search

schollz / find3

High-precision indoor positioning framework, version 3.
https://www.internalpositioning.com/doc
MIT License
4.63k stars 363 forks source link

Completely unstable... #194

Closed StefanoGiu closed 4 years ago

StefanoGiu commented 4 years ago

Completely unstable... The more I learn data for one room, the more other rooms accuracy drops down. When I start then in tracking mode, the room guessing is not accurate and the scan interval is very high (last seen is more than 5 minutes). As it is, it's completely useless...

schollz commented 4 years ago

Why?

StefanoGiu commented 4 years ago

Details are above in the text

schollz commented 4 years ago

@StefanoGiu Please add more info

nash-stokes commented 4 years ago

@StefanoGiu I had better accuracy when I had equal numbers of data points for each room. I'm not sure about your scanning time issue though, as you should be setting this yourself from the command line.

@schollz One issue I am having is that at a certain point, getting more data does not improve a room's accuracy at all, and in some cases even decreases it. Do you have any suggestions on the amount of data to collect for each room?

schollz commented 4 years ago

@nash-stokes Can you make a new issue with your question? Judging by the Github history of @StefanoGiu they only create issues in projects. I suspect the issue is not with the projects.

StefanoGiu commented 4 years ago

Why you closed this? Do you feel I'm trying to denigrate your project?? I was really happy when I was reading the way your project is supposed to work, but unfortunately, I had a very bad experience in tuning the learning up... If you want to collect my feedback, I'm more than welcome to help you understanding the issues I'm facing... otherwise, feel free to ignore it ;)