Entity

Time filter

Source Type


Valarmathi K.,Panimalar Engineering College
Asian Journal of Information Technology | Year: 2016

In WiMAX OFDMA networks, the existing resource allocation and scheduling techniques results in more overhead and delay. In order to overcome this problem, in this study, we propose HMM and fuzzy based scheduling and resource allocation for Downlink Traffic in WiMAX OFDMA networks. In this technique, new users are admitted after checking the predicted free slots using Hidden Markov Model (HMM) against the estimated slots of each user. Then the exact number of slots for each user is estimated using fuzzy logic inference considering QoS requirement such as bandwidth, delay and transmission rate of users over each subchannel. The scheduling of slots on each sub channel level is performed based on user priority which is estimated again using fuzzy logic inference considering the channel quality, queue length and urgency factor. By simulation results, we show that the proposed technique enhances the network throughout and minimizes the overhead. © Medwell Journals, 2016. Source


Babu R.S.R.,Sathyabama University | Deepa S.,Panimalar Engineering College | Jothivel S.,Sathyabama University
International Journal of Engineering, Transactions B: Applications | Year: 2014

This paper presents an implementation of open loop and closed loop control of quadratic boost converter (QBC) using PID-controller. QBC consists of boost converter and fly back converter driven by a single switch. QBC is designed especially for regulating the DC interface between various micro sources and DC-AC inverter to electricity grid. QBC, P, PI and PID-controller are modeled, compared and evaluated by MATLAB simulation. It has been found that the transient and steady state performance is improved using PID-controller. This converter achieves high step-up voltage gain with appropriate duty ratio and low voltage stress on the power switch. The simulated open loop and closed loop performance is verified experimentally. Source


Kannan P.,Panimalar Engineering College
International Journal of Applied Engineering Research | Year: 2014

The detection of moving objects from video sequences is a very significant process for object tracking, object focusing, and activity analysis in the field of video surveillance. Background subtraction plays a vital role in implementing such operation. The objective of proposed algorithm is to extract the foreground image by subtracting the background image. Along with the frame differencing method, Automatic Threshold Update (ATU) and Discrete Wavelet Transform (DWT) are included for the background subtraction. The Grey scale and RGB computations are evaluated with the help of eight methods for uniform light and bright light videos. The proposed frame difference method with Automatic threshold Uptation and Discrete Wavelet Transform gives better background subtraction based on analysis of accuracy and Reduction in computation time. © Research India Publications. Source


Subramanian S.,Panimalar Engineering College | Badrilal M.,Indian Institute of Technology Bombay | Henry J.,Veltech University
International Review of Electrical Engineering | Year: 2010

This paper presents using wavelet transform for differential protection and neural networks for fault classification. The wavelet transform is one of the powerful signal-processing tool is used to extract information's of differential current from wavelet coefficients. A two cycles of fault current data's are processed to wavelet transforms, which produces wavelet coefficients, and features like energy and standard deviations are calculated using Parseval's theorem. Extracted features are processed to Probability neural network and support vector machine for fault classification. Application of the proposed algorithm has been analyzed using different mother wavelets for signal processing. The application of the proposed simulation model has been studied by simulation of the fault (with and without noise) using MATLAB/SIMULINK software taking 2 cycles of data window (40 ms) containing 800 samples. Tripping signals are issued using the energy vectors that calculated from wavelet coefficients. Classification accuracy for the combination of wavelet transform and Support vector machine has been found better than the combination of wavelet transform and Probabilistic neural network. © 2010 Praise Worthy Prize S.r.l. - All rights reserved. Source


Devadoss A.K.,Panimalar Engineering College
Biomedical Research (India) | Year: 2016

Human beings have numerous sensors to sense the state of the environment and actuators to perform different actions. One such vital sensor is the Nostril. The nostrils are essential for life. Hence, coming in contact with hazardous gases causes danger to human life. In many developing countries, lthe sewers are still cleaned by unskilled labourers. Situations may arise where hazardous gas is emitted via sewage and can potentially endanger life. Furthermore, in coal mining, there is a possibility of hitting a source of natural gas which cannot be determined unless or until a sensor is utilized. To prevent such hazardous situations, this new gas detection system detects those types of gases, analyzes them for us and provides essential details about it. Our system is designed to track the presence of hazardous gases, identify the safety limit and calculate the level in that situation, thereby preventing hazards to human life. It allows detection of carbon monoxide and methane at the given time, along with their accurate concentration values in ppm. The basic idea of this system is to calculate a safe limit point 'x', when the system detects a value 'y' (near to and less than 'x') and then alert the environment. If the value exceeds 'x',then system should provide a higher level threat detection alert so that the persons immediately evacuate that area, thereby preventing any possible dangers. The alert messages are broadcasted using GSM technology and hence, can be used to notify other rescue workers about the potential hazard the worker is facing at the moment. © 2016, Scientific Publishers of India. All rights reserved. Source

Discover hidden collaborations