Narula Institute of Technology

About: Narula Institute of Technology is a based out in . It is known for research contribution in the topics: Quantum dot cellular automaton & Cognitive radio. The organization has 288 authors who have published 490 publications receiving 2258 citations. The organization is also known as: NiT.

Unvoiced Speech Recognition Using Dynamic Analysis of EMG Signal

Abstract: The current study is an initiative taken to recognize the speech by observing the muscle associated with speech (Locks et al. in Motriz, Rio Claro 21(1):15–22, 2015, [1]). It will realize unvoiced communication; guarantee the participation of inarticulate people in society. Electromyography (EMG) sensor is used to track and recognize various intramuscular signals for accurate and precise recognition of the subjects. In this aspect, letter ‘A’ and digit ‘1’ are performed in sign language by volunteers and the EMG signals are collected for four muscles, i.e. Lumbrical muscles, Hypothenar muscles, Thenar muscles and Flexor carpi muscles of palm and hands. The nonlinear study of these signals proved to be an efficient method of sign recognition.

Performance analysis of the CAC scheme CARETON under mixed cell environment in overlay networks

Abstract: The paper proposes a call admission control (CAC) scheme to offer best services to the users in the next generation heterogeneous networks. It optimises the available resources without compromising the interest of the service providers. It has combined the three commonly used CAC schemes – guard channel, buffer-based and priority-based schemes. It considers wireless overlay model for analysing the heterogeneity with detailed analysis of the WLAN-cellular mixed cell environment. The scheme considers differentiated services for the real time and non-real time calls to provide appropriate quality of service (QoS). The proposed scheme has been evaluated through simulation. The results show that the scheme outperforms some significant CAC schemes and shows even better performance under cellular-WLAN mixed environment due to the dual subscription of the users with dual mobile terminal.

Use of Convolutional Neural Network (CNN) to Detect Plant Disease

Abstract: The automatic and accurate detection of diseased leaves is a challenging job for researchers. It offers a promising step towards food security and agricultural growth. On contrary, the conventional manual interpretation is time-consuming and expensive. In this paper, it proposes a new approach to detect plant diseases using the deep learning Convolutional Neural Network. We have used 1900 images, taken from a public dataset to train our model. This deep learning model is designed to consist of 25-layer for plant disease classification. The trained model achieved 96.64% accuracy to detect the plant disease. The proposed deep learning convolution neural network model may have great potential in disease detection for current cultivation on large scale.

Overview of Various Bandwidth Enhancement Techniques for Ultra-Wide Band Antennas

Abstract: This article deals with the Ultra-Wide Band (UWB) antennas that are appraised in respect of different geometries, design parameters and their experimental results. Several types of UWB antennas including special horn, micro-strip patch radiators and array antennas in recent works are elucidated, while comparing their measured return loss, gain and radiation patterns. Projections to the future developments of UWB antenna technology are also specified.

Development of home intruder tracking system using face recognition

Abstract: The intrusion of burglars to the houses of common people in their absence intensifies the risk of financial loss in a great manner in recent days. Proper tracking to this type of intrusion and proper notification to the house members as well as the neighbour can reduce this type of malpractice. A system has been proposed in this paper to recognize the faces of intruders in order to identify their faces as illegal infiltrator and prevent them from any burglary activity by face recognition method.