Enabling probabilistic conflict prediction in smart space through context-awareness

This project proposes a data-driven approach to predict conflicts of IoT devices among multiple user based on their previous interaction histories. Here, a conflict is a situation when two users' preferences over the state of an IoT device differ. For example, a user wants to turn off the light while another user wants to maintain illumination with the same light.

Getting started

Python version

This project is tested with python3 only

Installing prerequisites

All the required packages are listed in the requirements.txt. The most simple way to install the depencies is by running

pip install -r requirements.txt

Prepare the data

• Find the REFIT dataset from this DOI. Download and unzip CLEAN_REFIT_081116.7z in the listed files.
• Create folders ./data/refit under the root directory of this project. And put all files from the downloaded dataset to ./data/refit. Expected file structure:

ConflictDetection
│   README.md
│   main.py    
│
└───data
│   └───refit
│       │   CLEAN_House1.csv
│       │   CLEAN_House2.csv
│       │   ...

• run data_prepare.py from the root directory: python3 data_prepare/data_prepare.py

Run example

To verify the project is set up correctly and to see some real conflict scenarios, run:

python3 main.py

Other than the debug messages (starts with DEBUG) the expected results is:

The number of conflicts we found for the devices are:
{'TV': 21, 'WashingMachine': 4, 'PC': 0}
Top 5 conflict scenarios with the highest probability for TV
House4 and House3 have 53.32% to have conflicts over TV at {'min_of_day#NUM': [1140.0, 1320.0], 'weatherDesc#CAT': {'Rain', 'HeavyRain', 'Fog', 'Cloudy'}, 'day_of_week#NUM': [0.0, 7.0]}
House4 and House3 have 44.29% to have conflicts over TV at {'min_of_day#NUM': [1080, 1140.0], 'weatherDesc#CAT': {'Rain', 'HeavyRain', 'Fog', 'Cloudy'}, 'day_of_week#NUM': [0.0, 7.0]}
House4 and House3 have 38.02% to have conflicts over TV at {'min_of_day#NUM': [1140.0, 1320.0], 'weatherDesc#CAT': {'Clear'}, 'day_of_week#NUM': [0.0, 7.0]}
House4 and House3 have 31.59% to have conflicts over TV at {'min_of_day#NUM': [1020.0, 1140.0], 'weatherDesc#CAT': {'Clear'}, 'day_of_week#NUM': [0.0, 7.0]}
House3 and House4 have 26.94% to have conflicts over TV at {'min_of_day#NUM': [1020.0, 1080], 'weatherDesc#CAT': {'Rain', 'Fog', 'Cloudy'}, 'day_of_week#NUM': [0.0, 7.0]}
House4 and House3 have 17.09% to have conflicts over TV at {'min_of_day#NUM': [1320.0, 1380.0], 'weatherDesc#CAT': {'Rain', 'HeavyRain', 'Fog', 'Cloudy'}, 'day_of_week#NUM': [0.0, 7.0]}
House3 and House4 have 11.30% to have conflicts over TV at {'min_of_day#NUM': [420.0, 600], 'weatherDesc#CAT': {'HeavyRain', 'Rain', 'HeavySnow', 'Snow'}, 'day_of_week#NUM': [0.0, 7.0]}
House3 and House4 have 11.14% to have conflicts over TV at {'min_of_day#NUM': [1320.0, 1380.0], 'weatherDesc#CAT': {'Clear'}, 'day_of_week#NUM': [0.0, 7.0]}
House3 and House4 have 10.17% to have conflicts over TV at {'min_of_day#NUM': [420.0, 600], 'weatherDesc#CAT': {'Fog', 'Clear', 'Cloudy'}, 'day_of_week#NUM': [0.0, 7.0]}
House3 and House4 have 8.27% to have conflicts over TV at {'min_of_day#NUM': [780.0, 1020.0], 'weatherDesc#CAT': {'Cloudy', 'HeavySnow', 'Rain', 'Snow', 'HeavyRain', 'Clear', 'Fog'}, 'day_of_week#NUM': [0.0, 7.0]}