Gehan A. J. Amaratunga

Learn More
The demand for increased information storage densities has pushed silicon technology to its limits and led to a focus on research on novel materials and device structures, such as magnetoresistive random access memory and carbon nanotube field-effect transistors, for ultra-large-scale integrated memory. Electromechanical devices are suitable for memory(More)
This paper proposes a high current impedance matching method for narrowband power-line communication (NPLC) systems. The impedance of the power-line channel is time and location variant; therefore, coupling circuitry and the channel are not usually matched. This not only results in poor signal integrity at the receiving end, but also leads to a higher(More)
In this paper, for the first time, an ensemble of deep learning belief networks (DBN) is proposed for regression and time series forecasting. Another novel contribution is to aggregate the outputs from various DBNs by a support vector regression (SVR) model. We show the advantage of the proposed method on three electricity load demand datasets, one(More)
A solution growth approach for zinc oxide (ZnO) nanowires is highly appealing because of the low growth temperature and possibility for large area synthesis. Reported reaction times for ZnO nanowire synthesis, however, are long, spanning from several hours to days. In this work, we report on the rapid synthesis of ZnO nanowires on various substrates (such(More)
When a carbon nanotube emitter is operated at high currents (typically above 1 microA per emitter), a small voltage drop ( approximately few volts) along its length or at its contact generates a reverse/canceling electric field that causes a saturation-like deviation from the classical Fowler-Nordheim behavior with respect to the applied electric field. We(More)
The information provided by the in-cylinder pressure signal is of great importance for modern engine management systems. The obtained information is implemented to improve the control and diagnostics of the combustion process in order to meet the stringent emission regulations and to improve vehicle reliability and drivability. The work presented in this(More)
Load demand forecasting is a critical process in the planning of electric utilities. An ensemble method composed of Empirical Mode Decomposition (EMD) algorithm and deep learning approach is presented in this work. For this purpose, the load demand series were first decomposed into several intrinsic mode functions (IMFs). Then a Deep Belief Network (DBN)(More)