Safety systems are there to be gamed...

Technology in driving is becoming more dominant...
User avatar
Horse
Posts: 1788
Joined: Mon Sep 28, 2015 9:20 am

Re: Safety systems are there to be gamed...

Postby Horse » Thu Jan 25, 2018 5:07 pm

jont- wrote:
Horse wrote:
GTR1400MAN wrote:And another.

Told by the system SEVEN times to retake control. Dozed off because of reduced workload?

http://money.cnn.com/2017/06/20/technology/tesla-autopilot-fatal-crash-warnings/index.html


"Hands on the wheel for 25 seconds in 35 minutes". That's not too shabby from the car's perspective!

But they can't spot a fire engine parked in their lane:
https://www.theregister.co.uk/2018/01/2 ... lot_crash/


It probably did 'spot it'. As several of the comments on the article linked earlier say: the car may have been doing 65mph at some point, but that probably wasn't a 65mph impact - the autonomous braking had probably done a fair job of slowing the car if not actually stopping it.



https://electrek.co/2018/01/22/tesla-mo ... ire-truck/
My own views. For better or worse :)

User avatar
Horse
Posts: 1788
Joined: Mon Sep 28, 2015 9:20 am

Re: Safety systems are there to be gamed...

Postby Horse » Thu Jan 25, 2018 7:16 pm

https://www.theverge.com/2018/1/4/16849 ... exus-lidar

The biggest improvement over previous versions of Toyota’s autonomous research vehicles is the ability to “see” farther in every direction. Thanks to four long-range LIDAR sensors attached to its roof, manufactured by a Portola Valley, California-based startup called Luminar, the vehicle now has a 200-meter range around a 360-degree perimeter, which Toyota argues makes it “one of the most perceptive automated driving test cars on the road.” By comparison, Velodyne’s powerful LIDAR, the HDL-64E, has a 120-meter range, while its most popular LIDAR, the VLP-16 Puck, has a range of 100 meters.

Of course, seeing 200m isn't the same as reacting to what it 'sees' . . .
My own views. For better or worse :)

Pyrolol
Posts: 68
Joined: Sun Oct 04, 2015 1:52 am
Location: San Francisco

Re: Safety systems are there to be gamed...

Postby Pyrolol » Fri Jan 26, 2018 1:39 am

LIDAR is much more capable than what Tesla do. Tesla attempt to figure out distance from a bunch of cheap sensors like cameras, LIDAR can actually measure distance and identify things that blend into the background for example.

User avatar
jont-
Posts: 924
Joined: Mon Sep 28, 2015 7:12 am
Location: Flatlands

Re: Safety systems are there to be gamed...

Postby jont- » Fri Jan 26, 2018 7:11 am

Pyrolol wrote:LIDAR is much more capable than what Tesla do. Tesla attempt to figure out distance from a bunch of cheap sensors like cameras, LIDAR can actually measure distance and identify things that blend into the background for example.

That's strange, I'm sure you and I have managed ok with just a pair of mk1 eyeballs.

User avatar
Horse
Posts: 1788
Joined: Mon Sep 28, 2015 9:20 am

Re: Safety systems are there to be gamed...

Postby Horse » Fri Jan 26, 2018 9:31 am

jont- wrote:That's strange, I'm sure you and I have managed ok with just a pair of mk1 eyeballs.


You think you do. See the cognitive bias thread :)

Of course, the lidar won't blink, won't need to refocus from near to far, won't have narrow foveal vision or suffer from temporary loss of vision during saccades :)
My own views. For better or worse :)

User avatar
jont-
Posts: 924
Joined: Mon Sep 28, 2015 7:12 am
Location: Flatlands

Re: Safety systems are there to be gamed...

Postby jont- » Fri Jan 26, 2018 9:33 am

Horse wrote:Of course, the lidar won't blink, won't need to refocus from near to far, won't have narrow foveal vision or suffer from temporary loss of vision during saccades :)

But won't work when it rains :lol:
https://cleantechnica.com/2016/07/29/te ... dar-right/

User avatar
Horse
Posts: 1788
Joined: Mon Sep 28, 2015 9:20 am

Re: Safety systems are there to be gamed...

Postby Horse » Fri Jan 26, 2018 11:28 am

jont- wrote:
Horse wrote:Of course, the lidar won't blink, won't need to refocus from near to far, won't have narrow foveal vision or suffer from temporary loss of vision during saccades :)

But won't work when it rains :lol:
https://cleantechnica.com/2016/07/29/te ... dar-right/


Interesting summary, thanks. Hmmm . . . loss of vision in rain and fog, so slow down? Who says AVs won't drive like people? ;)

Regardless of whether LIDAR is required or not, no solution based on a single-sensor set or even a dual-sensor set is likely to be viable. Each sensor type has strengths and weaknesses and amalgamating a single representation of reality from multiple sensors is required in order to avoid false positives and also false negatives.
My own views. For better or worse :)

User avatar
Horse
Posts: 1788
Joined: Mon Sep 28, 2015 9:20 am

Re: Safety systems are there to be gamed...

Postby Horse » Fri Jan 26, 2018 11:31 am

jont- wrote:
Horse wrote:Of course, the lidar won't blink, won't need to refocus from near to far, won't have narrow foveal vision or suffer from temporary loss of vision during saccades :)

But won't work when it rains :lol:
https://cleantechnica.com/2016/07/29/te ... dar-right/


'Technology' exists, has just been waiting for AVs and Lidar to come along ;)

Image

https://carpyscaferacers.com/insane-slicer-turbo/

Image
My own views. For better or worse :)

User avatar
Horse
Posts: 1788
Joined: Mon Sep 28, 2015 9:20 am

Re: Safety systems are there to be gamed...

Postby Horse » Fri Jan 26, 2018 12:09 pm

And the latest (!) entry in the 'nothing new about that!' category is . . .

. . . the programmable vehicle :)



http://www.leonardodavincisinventions.c ... incis-car/

Image
My own views. For better or worse :)

User avatar
Horse
Posts: 1788
Joined: Mon Sep 28, 2015 9:20 am

Re: Safety systems are there to be gamed...

Postby Horse » Fri Jan 26, 2018 12:47 pm

https://www.frontiersin.org/articles/10 ... 00122/full

Self-driving cars are posing a new challenge to our ethics. By using algorithms to make decisions in situations where harming humans is possible, probable, or even unavoidable, a self-driving car's ethical behavior comes pre-defined. Ad hoc decisions are made in milliseconds, but can be based on extensive research and debates. The same algorithms are also likely to be used in millions of cars at a time, increasing the impact of any inherent biases, and increasing the importance of getting it right. Previous research has shown that moral judgment and behavior are highly context-dependent, and comprehensive and nuanced models of the underlying cognitive processes are out of reach to date. Models of ethics for self-driving cars should thus aim to match human decisions made in the same context. We employed immersive virtual reality to assess ethical behavior in simulated road traffic scenarios, and used the collected data to train and evaluate a range of decision models. In the study, participants controlled a virtual car and had to choose which of two given obstacles they would sacrifice in order to spare the other. We randomly sampled obstacles from a variety of inanimate objects, animals and humans. Our model comparison shows that simple models based on one-dimensional value-of-life scales are suited to describe human ethical behavior in these situations. Furthermore, we examined the influence of severe time pressure on the decision-making process. We found that it decreases consistency in the decision patterns, thus providing an argument for algorithmic decision-making in road traffic. This study demonstrates the suitability of virtual reality for the assessment of ethical behavior in humans, delivering consistent results across subjects, while closely matching the experimental settings to the real world scenarios in question.
My own views. For better or worse :)


Return to “Technology”

Who is online

Users browsing this forum: No registered users and 1 guest