Solid as a rock.¯\_(ツ)_/¯
That's not a rack!
This is a rack!
This. It's testing/calibration equipment.Lidar / Radar etc for vision validation.
https://www.teslarati.com/tesla-lidar-sensors-spotted-testing-palo-alto/
Agreed! Awesome stuff, my CNC machine is made (mostly) out of it.Those crossbars look like 80/20 extrusions.
Just a tip on filming while driving!Looks to be modular post/channel beam design. But what are the round things on each side?
What sort of sensors? Havent you all seen the GPS sensors used on excavation equipment to ensure they are in the right place? Looks like that much more than any LIDAR, being a flat plate, a normal GPS configuration.
Those appear to be identical to the "high def" GPS antennas that we use with work. They are basically survey quality GPS, and are about 100 times more sensitive than what you have in your car. This gives you additional satellites to use in your position solution and thus a more higher resolution position. You can do (and it is now super common) to do DGPS on a single antenna. These are guaranteed to allow for SBAS (Space Based Augmentation System) which is nothing more than another sat that broadcasts correction factors for each normal GPS sat in timing and ionospheric interference, and sat health updates. The US SBAS system is called WAAS, in Europe it's called EGNOS. Japan, China and India have their own, and they all use the same format and frees, so one receiver can use it anywhere in the world. SBAS sats use a geo-statoinary orbit. Dual antennas are an indication of dual GPS systems. This is normally done for verification of signals. If the vehicle system reports 9 sats, then you better see at least 9 sats on a better antenna and receiver. If you receive a sat (PRN12 as an example) in the vehicle and the validation system DOESN'T see PRN12, then you know you have an issue. Offsetting the antennas has one of two or three purposes (or all). First is no interference between them. Second, differing reflection paths (from the vehicle and surrounds) can be compared. But most likely, it is accurate enough to show you which lane the vehicle is in, map that, and compare the video and GPS data in post-processing to tell you how the vehicle positioned itself within the lane. And in areas where the lane markings are less than perfect, this gives you GRADE A data and video to feed into the Tesla AI cluster to teach FSD how to handle all sorts of situations.Looks like differential gps
Well laid out and said. Kudos to you, sir.Those appear to be identical to the "high def" GPS antennas that we use with work. They are basically survey quality GPS, and are about 100 times more sensitive than what you have in your car. This gives you additional satellites to use in your position solution and thus a more higher resolution position. You can do (and it is now super common) to do DGPS on a single antenna. These are guaranteed to allow for SBAS (Space Based Augmentation System) which is nothing more than another sat that broadcasts correction factors for each normal GPS sat in timing and ionospheric interference, and sat health updates. The US SBAS system is called WAAS, in Europe it's called EGNOS. Japan, China and India have their own, and they all use the same format and frees, so one receiver can use it anywhere in the world. SBAS sats use a geo-statoinary orbit. Dual antennas are an indication of dual GPS systems. This is normally done for verification of signals. If the vehicle system reports 9 sats, then you better see at least 9 sats on a better antenna and receiver. If you receive a sat (PRN12 as an example) in the vehicle and the validation system DOESN'T see PRN12, then you know you have an issue. Offsetting the antennas has one of two or three purposes (or all). First is no interference between them. Second, differing reflection paths (from the vehicle and surrounds) can be compared. But most likely, it is accurate enough to show you which lane the vehicle is in, map that, and compare the video and GPS data in post-processing to tell you how the vehicle positioned itself within the lane. And in areas where the lane markings are less than perfect, this gives you GRADE A data and video to feed into the Tesla AI cluster to teach FSD how to handle all sorts of situations.
If that is the direction they are going, then the size and width and video system (think additional lower front cam and a higher standard rear cam) of the CT, along with the 2-5 centimeter accurate GPS data to train FSD AI. I'd put my bets on the latter possibility.