BLEWBAN_Dataset

BLEWBAN is a raw RF dataset of Bluetooth Low Energy (BLE) signals focused on Wireless Body Area Netwrok (WBAN). It consists of on-body and off-body recordings usign ESP32s in BLE mode. The advantages of this dataset is:

• It covers the entire bandwith of the BLE technology.(recorded at 2.44GHz at 100MSps)
• Recording in anechoic chamber to reduce unwanted signals or interference.
• On-body recording on 12 different locations including: both left and right head, arm, wrist, chest, front and back torso (waist).
• Off-body recording with the same devices at 7-different orientations

Google colab examples

• Tutorial
• CNN example

IEEE WCNC presentation YouTube Overview

How to access dataset

Raw recording files

the raw files are available on IEEE Dataport with the DOI : https://dx.doi.org/10.21227/mtg7-eb43. In the RAWFileProcessing folder the tutorial.ipynb notebook provides a user friendly introduction of how to manage handle the raw files.

MongoDB database

However, the Python tools in this repository provide a user-friendly access to the dataset stored in a MongoDB database. A MongoDB database that can easily run queries in multiple languages even on online platforms such as Google Colab (example here). Supported languages can be found here.

An example of how to perform a query and store the results in a Pandas data frame on a Google Colab is available !

Installing requierments

pip3 install pymongo
pip3 install pandas
pip3 install numpy

After succesfull installation import dependencies.

import pymongo
import numpy as np
import pandas as pd

Making connection to MongoDB database

There are two databases available.

• BLE
• BLE_metadata

BLE_metadata is the light version of the BLE excluding raw data. It contains basic time and frequency charectristics of the raw data and is much faster. On the other hand, The BLE database include raw recordings along with BLE_metadata.

For BLE use the following connection string

client = pymongo.MongoClient("mongodb://test:12345678910111213@SG-pine-beat-9444-57323.servers.mongodirector.com:27017/BLE")
BLE_WBAN = client["BLE"]

Or for BLE_metadata use the following connection string

client = pymongo.MongoClient("mongodb://test:12345678910111213@SG-pine-beat-9444-57323.servers.mongodirector.com:27017/BLE_metadata")
BLE_WBAN = client["BLE_metadata"]

Quering data

For example we are trying to filter off-body data recorded at atantenna 2 from device 1 in position 6 with 9dbm TX power.

filter = {'antenna': '2', 'dvc':'1', 'pos': '6', 'txPower': '9dbm'}
query = list(BLE_WBAN.offBody.find(filter))
df = pd.DataFrame(query)

the above store a dataframe with columns such as:

	_id	frame_origin_file	frame_nr	date	dvc	pos	SDR	test	txPower	antenna	...	Fc	gain	frameTime	lenFrame	frameChnl	frameDecode	bitLen	max_gradient_unwrapped_phase	I	Q
0	641aa1d333089480e2d0a5ca	/home/moh/...	1	jan 13 2023	1	6	1	offBody	9dbm	2	...	2439810000	31.5	[1478146.0, 1487348.0]	9202.0	3	0x5553d3d3c468a54267b1	[185, 92, 101, ...	[-0.014498871, 0.0127564659, ...	[-0.01559495, -0.0057069, ...	[-0.00430310, 0.014526810, ...
1	641aa1d333089480e2d0a5cb	/home/moh/...	3	jan 13 2023	1	6	1	offBody	9dbm	2	...	2439810000	31.5	[2978115.0, 2987351.0]	9236.0	12	0x5553d3d3c4b4f71616de0	[261, 96, 105, ...	[-0.013399868, 0.0109354282, ...	[0.00198370, -0.0071718, ...	[-0.01312295, -0.01129184, ...
2	641aa1d333089480e2d0a5cc	/home/moh/...	8	jan 13 2023	1	6	1	offBody	9dbm	2	...	2439810000	31.5	[4478106.0, 4487332.0]	9226.0	21	0x5553d3d3c40ce5d128c1	[256, 96, 105, ...	[-0.013268690, 0.0109537563, ...	[-0.00790429, -0.00302133, ...	[-0.00939970, -0.01232947, ...
3	641aa1d333089480e2d0a5cd	/home/moh/...	10	jan 13 2023	1	6	1	offBody	9dbm	2	...	2439810000	31.5	[5978104.0, 5987324.0]	9220.0	30	0x5553d3d3c4f4645966200	[251, 96, 104, ...	[-0.014260522, 0.0117406014, ...	[-0.00427259, -0.01013215, ...	[0.01345866, -0.005706963, ...

Data interpretaion

Since storing Complex numbers is not possible in the MongoDB database columns I and Q are real and imaginary numbers respectively. The code below will generate a new column named frame that is $I + j*Q$

df['frame'] = query.apply(lambda x: x['I'] + np.dot(x['Q'],1j), axis=1)

Examples of the BLE-WBAN

There is ongoing research on analyzing the on-body data. The plot below shows the classification results of 12 on-Body devices in moving and static body positions.