Time filter

Source Type

Akhand M.A.H.,Khulna University of Engineering and Technology | Murase K.,University of Fukui
International Journal of Neural Systems

An ensemble performs well when the component classifiers are diverse yet accurate, so that the failure of one is compensated for by others. A number of methods have been investigated for constructing ensemble in which some of them train classifiers with the generated patterns. This study investigates a new technique of training pattern generation. The method alters input feature values of some patterns using the values of other patterns to generate different patterns for different classifiers. The effectiveness of neural network ensemble based on the proposed technique was evaluated using a suite of 25 benchmark classification problems, and was found to achieve performance better than or competitive with related conventional methods. Experimental investigation of different input values alteration techniques finds that alteration with pattern values in the same class is better for generalization, although other alteration techniques may offer more diversity. © 2012 World Scientific Publishing Company. Source

Saha S.,Khulna University of Engineering and Technology
2014 International Conference on Informatics, Electronics and Vision, ICIEV 2014

Hypovolemia is a dangerous and acute disease for mankind which results form a decreased level of blood volume or a loss of blood. In extreme case, it can cause an irreversible cellular damage which can cause death. In this paper, a fuzzy control system is proposed to detect this hypovolemia class level. The four classes of hypovolemia can be detected by this proposed system for adult persons. This system analyzes five physiological data which come from the devices that are used for monitoring the patients and regarding all these data it'll give the hypovolemia class level of the patients. It'll empower the doctors to take quick decisions in case of surgeries and seizures. This fuzzy system can be embedded using FPGAs or CPLDs so that can also be used to give treatment to remote place patients. © 2014 IEEE. Source

Rana Md.M.,Khulna University of Engineering and Technology
Proceedings of 2010 13th International Conference on Computer and Information Technology, ICCIT 2010

In this paper least square (LS) and linear minimum mean square error (LMMSE) channel estimation (CE) techniques are presented for long term evolution (LTE) single carrier-frequency division multiplexing (SC-FDMA) systems. The main purpose of LTE is to increases data rate but energy utilization both on the mobile terminal as well as network side is important. For doing this, major challenges for LTE terminal are CE and equalization. This paper discusses the CE techniques and challenges imposed by developments in the LTE terminal implementation. Simulation results shows that the LMMSE CE algorithms outperforms the LS in term of mean square error (MSE) by more than around 3dB. Hence, based on a given LTE systems resources and specifications, a appropriate method among the presented methods can be applied. ©2010 IEEE. Source

Ahmed R.,Khulna University of Engineering and Technology | Reifsnider K.,University of South Carolina
International Journal of Electrochemical Science

Electrochemical Impedance Spectroscopy (EIS) is a powerful and proven tool for analyzing AC impedance response. A conventional three electrode EIS method was used to perform the investigation in the present study. Saturated potassium chloride solution was used as the electrolyte and three different material rods were used as working electrodes. Different configurations of electrode area were exposed to the electrolyte as an active area to investigate electrode geometry effects. Counter to working electrode distance was also altered while keeping the working electrode effective area constant to explore the AC response dependence on the variation of ion travel distance. Some controlled experiments were done to validate the experimental setup and to provide a control condition for comparison with experimental results. A frequency range of 100 mHz to 1 MHz was used for all experiments. In our analysis, we have found a noteworthy influence of electrode geometry on AC impedance response. For all electrodes, impedance decreases with the increase of effective area of the electrolyte. High frequency impedance is not as dependent on geometry as low frequency response. The observed phase shift angle drops in the high frequency region with increased working electrode area, whereas at low frequency the reverse is true. Resistance and capacitive reactance both decrease with an increase of area, but resistance response is more pronounce than reactance. For lower frequencies, small changes in working area produce very distinctive EIS variations. Electrode material as well as geometry was systematically varied in the present study. From these and other studies, we hope to develop a fundamental foundation for understanding specific changes in local geometry in fuel cell (and other) electrodes as a method of designing local morphology for specific performance. © 2011 by ESG. Source

Choudhury P.K.,Khulna University of Engineering and Technology
Journal of Optics (India)

Two adaptive Orthogonal Frequency Division Multiplexing (OFDM) modulation formats are experimentally characterized for high data-rate transmission (>10Gb/s) in Reflective Semiconductor Optical Amplifier (RSOA) based access network. The performance of these multi-level modulation techniques is severely affected by non-negligible chirp parameter of RSOA. Here, it is shown that the OFDM system based on water-filling algorithm is not able to guarantee full transmission capacity. In contrast, by using power pre-emphasis technique, transmission capacity is improved up to 17.1 Gb/s at the bit error rate (BER) within Forward Error Correction (FEC) limit over 25 Km fiber span by using commercially available RSOA of limited bandwidth (<1 GHz). © 2014, The Optical Society of India. Source

Discover hidden collaborations