PencilGeek
Well-Known Member
- Joined
- Feb 17, 2017
- Threads
- 3
- Messages
- 225
- Reaction score
- 195
- Location
- Northern California
- First Name
- Robert
- Vehicle(s)
- 2017 GT350R HR278, 2008 BMW M3
The vBox filters are simple averaging n-samples. There's probably more sophisticated filters that can be applied both in hardware and software. I forgot to mention that sampling rate has an effect as well. The slower the sampling rate, the more pronounced the spikes appear in the data. That's not saying they don't appear in higher sampling rates as well, but with higher sampling rates it's much easier to apply a best-fit algorithm to the data set. If I found the correct IMU on the GT350 CAN bus, they sample at 50 Hz. I haven't got back to that project recently, so I might have to update that figure when I finalize my findings.Its too bad they don't incorporate these filters you spoke of though into the software; would love to know how they work. Would be really cool to have a gauge that couldn't be contaminated with false data points so easily.
Fairchild has a really nice set of IMU's that they've put on the market in the last year or so. Internally, they sample at an incredible 2k sampling rate, but they limit the output data rate (ODR) to a very low 75 Hz. So by oversampling, they apply some really nice averaging filters onto the data. Their target market is drones and the like where you need high accuracy, but don't have the bandwidth to transmit over the air. It would be nice if they had a similar set of sensors that had a 400 Hz ODR because it would be ideal for virtual reality devices. But I digress...
Sponsored