Why should they try to do that on snow/ice covered roads? I mean, if its difficult and dangerous (as you imply) then they can just start with the easier trips and make loads of money. The hard stuff can come later..
This whole self-driving technology is worrying me a bit actually - I fully support this future, but I was thinking the other day that what will happen is that transport in the major hubs will get cheaper, but for rest of us it will get more expensive. For now, I'm ok but I am expecting a massive upheaval in the next 20 years
>Why should they try to do that on snow/ice covered roads?
Because, "Over 70 percent of the nation's roads are located in snowy regions, which receive more than five inches (or 13 cm) average snowfall annually. Nearly 70 percent of the U.S. population lives in these snowy regions." (http://www.ops.fhwa.dot.gov/weather/weather_events/snow_ice....)
If they're able to nail down "the easy stuff" that's all well and good, but it sounds to me like most of the country is only going to benefit from the current technology indirectly (maybe a third of the supply chain will be automatically driven). But this "in a few years nobody will own a car because we'll just rent from fleets that will come right to your doorstep!" thing I hear a lot still feels like a pipedream to those of us who are currently prepping for 6 months below freezing.
Who knows, maybe global warming will take care of this problem for us?
> a human driver accompanied the beer delivery, but his driving services were only used to get on and off the highway. The driver rode most of the way in the back of the truck, away from the controls.
> A 120-mile Colorado beer delivery, 45,000 cans of Budweiser beer
> Budweiser paid the market rate of $470 for the historic beer run
According to the DoT, they "took detailed measures to reduce any risks associated with the self-driving delivery project including testing validation, ride alongs and escorting the delivery" (source: https://www.trucks.com/2016/10/25/ubers-otto-hauls-budweiser... )
The article clearly states that no one was in a position to take over if the system failed. They were testing on public roads, and if something had gone wrong, we're talking about a loaded 18-wheeler: having escort vehicles would do nothing to mitigate a serious accident. Other people were on the road as well, none of whom volunteered to be part of the experiment. Short of closing the highway, there was no safe way of doing this. We're not talking about a mature technology here. We have no idea what the real, operational failure rate is. It was an irresponsible decision to put people's lives at risk for a corporate publicity stunt. FTA: "I think real implementation of technology like this is still five years off." Then so should this kind of experiment on public roads.
As a side note, I appreciate that you took the time to articulate your argument, rather than mindlessly downvoting. It makes the site much better when there's actual discussion.
Sure. Close down the road, get a bunch of professional drivers to simulate traffic (including poor driving), and have it at for a few hundreds of thousands of miles.
Let's be clear, this was not a "road test". This was a publicity stunt.
And I don't have a problem with you ranting, at least you made an effort to share your view.
They (Uber) already had trials in Pittsburgh. Presumably after the closed-track trials. This sound like a careful progression of testing and trials to me.
And every "Technology Demo" is of course a publicity stunt by definition.
I think you're badly underestimating the amount and nature of the data required. Think about space travel: after 60 years, the catastrophic failure rate is still about 1 in 200. This is due in part to the very low volume of flights over the decades. The Shuttle underwent what passed for extensive testing before its first "operational" flight, but was retired before it accumulated even as many hours as the average test flight program for a commercial airliner.
Before we run vehicles as potentially destructive as a fully loaded 18-wheeler autonomously on public roads, we should log tens of thousands of test hours under simulated real-world traffic conditions. That's expensive, hence Uber's willingness to roll the dice with other people's lives.
The burden is not on the public to show that an autonomous truck is unsafe; the burden is on the promoters to show that it is safe. That burden is not even close to being met, as the article acknowledges when they say the technology won't be ready for another 5 years. This was a stunt to help Uber maintain its valuation, and the DoT should've been looking out more for the public.
Sure, for the person performing it. In this case, people were put at risk who had no opportunity to opt out. And just to clarify, the problem isn't the autonomous driving demonstration---it's the lack of a backup in the form of a human driver in a position to take control if something went wrong. That's a shocking level of complacency given the inherent risks of a fully loaded 18-wheeler truck and the immaturity of self-driving technology.
Actually, I'd love to hear about ANY autonomous vehicle technology that can safely traverse winter roads.