Talk Morgan Forums Community The Malvern Arms Autonomous cars thoughts after Boeing crashes.

I have read the topic on the Boeing crashes and the MCAS system implementation and failings.
Yes an over reliance on computer based systems coupled with less than best training for system failures.
This has me wondering about autonomous cars.
One function they seem to stress is being quickly available and able to take manual control should the car sense a situation it cannot deal with effectively. This seems similar to pilot training to manually fly a plane when systems failure occurs.
Is there a parallel for car drivers?
I certainly think so. Being able to come from a semi switched off state to total control and quickly assess and react to why the car has instructed the driver to do so is potentially a hazard situation.
Are drivers capable enough?
Should training be introduced on these cars?
My answer is yes. Simulations of situations so that a driver has at least experienced how to react correctly as would skidpan training do.
Modern cars now have mainy driver aids and I believe this creates a sense of over reliance on them.

I agree, especially the last sentence when in the context of young drivers, eg lack of attention & late braking

Without a doubt! I imagine that in the event of a pilot needing to take control he/she has a few thousand meters between the plane and some disaster whereas in a car travelling at 70mph on the M25, mere inches would be more likely.
I very much doubt I could react quickly enough so thank you but no thank you.
My idea of an autonomous car is a taxi!!

At least a car doesn't think it is nose up and attempts to correct by putting the nose down and speeding up - or perhaps it could - hopefully different software...

Accidents happen because of the unexpected, no matter how smart the system is there will always be something that has not been anticipated by the developers/engineers so I expect the introduction of self driving cars to be something of an evolution, ie someone gets killed, the cause is investigated, the software solution is determined and downloaded globally overnight. This is why extensive tests are being carried by the manufacturers - to try and capture as many potential issues/ failings prior to going public.

I envisage a problem with this system is, the present Motor Manufacturers resistance to admitting major safety issues with their vehicles ( Bad Press?) and their further reluctance to issue recall notices (costs) . Test drivers are being sacrificed during the development testing process, these are admittedly at the pointy end - but and this is a big BUT . The Test Drivers are highly trained professional drivers, if these guys are crashing - what training is it possible to put together which will reduce the risk !

The biggest risk in most vehicles is the big sack of meat behind the wheel, the problem for autonomous vehicles is judging what these sacks of meat are randomly going to do next.

A pilot who uses his autopilot or other auxiliary systems remains permanently in a working condition, in his seat with the same body alignment. And if something goes wrong, he usually, not always, has time for corrections.

I suspect that the self-propelled car causes the occupants, especially the driver, to slide into a different mental state. Once a driver has got used to looking away from the road, talking face to face with other occupants, reading, playing with his mobile phone, there is no time at all left before he suddenly grasps a dangerous situation.

That's why the solution can only be completely or not at all automated. There is also another problem. If you really wanted to pay attention to how the car drives, you have to get very tired quickly if you have nothing active to do. It must be a boring and at the same time very strained feeling to drive like that.

Heinz "all or nothing". Does that mean all functions of the car or all of the cars on the road as they interact? I see an issue over the mixed population as well. I do not mean that as "so it cannot happen".

Don't worry. By the time they get autonomous cars up and running the new hyperloop network will be ready to go. You will hop into a pod at your local hyperloop station, key in you destination, and whoosh off you go at up to 4,000 miles an hour.

https://edition.cnn.com/2013/07/16/tech/innovation/elon-musk-tube-transport/index.html

Autonomous cars do have a driver, it's just that the driver is no longer on board; he / she programmed the driving some time before the car was manufactured and is not physically present to experience the outcomes. It's perfectly possible to fly an aircraft by remote control without a pilot on board but I suspect there would be market resistance to autonomous commercial airliners.