Driverless cars.

No, I'm not. A driverless car needs both, and cannot function without either one.

And because I know that I turn right just by the postbox after the Black Horse pub.

The nice man tells me. So shall we add speech recognition to the list of required technologies?

That should read: "cars than have been maintained by a team of technicians have a very impressive safety record"

Try extrapolating the MOT failure rate to see how well they'll be looked after in the real world

Given that the tech is being developed primarily by the US, where there is no MOT equivalent...

Not entirely true.

With some means of preventing it being pressed whilst in lane 3 of the M1 with a driver approaching from behind at warp speed in their Panzerwagen.

I'm not convinced they would even be capable of motorway operation and their programmed actions, acting in isolation might even lead to serious incidents.

An autohelm / self steerer on a boat doesn't absolve you from making effective and regular observations even when miles from nowhere. Autohelms for example obey the waypoint exactly, not a problem when it is a point in clear water but navigate to a buoy on a chart and with the right conditions that is *exactly* where you go, maybe at 15 knots directly into a half a tonne of floating metal, in a tupperware boat. An expensive repair, wet feet or a rapid one way trip to the bottom are possible outcomes. Don't think it can happen? Seen it with my own eyes in the Solent some 20 years ago from about 100 metres away. Force 3 or

Similar problem in the air with this near head on collision, in the arse end of nowhere in a near empty sky.

The pre - programmed default of 'it's all got too confusing for the system to work out wtf is happening' and coming to a complete halt or alerting the off duty driver to take over control could be way too late for any effective action. A driver, in control, using all visual and audible cues might have never got in the situation in the first place.

Include me out.

Personally, I'd rather drive myself if I

Both are needed but are different things.

A human driver needs both too. People that haven't yet become reliant on satnav have some sort of map in their head. And satnavs are everyday proof that machines make fewer navigation mistakes than humans.

You "know". By what internal mechanisms?

If the road is closed I'm not sure what verbal instruction can add. "The road is blocked, you can't go this way". Indeed.

Not at all sure that is proved. One trivial example: Just up the road from here, at least two entirely different satnavs, put in a left turn, right, right, left which adds a little diversion. Only, maybe, 100 yards or so extra, but entirely unneeded, unwanted, and on roads which really are back lanes rather than the main thoroughfares. A second slightly less trivial example: Also very close to here, a series of small "estate" roads, which satnavs won't even recognise when you are on them. They leave you in the "in the middle of a field" state. Not exactly difficult to get back onto recognised roads, but shows how satnavs can very quickly get lost. (These roads have been there for decades so newness is not an explanation.)

My partner has a valid licence and has insurance. But she does not drive because her health conditions make her feel less than fully in control. (None of the usual notifiable conditions.) So, for her own peace of mind and everyone's safety, she doesn't. A fully autonomous driverless car could help her enormously enabling travel when I am not available.

Nevertheless, I am deeply sceptical of the reality of driverless cars. Most especially mixed traffic with both driverless and driven vehicles interacting.

One of the most famous and appropriate ones for this conversation being the Mercedes and Volvo 'auto brake' fails on demo. runs in front of journalists.

And I never could get auto-park to work consistently when I had it.

What? Your saying I can't just go to sleep and let it chauffeur me to the destination?

Well that's just no good to me.

How's it going to get on with the random motorcyclist or even a cyclist?

Nothing is going to 100% successful. The main thing is that humans with expert knowledge can make machines get closer to 100% than humans without that expertise. How does someone find their way around there without any local knowledge? The difference in your example is that the satnav doesn't learn whereas a person won't necessarily make the same mistake twice.

What is lacking in standalone satnavs is a phone home feature. Cars are soon going to have a mandatory phone home feature, for reporting when they crash. There are plenty of things that can piggyback on that.

It sort of does; TomTom users (and I expect others) can submit alterations and additions. These eventually filter in to map updates.

Chris

You're just describing a simple navigation system that makes no attempt to supplement or replace human awareness. As I said earlier, navigation and awareness are two different things.

Walking speed. :-)

I can use voice control for phone calls and for media playing in my car.

Sometimes the bitch says "I do not understand the command"..

The Onstar System as installed in most GM cars in the US already has this feature by default.

Not just estate roads, I can very regularly corner right on one particular roundabout, on an A road, at a junction that has been unchanged for at least 50 years, and a few seconds later, phone based satnavs (demonstrated this numerous times on differing platforms and software) think I have missed the roundabout completely and proceed to calculate a new route which requires me to turn right a few hundred metres further on, which would be a valid route if I had indeed gone straight on rather than turn right at the roundabout.

It takes some 400 metres for the satnav to recognise that travel is in the direction originally directed in the approach to the roundabout, it then pisses about for a few seconds, recalculates the route and churns out the original series of directions it had stored just 30 seconds ago. There are no structures, military facilities or other transmitters nearby to cause GPS disruption and off the shelf stand alone GPS modules plugged into a laptop and dumped in the footwell shows plenty of satellites in view, and when the track is overlaid on google earth the path of the vehicle accurately follows the image of the road.

The only conclusion is that cornering at about 30 - 40 mph, occasionally with a bit of opposite lock, is somehow deemed an implausible manoeuvre:)