Loading...

A Framework For Situation Recognition in IOT Environments

Rajaby Faghihi, Hossein | 2018

782 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 51607 (19)
  4. University: Sharif University of Technology
  5. Department: Computer Engineering
  6. Advisor(s): Habibi, Jafar; Fazli, Mohammad Amin; Movaghar, Ali
  7. Abstract:
  8. According to the increase of interest in the internet of things, the importance of understanding the ongoing situation in an environment and acting upon them is largely considered. Due to the massive size of sensors used in a smart environment, there are plenty of data manufactured every second. Therefore, an interpretation of this data is mandatory to obtain valuable information to comprehend the situation ongoing in an environment. Situation recognition can bring a better understanding of the interaction between the smart actuators and sensors with the living agents in the environments.This will enable the system to help people do their routines, save energy, and accomplish more complicated tasks easily. Some significant characteristics of such environments are the multi-agent manner, multi-sensor, uncertainty, and dynamic nature.Some existing methods have utilized machine learning techniques to interpret data received by sensors while others have used semantic relations and human knowledge. In this thesis, we have discussed a hybrid approach jointly with an architecture for the smart environments to be able to interpret situations while facing the challenges mentioned above. Regarding the Framework, The most critical factor in our design is the ability to extend the system in every aspect of its structure. We have also discussed the human interaction and role, the system security over situation execution, data gathering methodology, and data interpretation.Our framework is a three-layer structure containing Sensor layer, Situation Layer and Workflow layer. The Situation Layer is also designed so that it can adapt to the environment by using the hybrid approach we developed utilizing situation templates and decision tree.Finally, we have evaluated our method in two ways. First, we have used the ATAM method to evaluate the architecture. Secondly, we have used the simulation process to evaluate the accuracy of the hybrid learning approach.Our evaluation led us to the result that the proposed framework works better than the reference framework in the cases of extensibility and adaptability. The
    proposed model has shown a better accuracy in comparison with the situation templates
  9. Keywords:
  10. Semantic Web ; Internet of Things ; Situation Recognition ; Framework

 Digital Object List

 Bookmark

...see more