Question on live demo

[GpaDoctor]

I am quite new to this. I would like to try it out and run the live demo. Can you provide more guidance on how do I set it up for my phone and watch to provide imu readings and have it run in Unity?

[Yefancy]

I have successfully run the live demo with my own data collector. The only thing I can suggest is that the acc you get from your devices should not include gravitity and the quat you get should be ordered in wxyz. But the actual behaviour may not as good as the teaser. I suppose there is a bug related to the datasets process #3

[jql123456]

Hi, I noticed that you said you implemented a real-time demo. Can you tell me what type of IMU you are using and the sampling rate of the IMU? How is the inference performance, assuming that only three IMUs are used? @Yefancy

[Yefancy]

Hi, I noticed that you said you implemented a real-time demo. Can you tell me what type of IMU you are using and the sampling rate of the IMU? How is the inference performance, assuming that only three IMUs are used? @Yefancy

I collected the imu data from my iPhone and iWatch. sampling rate is 30Hz. I cannot say the performance is bad, but its not that really ideal as my expectation. I only use two of them, I think the head one is not tahy really necessary. bc from the teaser, he only used two of them aswell.

[jql123456]

Hi, I noticed that you said you implemented a real-time demo. Can you tell me what type of IMU you are using and the sampling rate of the IMU? How is the inference performance, assuming that only three IMUs are used? @Yefancy

I collected the imu data from my iPhone and iWatch. sampling rate is 30Hz. I cannot say the performance is bad, but its not that really ideal as my expectation. I only use two of them, I think the head one is not tahy really necessary. bc from the teaser, he only used two of them aswell.

Hello, I have another question. When you use two devices, is the estimation of trans accurate? In my own experiments, the estimation of trans is not very ideal.

[Yefancy]

Hi, I noticed that you said you implemented a real-time demo. Can you tell me what type of IMU you are using and the sampling rate of the IMU? How is the inference performance, assuming that only three IMUs are used? @Yefancy

I collected the imu data from my iPhone and iWatch. sampling rate is 30Hz. I cannot say the performance is bad, but its not that really ideal as my expectation. I only use two of them, I think the head one is not tahy really necessary. bc from the teaser, he only used two of them aswell.

Hello, I have another question. When you use two devices, is the estimation of trans accurate? In my own experiments, the estimation of trans is not very ideal.

not very ideal but reasonable

[VascoXu]

Hello! We are using IMUs in iPhone and Apple Watch. You can access them in the Core Motion API. We have a basic implementation of an iPhone and Watch app that you can install onto your devices here: https://github.com/SPICExLAB/SensorTracker