Slashdot Mirror
Vigilante Engineer Stops Waymo From Patenting Key Lidar Technology (arstechnica.com)
An anonymous reader quotes a report from Ars Technica: A lone engineer has succeeded in doing what Uber's top lawyers and expert witnesses could not -- overturning most of a foundational patent covering arch-rival Waymo's lidar laser ranging devices. Following a surprise left-field complaint by Eric Swildens, the US Patent and Trademark Office (USPTO) has rejected all but three of 56 claims in Waymo's 936 patent, named for the last three digits of its serial number. The USPTO found that some claims replicated technology described in an earlier patent from lidar vendor Velodyne, while another claim was simply "impossible" and "magic." The 936 patent played a key role in last year's epic intellectual property lawsuit with Uber. In December 2016, a Waymo engineer was inadvertently copied on an email from one of its suppliers to Uber, showing a lidar circuit design that looked almost identical to one shown in the 936 patent.
The patent describes how a laser diode can be configured to emit pulses of laser light using a circuit that includes an inductor and a gallium nitride transistor. That chance discovery helped spark a lawsuit in which Waymo accused Uber of patent infringement and of using lidar secrets supposedly stolen by engineer Anthony Levandowski. In August 2017, Uber agreed to redesign its Fuji lidar not to infringe the 936 patent. Then, in February 2018, Waymo settled the remaining trade secret theft allegations in exchange for Uber equity worth around $245 million and a commitment from Uber not to copy its technology. "This includes an agreement to ensure that any Waymo confidential information is not being incorporated in Uber hardware and software," said a Waymo spokesperson at the time. That redesign now seems to have been unnecessary, says Swildens, the engineer who asked the USPTO to take a closer look at 936. "Waymo's claim that Uber infringed the 936 patent was spurious, as all the claims in the patent that existed at the time of the lawsuit have been found to be invalid," he said. Uber told Ars that despite the ruling, it would not be redesigning its lidars yet again. Swildensj, an employee at a small cloud computing startup, reportedly "spent $6,000 of his own money to launch a formal challenge to 936," reports Ars. "In March, an examiner noted that a re-drawn diagram of Waymo's lidar firing circuit showed current passing along a wire between the circuit and the ground in two directions -- something generally deemed impossible. 'Patent owner's expert testimony is not convincing to show that the path even goes to ground in view of the magic ground wire, which shows current moving in two directions along a single wire,' noted the examiners dryly."
"As I investigated the 936 patent, it became clear it was invalid due to prior art for multiple reasons," Swildens told Ars. "I only filed the reexamination because I was absolutely sure the patent was invalid."
The patent describes how a laser diode can be configured to emit pulses of laser light using a circuit that includes an inductor and a gallium nitride transistor. That chance discovery helped spark a lawsuit in which Waymo accused Uber of patent infringement and of using lidar secrets supposedly stolen by engineer Anthony Levandowski. In August 2017, Uber agreed to redesign its Fuji lidar not to infringe the 936 patent. Then, in February 2018, Waymo settled the remaining trade secret theft allegations in exchange for Uber equity worth around $245 million and a commitment from Uber not to copy its technology. "This includes an agreement to ensure that any Waymo confidential information is not being incorporated in Uber hardware and software," said a Waymo spokesperson at the time. That redesign now seems to have been unnecessary, says Swildens, the engineer who asked the USPTO to take a closer look at 936. "Waymo's claim that Uber infringed the 936 patent was spurious, as all the claims in the patent that existed at the time of the lawsuit have been found to be invalid," he said. Uber told Ars that despite the ruling, it would not be redesigning its lidars yet again. Swildensj, an employee at a small cloud computing startup, reportedly "spent $6,000 of his own money to launch a formal challenge to 936," reports Ars. "In March, an examiner noted that a re-drawn diagram of Waymo's lidar firing circuit showed current passing along a wire between the circuit and the ground in two directions -- something generally deemed impossible. 'Patent owner's expert testimony is not convincing to show that the path even goes to ground in view of the magic ground wire, which shows current moving in two directions along a single wire,' noted the examiners dryly."
"As I investigated the 936 patent, it became clear it was invalid due to prior art for multiple reasons," Swildens told Ars. "I only filed the reexamination because I was absolutely sure the patent was invalid."
Is this the correct word for this?
If you must moderate, please moderate as irrelevent, not something bad, because I'm sure someone will find this interest
> current passing along a wire between the circuit and the ground in two directions
Now we can finally get rid of all those nasty high voltage wires and use only ground to transmit power everywhere.
How many years will patents delay self-driving cars? 5? 10? 30? How many preventable road deaths will happen in the intervening period?
I am not a health and safety fanatic, but there comes a time when you have to question whether the USPTO should wield as much power as it evidently does.
And yes, I am aware that the USPTO, on this one occasion, did reject a large number of spurious patents.
Like if the circuit oscillates. I didn't see it but they did mention an inductor and "pulses"...
Strange things are afoot at the Circle-K.
Adolph, is that you?
Patent owner's expert testimony is not convincing to show that the path even goes to ground in view of the magic ground wire, which shows current moving in two directions along a single wire,' noted the examiners dryly.
"We'd like to submit an amendment: ... moving in two directions along a single wire over the Internet."
"Oh, that makes much more sense now. Patent upheld. Sorry if we sounded dry about it earlier."
Maybe the other self driving car companies should write him a cheque - he saved them millions.
I should use this sig to advertise my book ISBN-13 : 978-1501515132.
Build a working prototype or GTFO.
Congress: Make it so.
How many patents are out there that are also bullshit magic that didn't get noticed ?
Not news I know, but there was someone who's job it was to investigate these patents, to understand them and look for prior art.
The fact that a third party had to commit $6000 and a huge amount of personal time comes down to the fact that the patent examiner clearly didn't do their job.
While the work of Swildens should be celebrated, we shouldn't lose sight of the failings of the USPTO which required it.
"The patent describes how a laser diode can be configured to emit pulses of laser light using a circuit that includes an inductor and a gallium nitride transistor."
Every time I read stuff like this, I can only think I spent my life doing it wrong. I should of been figuring out how to get patents on stuff everybody knew about and was using anyway. Wow when a laser diode goes forward biased it will drain the stored energy from an inductor SMH.
It looks like they are using the reverse recover charge in the diode. as an undergrad I though I discovered the effect in 1983, but soon learned that it was well known. (used it to trigger a krytron for a pockels cell driver). (one small diode and inductor to make a KV pulse).
People who do pulse power systems know about a wide variety of tricks.
If they are looking for a fast (ns), high current (~100A, 100V), switches, there are much better tricks these days. We just dusted off an old 1KV, 20A, 100ps rise / fall, 2ns wide pulser we built about 10 years ago (tiny circuit) along with it big brother - 6KV, 120A, 200ps pulser.
I think that the ability to file a request to have a patent reviewed is very new. And good. There are still restrictions, I think there is a short time limit for such reviews.
Previously, I think the only way to have a patent reviewed was to have the owner sue you, and you lose everything if the jury in E. Texas rules against you.
From the article:
Swildensj, an employee at a small cloud computing startup, reportedly "spent $6,000 of his own money to launch a formal challenge to 936,"
The USPTO found that some claims replicated technology described in an earlier patent from lidar vendor Velodyne, ...
Shouldn't it be the job of the patent office to check new patent applications against existing patents or publications, like ... when applied and not when challenged?