"Right to Drive" vs. autonomous vehicles

In reply to Ian F :

I'm here to point out that all of the safety tech for driver's aids is already tied to options now...

 

Ian F said:

The article ignores the 800 lb gorilla that can push the general public towards self-driving cars far faster than any laws could: Insurance costs.  When self-driving cars become more common and real-world safety numbers can be quantified, the insurance companies will have the power to make human driven cars unaffordable to all but the wealthy, save occasional-use classics. 

Federal, state and local governments won't need to be involved.

 

 

This argument does not make any sense. Insurance cost is entirely relative to risk and the cost of those risks.  How would a large number of autonomous cars increase the risk of driving a human piloted one?

I would pretty much guarantee insurance rates would go down very noticeably for human driven cars since you effectively remove almost all the risk of being hit and make it significantly harder to hit other cars when most cars are self-driving,

The bigger risk to human driven cars would likely be restrictions from driving in certain area/roads, such as roads that no longer use signals.

In reply to aircooled :

You are applying a path of logic I've never seen insurance companies use. Why are some cars more costly to insure than others? Because data shows they have greater risks to claims. If autonomous car data show a lower risk, then by default human driven cars will have greater risk, and thus higher costs - even if the actual risk doesn't really change. Insurance companies will still want to make money and are driven by greed more so than logic or a desire to serve their customers. 

In reply to Javelin :

But increasingly less so: back up cameras, for example and other safety items that are becoming standard on more models every year.

I could see Jay_W's scenario becoming reality. One reason why I could imagine an autonomous car for person commuting, but I have no intention of getting rid of my other cars for more frivolous uses.

Brett_Murphy said:

A long time ago, many cities used to have trolleys and trams that went around on tracks. As cars became more popular, away went the trolleys and trams.

For all the talk of self-driving cars, maybe the better solution is to bring back the trolleys and trams. 

I was growing up just as that happened. It took another decade  and a half to remove the tracks completely. 

But coverage was limited and only at certain times. If you didn’t live at one end of a line, one where you worked/ shopped:/ attended church etc at the other end the trip required waiting. 

What could be done in your own car took 3 -4 times as long or longer via the street car.  So GM spent much of their profit buying local politicians who supported having street car lines pulled up at tax payers expense. The money spent was recovered with the buses GM sold and profit from additional cars just became further incentive to further buy politicians. 

Not the first case nor the last case of the power of money to corrupt politicians.  

Ian F said:

In reply to aircooled :

You are applying a path of logic I've never seen insurance companies use. Why are some cars more costly to insure than others? Because data shows they have greater risks to claims. If autonomous car data show a lower risk, then by default human driven cars will have greater risk, and thus higher costs - even if the actual risk doesn't really change. Insurance companies will still want to make money and are driven by greed more so than logic or a desire to serve their customers. 

You're both talking about the same path of logic. Lower risk = lower premiums, higher risk = higher premiums. They make money either way. So the autonomous cars would - in theory - have lower risk and so lower premiums. The human driven cars could stay at the same rates or even go down and still be sigificantly higher than the autonomous cars. If there are no or minimal claims then the premium is all profit for the insurance company. Currently, the vast majority of auto premiums are paid out in claims so it's not like they're raking it in on the premiums themselves. Insurance is not a very opaque industry, really.

In reply to dculberson :

Safety plays a big role in determining insurance rates, but so do repair costs. All of the sensors currently used to facilitate autonomous driving are costly to repair. Even if the autonomous vehicles don't crash on their own, they'll have higher repair costs for the times when humans crash into them, trees fall on them, they hit a rain filled pothole that damages things, or a mountain goat gets defensive and rams your car repeatedly.

Time, and scale should help to bring those costs down somewhat, but they'll always be more costly than vehicles without the tech. And as the tech advances and older sensors become obsolete, they may be impossible to source. And the newest tech is nearly always more expensive initially than the older tech. There may still be enough of a net gain for autonomous vehicles to have lower rates than human operated ones, but I don't know if we can say for certain what impact they'll have.

At the very least, there's a lot of uncertainty around how these things will be insured, and what impacts they might have across the board.

In reply to STM317 :

Repair costs only come into play once there's an accident, and I'm just not sure the sensors are really going to add that much to a repair. We're already at thousands of dollars for minor body damage, a few hundred dollars in sensors is a drop in the bucket relative to that.

In reply to dculberson :

It might take awhile for a LIDAR sensor to be "a few hundred dollars". Sold individually, the industry standard is currently about $75k each.

They say that they're sold in bulk to manufacturers at lower costs, but still.

In reply to STM317 :

I'm quite certain that a regular car sold to an individual will not have $75k lidar sensors on it.

OEMs are already selling $300 solid-state LIDAR sensors with 90deg of vision (of which a car would need 3~4).

I'm going to start posting photos of roads that fully autonomous cars won't have a chance in hell at, starting with the parking lot at my work...

dculberson said:

In reply to STM317 :

Repair costs only come into play once there's an accident, and I'm just not sure the sensors are really going to add that much to a repair. We're already at thousands of dollars for minor body damage, a few hundred dollars in sensors is a drop in the bucket relative to that.

Just playing the devils advocate here though. I have been unimpressed by a typical body shop employee’s  ability to deal with electronic issues in cars.  I’m not sure the cost of components is important, more likely the cost of trouble shooting and proper diagnosis of which components need replacing. If it’s anything more complex than plug and play, I think we’re adding an outside contractor

dculberson said:

In reply to STM317 :

I'm quite certain that a regular car sold to an individual will not have $75k lidar sensors on it.

No way. The numbers don't work. But how far are we from the numbers working, and what how bumpy is the path to get to that point?

GameboyRMH said:

OEMs are already selling $300 solid-state LIDAR sensors with 90deg of vision (of which a car would need 3~4).

But are they useful for this application? Do they see far enough? Do they use too much energy? Is the laser wavelength perceptible to humans or not?

I'm not seeing anything close to that price wise in some quick searches. This looks promising, but they only claim to be "cheaper to make" than Velodyne's competitive product. That leaves a lot of room for pricing.

Innoviz units are already in BMW's newest cars.

dculberson said:

In reply to STM317 :

I'm quite certain that a regular car sold to an individual will not have $75k lidar sensors on it.

Not $75k, but the RLX front radar sensor was over $4k just a few years ago. 

Javelin said:

I'm going to start posting photos of roads that fully autonomous cars won't have a chance in hell at, starting with the parking lot at my work...

my road.

Where is the road? 

In the middle, right? It is basically a service alley. What happens when someone is coming at you? Well bear right of course.

Where you are now crashed into this concrete pad of death (shoe for scale, I'm out of bananas).

Right now the default mode is probably stop and do nothing and have everybody honk at you angrily.

Where does it park? In the lines like you're supposed to, or the opposite like the human driven 25 year old car?

Here's a good one.

This street is a one way, good to know. (Bonus - hit the jaywalkers or the car next to you?)

Yet one aisle over, it's not marked! Well the car knows from the on-board local map right? Except I ordered it from a different town, and it can't get onto the network from here. Car makes a right, does a head on collision and kills someone. Who's at fault?

Note that I haven't even left the parking lot yet!!

In reply to Javelin :

How do you know how the car will react? This whole thing seems pretty silly, honestly. I get it, driving is challenging. Don't you think the companies developing these things already know that?

In reply to Javelin :

Area identified. A few scans and some processing. Information uploaded to all cars. 

Or - the 360 deg scanning and processing power in each car improves to the point where it can "see" these conditions and calculate what to do.

it's just another problem to solve. And not an insurmountable one. Do you think yours is the only parking lot in the world that looks like that?

red_stapler said:

Bingo bango.

There's a reason all of the semi-autonomous stuff has been in very restricted and highly regulated areas.

In reply to Ian F :

My blanket statement was not about every individual. There are many exceptions. 

But the corporate voice of our GRM community is really negative on this subject every time it comes up. There are far more comments in this thread that basically say “I don’t want it”, or “It won’t work” than comments like “Hey guys, I have an idea to address that problem”

It was not an attack on anyone.  I’d rather it be an encouragement for us to work toward solutions.