Context Aware Smart Environment for the Internet of Things

Abstract

The concept of the Smart Environment is based on Internet of things (loT) infrastructure. The loT infrastructure is composed of the sman objects. These are the objects that possess the ability to extract information from the environment where they are deployed, they might share this information with other smart objects or sends this directly on the designated server. Smart environment is composed of hundreds and thousands of smart objects each having different nature of job, consequently their data is also heterogenous in nature. These smart objects generates a massive amount of data refetTed as "BIG DATA''. This heterogenous nature of data calls for an automated system in place to classify this data in real time into their respective classes so that the decision making can be perfom1ed in much improved way. To address these issues this project presents the concept of an inte lligent system using machine learning algorithm. So, for the classification of this data we have used "Support Vector Machine" which is a supervised machine learning algorithm. The algorithm works in a way that its first trained on the basis of training set and after the training stage the real time data is provided to the SVM for the purpose of classiflcati on. This classified data play a vital role in performing various tasks and actions and is also stored in the database. A web application is also being developed which serves as a main dashboard where all the realtime and his torical data can be seen.