Technology & Inovations
The Issues with Today's Technology
The inherent problem with today’s video processing is that the camera captures a 3-dimensional environment onto a 2-dimensional receptor and then requires a processor to interpret that 2-dimensional image back into a 3D image. This is nearly an impossible task. As an example, if the camera sees a bird twenty feet away, it cannot easily distinguish that from an airplane two miles away without more information. What’s more problematic, if you paint a mural of a road on a brick wall extending into the distance, a self-driving car will crash into it thinking it’s simply driving down a road. Our innovation technology basically inputs the image as 3-dimensional, by creatively leveraging natural optical and geometric capabilities inherent to a camera. It then resolves the third dimension (distance from camera to object), and saves each object with all of its unique properties including distance, location in space, velocity vector, size dimensions, among others.
How We're Different
The design does not need machine learning to parameterize an object. It therefore uses extraordinarily less RAM. (It currently uses 2 Megs vs. an average of 5 Gigs for deep learning). The entire system uses exponentially less power. Currently, it uses 6 watts vs. 1.2kw (Nvidia B200). All of the algorithms reside in programmable logic (hardware) and therefore run at lightning-fast speeds (nanoseconds vs. milliseconds). Some of the video processing requires multiple frames of video and it is therefore the limiting factor in its speed. However, there are current plans which entail creating smaller frame sizes in the camera that will greatly accelerate the frame rate and nearly eliminate this limitation.
One of the applications of this technology is for self-driving cars, and because its response time is so quick and its ability to determine the distance between it and the car in front of it, we have a motivational mantra in our business of “80-8”. It refers to “eighty miles an hour, 8 feet apart.” That’s a goal at Kemper Engineering. Imagine traffic jams being a thing of the past. If every car could move at 80 miles an hour at a relatively close distance apart, the cars would quickly move along, being able to slow down or speed up with zero reaction time. Or imagine being in a left turn lane with twenty cars ahead of you and when the arrow turns green, all of the cars would start moving at the same time, and rather than the light turning red before even reaching the intersection, you’d be going through the intersection with plenty of time to spare. Traffic congestion would be a thing of the past.
Our Patent
At the time of writing this paper, we have already secured a provisional patent and have completed the paperwork for a non-provisional patent on our design and have submitted it to our attorney and subsequently he will be submitting it to the United States Patent and Trademark Office and thereby we will be receiving patent pending status.
