The first cars bold enough to drive themselves
- Thread starter JournalBot
- Start date
Radio controlled cars really have nothing whatsoever to do with autonony, no matter how long ago they were first experimented with.
Upvote
78
(101
/
-23)
I'm going to chalk this up as another example of neat ideas that people want (less stressful driving!) but no one is willing to pay for. The technology was clearly there - we just need some cooperation to lay the groundwork (sometimes literally) and funds to support the system.
That didn't happen and so we have many parties trying to make independent systems with vastly different takes on safety. Humans once again getting in the way of themselves.
That didn't happen and so we have many parties trying to make independent systems with vastly different takes on safety. Humans once again getting in the way of themselves.
Upvote
18
(24
/
-6)
I would not say that. They both basically need to have the machinery in place to replace the human body - that is, in some way substitute for arms that control the steering and feet to step on the pedals. The radio-controlled car receives signals, the rest must be done by the car. In the autonomous car the signals are generated by a computer (simplistically speaking) instead of coming from the outside. But aside from that difference, they have many engineering challenges in common that separate them from a conventional car.
Upvote
50
(69
/
-19)
i think they're using "detroit" as a synonym here for "car industry."
Upvote
31
(34
/
-3)
"1928. Francill, who styled himself “America’s Radio Wizard,” demonstrated how radio control could move Ford automobiles without a driver" in the sense of "without a driver inside the car" and as a wise teacher of mine used to say, slowly slowly...
Upvote
24
(24
/
0)
Now we just have parking lots of autonomous vehicles honking at each other…
Let’s hope they continue to evolve. The median driver is pretty bad and half of the rest of them are worse.
Let’s hope they continue to evolve. The median driver is pretty bad and half of the rest of them are worse.
Upvote
39
(39
/
0)
Which is an absurdly American centric take.
Americans didn't invent cars, nor do they own the concept.
Upvote
62
(73
/
-11)
I know it's basically required for websites to make money, but please stop making click-bait headlines. I've been an Ars reader since the sites inception, and I beg you to do better than this.
Last edited:
Upvote
40
(52
/
-12)
It was not the point. The idea was that a sitting driver was no longer necessary for the car to move.
Upvote
4
(23
/
-19)
that's the idea, sure. but that's a different idea than autonomy.
Upvote
32
(35
/
-3)
I think they were a first step towards building cars that can be driving autonomously by building control systems that aren't actuated by humans. Perhaps not a necessary first step, but they were the actual first step since autonomous, wire guided cars weren't demonstrated until 30 years after the remote control cars.
Upvote
23
(29
/
-6)
No? In the context of these vehicles, they're only proof that the very much human driver did not have to be physically in the vehicle. It's neat stuff, but it doesn't at all justify this sentence.
Upvote
26
(32
/
-6)
Darpa had a self driving vehicle program as well back in the 80s. There were two prototypes I'm aware of. Navlab (a semi-autonomous van that carried its computers and generators) and the ALV. A fully autonomous unmanned multiwheeled vehicle. By the mid 90s they had driven Navlab 5 over 2800 miles across the US fully autonomously, with only 50 miles human driven. Fascinating stuff.
Last edited:
Upvote
35
(35
/
0)
Nice article, but 2 important corrections:
- the title "The first cars bold enough to drive themselves" is totally wrong. The 1904 cars in the article were driven by remote operators, not by the cars themselves!!! I got bait, so I guess the title worked. But I expect more from ars.
- "By 1904, he was using the Telekino to direct a small, three-wheeled vehicle from nearly 100 feet away. It was the earliest recorded instance of a vehicle being controlled by radio" seems to forget the Nikola Tesla remotely controlled boat he demonstrated in 1898.
Last edited:
Upvote
69
(69
/
0)
That's an excellent analogy, and why this article feels dishonest/clickbaity
Upvote
28
(29
/
-1)
The Mechanical Turk was not a robot… and in the same way, this entire article suffers from a fundamental misunderstanding of the word “autonomous” in this context.
The majority of the article is a discussion of “remote control” systems, which are not autonomous.
The author is confused that “the car has no driver inside of it” is not the same thing as “autonomous” but unfortunately they’re entirely different concepts.
The majority of the article is a discussion of “remote control” systems, which are not autonomous.
The author is confused that “the car has no driver inside of it” is not the same thing as “autonomous” but unfortunately they’re entirely different concepts.
Upvote
46
(46
/
0)
"Inexplicably?" How the hell am I supposed to get my fresh Sunday-morning-drive OJ otherwise? From a carton? Like the riff raff over in the Steering Wheel Required lane? Pfft.
Upvote
28
(28
/
0)
Ninja'd, but here's a link to Tesla's patent on his "Method of and apparatus for controlling mechanism of moving vessels or vehicles" (Google Patents)
IANAL, but on a quick read, the claims seem broad enough to cover a boat or a car.
Upvote
24
(24
/
0)
Upvote
0
(7
/
-7)
Quevedo's "chess-playing machine" has nothing to do with von Kempelen's Mechanical Turk. Quevedo's "ajedricista" was a genuine mechanical computer, with hardware that could play a king-rook-vs-king endgame (always as white - the side with a rook). Of course, that's not chess, that's a tiny subset of chess.
Even for this extremely limited task, the algorithm was chosen for simplicity, not effectiveness. The machine frequently lost due to the fifty-move rule.
Still, this was pretty impressive for the time, and it wouldn't have been crazy to see it as a proof of principle. In principle you could build a bigger, more complicated machine that could handle a bigger, more complicated subset of chess, and do this iteratively until you had a machine that could play an actual game of chess. (Unfortunately, that has proven over-optimistic. To design the hardware, you need to have an algorithm for the hardware to implement. That is possible if you're OK with playing chess very badly, but in the search for best play we've largely given up on asking a human mind to formulate a single best algorithm. We make self-learning machines instead. And, of course, we implement algorithms in software, not hardware, whenever possible.)
Upvote
31
(31
/
0)
YetAnotherAnonymousAppellation
Ars Praefectus
By the author's definition, an animal-drawn cart following a walking human is an "autonomous vehicle." They were around for millennia before his "first" example.
The best thing that can be said about this article is that it is spurring discussion. The worst thing is that it is garbage.
The best thing that can be said about this article is that it is spurring discussion. The worst thing is that it is garbage.
Upvote
40
(40
/
0)
The content of the article is neat, but the context and headline are bad. Autonomy and wireless are distinct and unrelated things. Nobody is looking at remote vehicle operation as a solution to autonomous operation—only as a backup for the lack of onboard controls during the times the autonomous system gets stuck.
Radio operated vehicles have been used for many decades in stunts for movies and tv shows, and that’s really all that most of these examples are. Framing this as being related to autonomy is dishonest clickbait.
Speaking of autonomy, though (Ars brought it into an unrelated subject, not me): what’s going on with the authors of the piece you pulled due to it containing LLM hallucinations?
Radio operated vehicles have been used for many decades in stunts for movies and tv shows, and that’s really all that most of these examples are. Framing this as being related to autonomy is dishonest clickbait.
Speaking of autonomy, though (Ars brought it into an unrelated subject, not me): what’s going on with the authors of the piece you pulled due to it containing LLM hallucinations?
Upvote
18
(20
/
-2)
Didn’t Nikola Tesla invent radio controlled vehicles nearly a decade earlier? That happened in NYC.
Upvote
12
(12
/
0)
Upvote
12
(12
/
0)
Even sentences like "This was a place where invention was not merely encouraged, but expected." and "What he wrought was brilliant, if crude—and certainly ahead of its time."?
Upvote
-13
(3
/
-16)
The DARPA competition was pretty exciting to be honest. We need more scientific "art of the possible" challenges.
Upvote
8
(8
/
0)
Interesting to dig farther back into the human desire for automation. But this has to be one of the most poorly written ars articles I have seen.
The first bit that rankles me is: "But he had shown that a machine could be guided by signals. It would be more than a century before that notion would reach fruition. But that doesn’t mean others didn’t try." Sloppy at best. Sure, I get it, sort of...largely the article is about autonomous vehicles. But radio control came to useful fruition far, far earlier that a century after 1904.
Next bit that really jumped out at me: "These vehicles may seem like novelties today, but they’re early proof that the automobile can be guided by something other than humans." I'm sorry, but a vehicle being remotely controlled is still demonstrably being guided by a human. Something like "...guided by something other than human control of mechanical devices in the vehicle" would get closer. But really, as written, just sloppy.
The caption to the Firebird concept car ("The Firebird II concept from 1956 could drive itself on special roads. Credit: General Motors") broaches one of what seems to me (and I am not in the field so it's my admittedly limited knowledge) one of the biggest challenges facing level 5 autonomous driving: making it happen with no modification or accommodation to the existing driving infrastructure (i.e., roads, markings, traffic signals etc.). But maybe that is fodder for a future article. (It also reminded me of the bit in Isaacson's biography of Musk where Tesla engineers somehow manage to get a municipality to repair the road markings so Musk will quit berating about his car's FSD failing on his daily commute.)
Also, very impressed that Dickmanns was able to produce some useful driving autonomy with 1980's level compute and optical capabilities.
Overall, interesting topic and historical content. Sad that the soppy writing rankled me so much that I really failed to enjoy the article as much as I might have.
Finally, a story about my stepfather (Ph.D. in physics and EE) related to the topic. Decades back (1970's?) he was involved in a DOT sponsored project to develop an automated braking system for cars. (The kind of thing that now exists to stop the car before you hit the kid chasing a ball into the street.) His part of this was developing a millimeter wave radar system. The team built and fitted the system to a car and he was given a ride...straight at a brick wall. It worked but scared him sh**less. But the system was not practical because it stomped on the brakes due to rain and snow and...
The first bit that rankles me is: "But he had shown that a machine could be guided by signals. It would be more than a century before that notion would reach fruition. But that doesn’t mean others didn’t try." Sloppy at best. Sure, I get it, sort of...largely the article is about autonomous vehicles. But radio control came to useful fruition far, far earlier that a century after 1904.
Next bit that really jumped out at me: "These vehicles may seem like novelties today, but they’re early proof that the automobile can be guided by something other than humans." I'm sorry, but a vehicle being remotely controlled is still demonstrably being guided by a human. Something like "...guided by something other than human control of mechanical devices in the vehicle" would get closer. But really, as written, just sloppy.
The caption to the Firebird concept car ("The Firebird II concept from 1956 could drive itself on special roads. Credit: General Motors") broaches one of what seems to me (and I am not in the field so it's my admittedly limited knowledge) one of the biggest challenges facing level 5 autonomous driving: making it happen with no modification or accommodation to the existing driving infrastructure (i.e., roads, markings, traffic signals etc.). But maybe that is fodder for a future article. (It also reminded me of the bit in Isaacson's biography of Musk where Tesla engineers somehow manage to get a municipality to repair the road markings so Musk will quit berating about his car's FSD failing on his daily commute.)
Also, very impressed that Dickmanns was able to produce some useful driving autonomy with 1980's level compute and optical capabilities.
Overall, interesting topic and historical content. Sad that the soppy writing rankled me so much that I really failed to enjoy the article as much as I might have.
Finally, a story about my stepfather (Ph.D. in physics and EE) related to the topic. Decades back (1970's?) he was involved in a DOT sponsored project to develop an automated braking system for cars. (The kind of thing that now exists to stop the car before you hit the kid chasing a ball into the street.) His part of this was developing a millimeter wave radar system. The team built and fitted the system to a car and he was given a ride...straight at a brick wall. It worked but scared him sh**less. But the system was not practical because it stomped on the brakes due to rain and snow and...
Upvote
23
(23
/
0)
Well now you're getting into the distinction between entirely written by AI, versus lightly touched up by AI and all of the messy points on the spectrum between the two which is different thing than what you asked about initially. I hope we never get the former from ars, but the latter is probably unavoidable and probably unoffensive.
By unavoidable, I mean that many cms/word processing apps will probably be labeling any kind of grammar/spell checker as AI shortly due to leadership FOMO, and as such any article that is so much as spell checked could credibly be described as having some level of AI involvement, assuming the spell checker changed something.
Upvote
3
(3
/
0)
I suppose it's how you define "vehicle," but Tesla demonstrated a remote controlled model boat in 1898.
Upvote
10
(10
/
0)