3 ms·
My former job was systems programming on precisely this sort of project. We ran embedded ARM based Linux systems in cars, motorcycles and over the road trucks t
by 16s 14y ago
My former job was systems programming on precisely this sort of project. We ran embedded ARM based Linux systems in cars, motorcycles and over the road trucks to gather research data from the vehicle CAN. The volume of data and variables collected was immense. Lateral acceleration (swerve), turn signals, seat-belts engaged, gear position, vehicle speed, individual wheel speed, engine rpm, GPS coordinates, steering wheel angel, etc, etc, etc. We were collecting petabytes of data for research. We had cameras on the vehicles as well so that we could correlate all of the collected data with what the driver was doing, where they were looking, etc. I doubt the cameras will exist in the mandated black-box systems, but who knows.
I only say this because there are tons of research projects in this area right now and likely will be for many years to come.
- YokoZar 14y agoIf we had steering wheel angels we wouldn't need black boxes ;)
- beagle3 14y agoHow often did you sample most of those variables? I don't doubt the petabytes of data, but I'd be surprised if data that wasn't visual, all included, was more than a couple of terabytes (if even that .. a terabyte of textual data is a ridiculously large amount). Counting the gzipped data is the only meaningful thing here, btw - because a verbose XML document with repeating data "<absoluteLocation>17,34,25</absoluteLocation><absoluteLocation remark="notMoved">17,34,25</absoluteLocation>" is not more informative than "l=17,35,25", despite being 10 times longer. Video ... that's a completely different story.