Comments (2)
I think 5 minutes - but yeah the accuracy should go up a little bit if you do spend more time. There are WiFi fluctuations that can be pretty rampant.
from find3.
I am reopening this issue rather than creating a new issues for basically the same title. If you would prefer new issue instead let me know.
I read https://www.internalpositioning.com/doc/tracking_your_phone.md under the Learn some Locations section that it is recommended to train location for 5 minutes per room.
Then I read on https://github.com/schollz/find3/blob/master/doc/tracking_your_computer.md For your tracking scans to work, you must go to each room and run the learning command for about 10 minutes.
Then I read on https://github.com/schollz/find3/blob/master/doc/passive_tracking.md Leave the device in the location for about 30 minutes to collect a good amount of fingerprints. This will allow you to collect about 15 pieces of sensor data per location.
So different documentation recommends 5 minutes, 10 minutes, and 30 minutes per room. Granted, the 30 minute recommendation appears to be for passive scanning rather than learning by phone. Perhaps it would be beneficial to change the documentation to match each other or at very least have the information regarding time to train for various methods in one spot.
Is it possible to bypass the rate that android scans naturally, to decrease the time it takes to learn each room? - Like what would happen if we scan every second for the wifi signals?
I will hint that training over longer periods of time rather than all in a quick setting seems to produce significantly better results. I.e. if you scan in 5 minutes, chances are someone hasn't opened or closed different doors. It seems to help to take readings in different environments.
Perhaps the most significant environment change is humidity. Train on a dry day, then train while its raining, then train again the day after it rains.
Perhaps on the app, instead of training for 5,10,30 minutes at a time, just have a button that takes one reading per press. Tell the user to train each room twice a day, once in morning, once at night, over the course a week. That's 14 well difference readings that take into account doors opening and closing, and different humidity levels. All of which would not be accounted for if you just train in one setting.
from find3.
Related Issues (20)
- server crashing on own server install HOT 3
- [Docker] No EntryPoint defined in Dockerfile HOT 2
- sum of percentage HOT 1
- internalpositioning.com is unreachable HOT 3
- Broken link on GitHub about section HOT 1
- couldnt find getting magnetic data in android app find3 HOT 1
- single quote in device name breaks sql queries HOT 1
- How often does a push notifications to MQTT happen? HOT 1
- broken link in repo about tab HOT 1
- Unable to see any raw data
- Suggestion / Improvement HOT 2
- Why ? HOT 1
- Using multiple raspberry pi's to monitor the WiFi
- Feature request: Openhab compability
- Google Play, cannot find FIND3 HOT 2
- Markupsafe breaks with latest version of setuptools
- FIND3 scanner app getting crashed
- Location Error
- Need information
- find3 like a Wifi Heat Map
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from find3.