Design and Implementation of Tensor based Adaptive Algorithms for Applications to Big Data in IoT and Wireless Sensor Networks

In recent times where electronic ga dgets are common in every aspect of our daily life, there is a huge amount of data stemming from different digital devices which is termed as Big Data. Big Data has transformed and revolutionized various fields such as business, healthcare, education, telecommunication industry, transportation, energy sector, just to name a few. For example, wearable devices and sensors are used in the healthcare industry to provide real-time analysis to the electronic health record of a patient. Similarly, in the recent technology of Internet of Things (IoT), there is generation of massive volumes of structured and unstructured data. To analyse and examine performance of such Wireless Sensor Networks (WSN), huge volumes of data need to be processed. In this regard, there are lot of challenges in dealing with the analysis of Big Data as the conventional signal processing tools are inefficient. Hence, there is a need to develop efficient signal processing tools in this regard. Recently, it is found that the mathematical tools based on Tensors are very effective in dealing with Big Data applications such as in Deep Learning applications to image and video processing. Motivated by this, Tensor Least Mean Squares (TLMS) is recently proposed in 2020 for adaptive processing of Big Data. However, there are speed limitations of the TLMS. In this project, we propose to design various Tensor based adaptive algorithms with faster speed for Big Data applications. Moreover, we propose to develop Tensor based adaptive algorithms such as Incremental TLMS and Diffusion TLMS algorithms for distributed processing in IoT and wireless sensor networks. Analytical performance analysis such as mean and mean-square-error (MSE) analysis of these algorithms will also be carried out to guarantee reliable performance. Fina lly, these algorithms will be tested on various challenging Big Data applications by implementing IoT prototype testbeds