Filtering strategies for RSSI signal prediction

Implementation of all filtering strategies described in [1] to filter a noisy RSSI signal.

This is:

• Grey filter
• Fourier Transform filter
• Kalman filter
• Particles filter

Although [1] refers to a RSSI signal, this implementation can be runned with any time series.

Getting Started

Clone repository

~ $ git clone https://github.com/philipiv/rssi-filtering-kalman.git
~ $ cd rssi-filtering-kalman

Project requirements

It is strongly advised you work in a virtual environment.
First step is to create one and install all necessary project requirements.

~/rssi-filtering-kalman $ virtualenv env --python=python3
~/rssi-filtering-kalman $ source env/bin/activate
~/rssi-filtering-kalman $ pip install -r requirements.txt

Execution

~/rssi-filtering-kalman $ cd scripts
~/rssi-filtering-kalman/scripts $ python main.py [--file /path/to/file]

Optionaly, you can set the path to a file containing your data, default path is ../data/sample.csv.

For example:

~/rssi-filtering-kalman/scripts $ python strategy.py --file ../data/sample.csv

Results

After execution, the script output is a Figure containing original signal and output to all filters.

When executed with the sample data the output looks like this:

References

[1] P. Bellavista, A. Corradi and C. Giannelli, "Evaluating Filtering Strategies for Decentralized Handover Prediction in the Wireless Internet," 11th IEEE Symposium on Computers and Communications (ISCC'06), Cagliari, Italy, 2006, pp. 167-174.