What can AI tell us about patent US10,000,000?

August 8th 2018   Self driving cars are coming! 

In fact, they are already here, with some premium European cars and also Tesla cars offering a degree of self-driving. Meanwhile online 'taxi' companies such as Uber are pushing hard in this direction, because if they can pull the cost of the driver out out of their fares, they can charge less and grow volumes, which in turn could decrease ownership of private cars, which in turn could increase demand for their services.

Given the highly variable nature of human driving, it is easy to believe that self-driving cars could be safer in many cases. But this does require cars with the ability to sense the world around them. Currently this is done using cameras and radar - and these scanning technologies are increasingly found in cars with automatic braking and 'active' cruise control. The next step, however, will likely require 'LIDAR (Light Imaging Detection And Ranging also known as LADAR), which uses the reflections from pulses of light to create a three dimensional view of the world. LIDAR is already used for surveying applications, and will become a core technology for self driving cars.


Given this potential demand, we would expect Lidar to be under significant development, and this is exactly what we see, with the number of published patents referring to this technology on the rise:








Given this growth, it is not too surprising that when the USPTO came to assign the milestone number "US10,000,000" to its 10 millionth granted utility patent, it applied it to a a patent, filed by Raytheon, that claimed an improvement in Lidar:  Coherent LADAR using intra-pixel quadrature detection". This patent has a priority date of March 2015, and the specific improvement is aimed at improving accuracy by reducing the errors caused by interference between light waves.

Screenshot 2018-08-08 at 11.06.05.png


What can the innovative AI patent search engine Ambercite tell us about this patent?


Ambercite has been developed to predict the most similar patents to one or more starting patents, and by looking at these similar patents, we can learn a lot about a patent and the field of technology.

Screenshot 2018-08-08 at 11.16.51.png

Running a patent search simply requires relevant patent numbers, using the default settings in this case:












In this case, given that Lidar could be such an important technology, a search for the most similar 50 patents may not be enough.

Screenshot 2018-08-08 at 11.20.00.png

There is an option to request between 5 and 2000 patents - if we choose to request 500 patents, as shown to right, the prediction engine will return 294 similar patents.

These can easily be exported to Excel for further analysis. 










Among other things, the results will tell you:

  • The most similar patent (according to our AI algorithms) was US9871339, which actually had a later priority date of December 2016, but was granted earlier.







Screenshot 2018-08-08 at 11.21.59.png
  • Lockheed Corporation was the leading owners of similar patents, with 18 patent families, followed by Raytheon (14), Digilens (13) and Daylight solutions Inc (13)

  • US10,000,000 is listed with 5 known backward citations and 2 known forward citations. In addition, Ambercite was able to identify 223 other similar patent families. This is a big benefit of Ambercite, in that it will instantly find a much much greater range of relevant patents than conventional searching.

  • Among the highest ranked of the similar but 'unknown' (not cited) patents is US8179521, filed with a priority date of 2009 by ONERA for Measurement of speed or vibration characteristics using a LIDAR device with heterodyne detection, and which discloses similar subject matter, i.e the ccorrection of errors caused by interference in Lidar waves.

  • The five most recent and similar patents are all filed by Luminar Tech Inc - whose website demonstrates the use of Lidar for cars.


Would you like to try Ambercite for yourself?

There are may other insights to be gained from Ambercite patent analysis, both on this patent, and on any other patent or patents you look at.

Ambercite now offers a free trial version (some restrictions apply) if you would like to see this capability for yourself. To register online and start using the trial version, please click the button below, and register in the normal manner. 


Mike Lloyd