liuzhidan / gnome

Gnome

This is the basic version of "Gnome: A Practical Approach to NLOS Mitigation for GPS Positioning in Smartphones". Some functions are not optimized. We will keep updating the repo.

Requirement

• Python 2.7.x
• OS X or Ubuntu
• Android 7+

Code Structure

• cloud/: Cloud part of Gnome. It downloads data from Google Street View in the specified area, and compute GPS path inflation.
• mobile/: Mobile part of Gnome. It takes input from Android GPS metadata and the path inflation model from cloud, and computes the adjusted location.

Workflow

Run Cloud

• Go to cloud/ dir, fill alps/config/query_google_key.info with your own google key.
• Run python run.py lat_bottom lon_left lat_top lon_right. The last four parameters specifies the region where Gnome generates the path inflation model. The output is written in inflation.pkl Sample command: python run.py 34.045175 -118.260346 34.049438 -118.256356

Run Mobile

The mobile part of Gnome is running as a python library. Your code should include the Gnome class in mobile/main.py and GPSmeta in mobile.meta.py. Here is the example:

from mobile.main import Gnome
from mobile.meta import GPSmeta

gnome = Gnome(path_to_inflation_model)
gnome_meta = GPSmeta()					# input path to gnss logger file (optional)

while reading_gps:
	position_adjusted = Gnome.update(your_own_data)		# pass data from your own app
	position_adjusted = Gnome.update(gnome_meta.read())	# pass data from gnss logger file 
	 

• Gnome is the main object to compute the new location. Use its update function to get new location reading
• GPSmeta contains raw GPS measurement like C/N0, Pseudorange rate, clock, etc. Please read mobile/meta.py to check how to initiate GPSmeta. You can get the raw GPS measurement data from Android APIs in your app. For API details please read the source code of Android GNSS Logger. GPSmeta also enables you to initiate the meta object with the output file of Android GNSS logger. The Android version of Gnome mobile will come in future.