Machine Learning On Real-Time Data to Enhance Home Automation

  IJETT-book-cover  International Journal of Engineering Trends and Technology (IJETT)          
© 2016 by IJETT Journal
Volume-42 Number-1
Year of Publication : 2016
Authors : Sameer Shaikh, Ayyaj Attar, Siddik Pathan, Rukaiya Shaikh
DOI :  10.14445/22315381/IJETT-V42P202


Sameer Shaikh, Ayyaj Attar, Siddik Pathan, Rukaiya Shaikh "Machine Learning On Real-Time Data to Enhance Home Automation", International Journal of Engineering Trends and Technology (IJETT), V42(1),6-8 December 2016. ISSN:2231-5381. published by seventh sense research group

Traditional home automation systems are mostly hard-coded or require manual automation plan generation by users. This requires interaction between a control system (an app) and the user, requiring quite some effort and time to be put in manual planning of the home environment. This project intends to explore combining ML and real-time usage data to generate personalized and time-variant home automation plans. These plans will save the user time and effort, leading to a smoother ML driven home automation experience. We collect streaming usage statistics from smart-home occupants and store it on a centralized server. Simultaneously, we also collect “external” data (which may consist of environmental factors like natural light intensity, wind speed, et cetera) which may influence occupants’ usage behavior. These datasets are combined, with data timestamps as a unique identifying field, into a super-set. It’s then fed into a Machine Learning system to correlate user habits with time of the day and the external factors. The correlation hence established will be updated as new data coming into the system in real time or if it crosses a certain percentile ratio threshold. This correlation will be used to generate personalized automation plans for individual occupants. Hence, by combining real-time usage data from a conventional home automation system and Machine Learning, we will be able to provide smoother and more comfortable environment to the users, as the burden of plan generation will be greatly reduced.


[1] Parisa Rashidi, and Diane J. Cook, Keeping the Resident in the Loop: Adapting the Smart Home to the User, in IEEE Transactions On Systems, Man, And Cybernetics, 2009. [2] Panos Luridas, Christof Ebert, Machine Learning, IEEE Computer Society, 2016.
[3] Jérémie Saives, Clément Pianon, and Gregory Faraut, Activity Discovery and Detection of Behavioral Deviations, in IEEE Transactions On Automation Science And Engineering, 2016.
[4] Christopher Osiegbu, Seifemichael B. Amsalu, Fatemeh Afghah, Daniel Limbrick and Abdollah Homaifar, Design and Implementation of an Autonomous Wireless Sensor-based Smart Home, IEEE, 2015.
[5] Wesllen S. Lima, Eduardo Souto, Richard W. Pazzi, Ferry Pramudianto, User Activity Recognition for Energy Saving in Smart Home Environment, IEEE Symposium on Computers and Communication, 2015.
[6] Vahid Ghasemi, Ali Akbar Pouyan, Activity Recognition in Smart Homes Using Absolute Temporal Information in Dynamic Graphical Models, IEEE Computer Society, 2015.
[7] Beichen Chen, Zhong Fan, and Fengming Cao, "Activity Recognition Based on Streaming Sensor Data for Assisted Living in Smart Homes", IEEE Computer Society, 2015.
[8] K.S.Gayathri, Susan Elias, S.Shivashankar, Composite activity recognition in smart homes using Markov Logic Network, IEEE Computer Society, 2015.
[9] Labiba Gillani Fahad & Muttukrishnan Rajarajan, Anomalies detection in smart-home activities, in IEEE 14th International Conference on Machine Learning and Applications, 2015.
[10] Z. Meng, and J. Lu, A Rule-based Service Customization Strategy Context-aware Automation for Smart Home, in IEEE Transactions On Mobile Computing, 2014.
[11] Chiming Chang, Paul-Armand Verhaegen, Joost R. Duflou, A Comparison of Classifiers for Intelligent Machine Usage Prediction, IEEE Computer Society, 2014.

machine learning, home automation, real-time analytics, simple reflex agent, learning agent, anomaly detection, activity detection.