Power/Performance Bits: Nov 28


Deep learning to detect nuclear reactor cracks Inspecting nuclear power plant components for cracks is critical to preventing leaks, as well as to control in maintenance costs. But the current vision-based crack detection approaches are not very effective. Moreover, they are prone to human error, which in the case of nuclear power can be disastrous. To address this problem, Purdue Universit... » read more

Manufacturing Bits: Nov. 21


Germanium-on-mica Germanium is an element that can be used in various applications in electronics, such as optoelectronics, semiconductors and others. For example, silicon-germanium (SiGe), an alloy of silicon and germanium, is used for making RF chips. In future finFET transistors, some are exploring the idea of using pure germanium for the PFET structure to boost the electron mobility in ... » read more

System Bits: Nov. 21


MIT-Lamborghini to develop electric car Members of the MIT community were recently treated to a glimpse of the future as they passed through the Stata Center courtyard as the Lamborghini Terzo Millenio (Third Millennium) was in view, which is an automobile prototype for the third millennium. [caption id="attachment_429209" align="alignnone" width="300"] Lamborghini is relying on MIT to make i... » read more

Power/Performance Bits: Nov. 21


Greener greenhouses Researchers at the University of California, Santa Cruz are testing greenhouses capable of generating some of their own energy, without hampering plant growth. Greenhouses use electricity to control temperature and power fans, lights, and other monitoring systems. Electricity-generating solar greenhouses utilize Wavelength-Selective Photovoltaic Systems (WSPVs), a novel ... » read more

Manufacturing Bits: Nov. 14


GaN for electric cars Leti is coordinating a new European project to improve the drivetrain in electric vehicles. The so-called ModulED project will focus on the development of gallium nitride (GaN) technology for electric vehicles. The goal is to use power-based GaN devices for the motor, enabling a change from direct current to alternating current. The three-year, €7.2 million proje... » read more

System Bits: Nov. 14


Tracking cyber attacks According to Georgia Tech, assessing the extent and impact of network or computer system attacks has been largely a time-consuming manual process, until now since a new software system being developed by cybersecurity researchers here will largely automate that process, allowing investigators to quickly and accurately pinpoint how intruders entered the network, what data... » read more

Power/Performance Bits: Nov. 14


Bacteria power wastewater cleanup Researchers at the King Abdullah University of Science and Technology (KAUST) are exploring ways to detoxify warm, salty industrial wastewater while simultaneously generating electricity. They are using bacteria with remarkable properties: the ability to transfer electrons outside their cells (exoelectrogenes) and the capacity to withstand extremes of temperat... » read more

Manufacturing Bits: Nov. 7


Making a superbeam Lawrence Livermore National Laboratory (LLNL) has combined several lasers to create what it calls a superbeam. The move represents a possible breakthrough in the arena. In theory, lasers can be combined. But the laser beams tend to pass through each other, thereby making a combined laser or a superbeam nearly impossible. With the help of plasma optics, however, LLNL ha... » read more

Power/Performance Bits: Nov. 7


Speeding up MRAM Researchers at UC Berkeley and UC Riverside developed an ultrafast method for electrically controlling magnetism in certain metals, which could lead to increased performance for magnetic RAM. While the nonvolatility of MRAM is a boon, speeding up the writing of a single bit of information to less than 10 nanoseconds has been a challenge. “The development of a non-volatile... » read more

System Bits: Nov. 7


Exposing logic errors in deep neural networks In a new approach meant to brings transparency to self-driving cars and other self-taught systems, researchers at Columbia and Lehigh universities have come up with a way to automatically error-check the thousands to millions of neurons in a deep learning neural network. Their tool — DeepXplore — feeds confusing, real-world inputs into the ... » read more

← Older posts Newer posts →