California DMV Proposal: Self Driving Cars Will Require a Driver

Google: "gravely disappointed"

[engine]

According to draft proposal from the California Department for Motor Vehicles the recommendations require that self driving cars must have driver controls in the advent of a system failure. The idea being that the driver can take over in the event of a problem.

Google, one of the most outspoken driver-less car developers says it is "gravely disappointed"

Of course, they are not the only ones that are testing driver-less vehicles.
Until the technology is proven, I suspect there's going to be a level of uncertainty. There are so many things that could go wrong with the technology: Systems failure, hacking, mechanical failure, bad weather, natural disasters, etc.
However, it does seem that a driver in the cab on standby is just going to be a waste of the technology, so i can understand the negativity from developers.

In a statement today, Google decried the proposal, saying California’s rules would hold back a technology with the potential to prevent car crashes and improve the mobility of people who currently cannot drive.

“Safety is our highest priority and primary motivator as we do this,” spokesman Johnny Luu wrote in an e-mail. ”We’re gravely disappointed that California is already writing a ceiling on the potential for fully self-driving cars to help all of us who live here.” U.S. DMV: Self Driving Cars Will Require a Driver [autonews.com]

[tangor]

Driverless cars threaten a government revenue stream: license renewals. Think about it. :)

Additionally, driverless cars can lead to car fleets, on call use, and fewer "property" owners, who are also taxed.

[LifeinAsia]

Driverless cars also negate one of the big incentives for people to use mass transit- being able to do work/sleep/etc. during the commute instead of paying attention to the road.

I think this also a preemptive challenge to an uber Uber- a fleet of driverless Uber vehicles.

[lawman]

Revenue stream? Holy cow, some municipalities wouldn't exist without money from traffic violations.

[engine]

I'm certain the authorities will find a suitable alternative tax, so no fear there. Note, these are just proposals at this stage.

[tangor]

Yes, but those alternatives will not reach down to the personal level, where gubermint (sic) likes to control.

Meanwhile, driverless cars will all but kill the automobile Insurance market and MADD will lose it's political clout.

Worse, all sense of privacy in travel will disappear. The gubermint (or their associated "traffic control") will know exactly who, what, and where anyone or anything is transported.

[ken_b]

But right now the idea of driverless cars is driving an increase in the sale of tinfoil.

[piatkow]

According to draft proposal from the California Department for Motor Vehicles the recommendations require that self driving cars must have driver controls in the advent of a system failure. The idea being that the driver can take over in the event of a problem.

I would have thought that a manual override was a no brainer. No system is perfect and all that is needed is a control to allow the vehicle to be moved to a safe pace at low speed.

[lawman]

Not sure about manual override. Too many nincompoops will be overriding and that would defeat the benefit of the self driving cars. These vehicles should be able to handle all situations. If there is a system problem they should be designed to pull over at the first safe spot and notify OnStar or whoever. If car isn't ready for prime time, don't put it on the market.

[piatkow]

After 30 years working in IT one thing that I do know is that no system will ever handle ALL situations,.

Once you are off the highway you may need to move the vehicle to another part of the site or the site may have no highway quality signage and markings for the car's sensors to respond to. The ability to drive manually at 70mph would be unecessary but at some point you will need to undertake a manouvre at walking pace which the system won't understand,.

[toidi]

one thing that I do know is that no system will ever handle ALL situations,. 

and no system is imune from crash/freeze/bugs/hacks/stuff we haven't even heard of yet.

[lawman]

Piatkow, not sure your stated problems are insurmountable, but if you're saying manual override could be implemented only in very limited situations and have the system take over when the situation no longer exists, I could live with that.

Like I said toidi, if it's a system problem have it designed to pull over and notify. If it's a hack, it's doubtful even manual override would work in that situation.

[toidi]

If the system crashes, how will it know to pull over? How will it know what a safe space to pull over is? The side of the road could be two lanes over. Worse, the system doesn't crash, it just starts acting weird, mud obscures the sensors, a squirrle chews through some wires.

[lawman]

Maybe we should just pass this thread on to the ones trying to make autonomous cars work so they can give up and stop wasting time.

[LifeinAsia]

As a point of reference, I'm guessing that every day we have hundreds, if not thousands, of airplanes flying around on autopilot, many of they carrying hundreds of passengers. Although takeoffs and landings are done manually by the pilots, it's my understanding that the newer ones can handle those functions (at least landings) as well.

We're talking about moving a multiple ton object in 3 dimensions and having to take into account wind velocity and other factors that a self-driving car wouldn't have to. Granted, a 747 doesn't have to worry about a deer running across the road, a patch of black ice, or a drunk driver weaving into its flight path. But I'm sure there are some safeguards built in to take into account an unexpected object in the flight path.

It would be interesting to hear some input from someone with aircraft autopilot experience (designing/building/testing, not just using) and get their take on the differences between self-driving cars and self-flying planes.

[piatkow]

Piatkow, not sure your stated problems are insurmountable, but if you're saying manual override could be implemented only in very limited situations and have the system take over when the situation no longer exists, I could live with that.

Just what I was saying. I suppose it is a lifestyle issue a well. As a music publishing professional I attend a number of festivals every year and some involve going quite a distance off road to park in a field. I can't see an automated system ever following directions from a camp site steward directing you to your pitch.

[toidi]

Maybe we should just pass this thread on to the ones trying to make autonomous cars work so they can give up and stop wasting time.

They are not wasting time and i do not want them to give up. They are trying to rush them onto the market before the technology is ready. When they can create crash free systems then they will be getting close. In the meantime, the r&d is best kept to test tracks not the open road.

[lawman]

Thanks for clarifying toidi. It sounded an awful lot like you meant that no system is imune from crash/freeze/bugs/hacks/stuff we haven't even heard of yet.

[Leosghost]

Paging Mr Rumsfeld..Schrödinger's cat may or may not be in Mr Heisenberg's self driving car..