In the Social Internet of Things (SIoT),  the social relationship is built on  the object’s relationship in a smart environment. There are many challenges in social IoT, the major being object mobility, scalability, and pattern analysis in a smart environment. The smart objects have limited intelligence to make decisions to predict the corresponding data. This seminar focuses on giving insights on the model proposed to provide the services to the user in a given environment. The proposed model provides services by adding relationship knowledge to the existing model using R-LSTM network to determine the right data and predicting responding objects and relationship between the objects. The working principle of the proposed model and the results obtained will be discussed during this talk. The dataset used and the proposed model will be demonstrated using Jupyter notebook. 

Hyper scaling of IoT sensors leads to the connected device segments like smartphones, smart watches, smart home, smart city and many more. These hyper scaling devices which are in a social relationship form a social IoT. The Social Internet of Things (SIoT) is a network of interconnected heterogeneous or homogeneous objects with social relationships as in humans.

The Internet of Things or IoT is a network of interconnected heterogeneous objects that are uniquely identifiable which provide data transferability without the need for human-to-computer or human-to-human interaction. Figure 1 shows the conceptual SIoT environment.  

Figure 1: Conceptual SioT Environment 

It consists of a social network having intelligent objects, known as the social internet of things, which is a mapping between objects and the social network of humans. In SIoT, objects have the ability to interact and behave in a social manner like human relationships.

SIoT has 10 types of relationships namely parent object relationship, owner object relationship, guardian object relationship, social object relationship, sibling object relationship, guest object relationship, service object relationship, strange object relationship, co-location object relationship, and co-work object relationship. 

The proposed methodology will discussed is as shown in figure 2. 

Figure 2: Proposed methodology

The proposed work is carried out using relationships between the objects on the LSTM model. It predicts the time of forecasting services in terms of values each device obtains like closeness of classes in services like air quality, weather, temperature, tra c, people presence and parking status within a smart city. The proposed R-LSTM based service oriented knowledge model uses the relationship between response is device and requesting devices to predict the services to the users with accuracy of 98.75% and loss of 0.37%. It requires an object profiling in every object otherwise relationship can not be identified is the only one limitation in this work. In the future, analyze real time objects using deep learning models.

The presentation covers as follows:

Theoretical knowledge of SIoT and LSTM:

1. Introduction on social objects characteristics, challenges and few applications.
2. Related Works on LSTM modules work.
3. Problem Statement of this work focused on developing relationship models based on device mobility that predicts the responding devices to provide services for a smart city environment application.
4. Design and Methodology of the proposed work of R-LSTM.
5. Results and Discussion about R-LSTM.
6. Conclusion and Future Work

Demonstrate the Usage of few packages in the R-LSTM 

Speaker: Dr.S.P. Shiva Prakash, Associate Professor, Department of Information Science and Engineering, JSS Science and Technology University 

Seminar in English.

Beginning at 16:30 Moscow time zone, June 7.

The workshop will take place on Zoom. The link will be sent automatically 1 hour before the start.