource Management and human factors.

 

https://en.wikipedia.org/wiki/Crew_resource_management

 

https://www.hf.faa.gov/role.aspx

 

Eastern Air 401:

 

CRM is monumental in preventing crashes. and is 100% something that A) is a human problem and B) is exacerbated by adding additional people to the flight-deck.

failure to cooperate or to follow procedure is not the same thing as doubling the problems humans have.

 

In your propositions either the captain won't listen to the FO or the FO is handed control due to the captains insecurities.  In both situations only one human is in control and thus on one set of human variables apply.   In order for there to be a doubling of human variables both the captain and the FO would have to be each trying to fly the plane at the same time and ignoring the other.

[Takumidesh]

2 hours ago, mr moose said:

In order for there to be a doubling of human variables both the captain and the FO would have to be each trying to fly the plane at the same time and ignoring the other.

Airbus specifically considered this problem when designing the A320, putting an override on the flight stick to explicitly remove control from the other pilot.
 

Here is a very recent incident (however it did not result in a crash) where the pilots did quite literally what you just described.

 

https://nypost.com/2022/04/29/air-france-pilots-almost-crash-boeing-777-at-charles-de-gaulle/

 

Quote

In a preliminary report, France’s air-accident investigation agency said Thursday that the pilots “simultaneously made inputs on the controls” during the incident.

“The captain held the control column in a slightly nose-down position while the co-pilot made several, more pronounced, nose-up inputs,” the report said.

 

[mr moose]

20 hours ago, Takumidesh said:

Airbus specifically considered this problem when designing the A320, putting an override on the flight stick to explicitly remove control from the other pilot.
 

Here is a very recent incident (however it did not result in a crash) where the pilots did quite literally what you just described.

 

https://nypost.com/2022/04/29/air-france-pilots-almost-crash-boeing-777-at-charles-de-gaulle/

 

 

So the columns became decoupled due to different forces.  It sounds like the hardware did exactly what it was supposed to do and the pilot not following protocol was the issue.

[Heliian]

Most of these evtol things are scams. 

 

The legitimate ones being designed and tested will help future generations perfect and implement the technology.  

 

Forget about this electric vtol dream until incredibly light and powerful and efficient batteries are produced.  

 

Just look at the current tiltrotor aircraft, not quite ready for the civil market yet.  https://en.m.wikipedia.org/wiki/AgustaWestland_AW609

 

The v22 osprey and its replacement, the v280 valor, are the next best examples of the possibilities and those took decades and billions.