By mimicking cells, MIT researcher designs electronic circuits for ultra-low-power and biomedical applications.
A single cell in the human body is approximately 10,000 times more energy-efficient than any nanoscale digital transistor, the fundamental building block of electronic chips. In one second, a cell performs about 10 million energy-consuming chemical reactions, which altogether require about one picowatt (one millionth millionth of a watt) of power.
MIT's Rahul Sarpeshkar is now applying architectural principles from these ultra-energy-efficient cells to the design of low-power, highly parallel, hybrid analog-digital electronic circuits. Such circuits could one day be used to create ultra-fast supercomputers that predict complex cell responses to drugs. They may also help researchers to design synthetic genetic circuits in cells
In his new book, Ultra Low Power Bioelectronics (Cambridge University Press, 2010), Sarpeshkar outlines the deep underlying similarities between chemical reactions that occur in a cell and the flow of current through an analog electronic circuit. He discusses how biological cells perform reliable computation with unreliable components and noise (which refers to random variations in signals — whether electronic or genetic). Circuits built with similar design principles in the future can be made robust to electronic noise and unreliable electronic components while remaining highly energy efficient. Promising applications include image processors in cell phones or brain implants for the blind.
Sarpeshkar, an electrical engineer with many years of experience in designing low-power and biomedical circuits, has frequently turned his attention to finding and exploiting links between electronics and biology. In 2009, he designed a low-power radio chip that mimics the structure of the human cochlea to separate and process cell phone, Internet, radio and television signals more rapidly and with more energy efficiency than had been believed possible.
That chip, known as the RF (radio frequency) cochlea, is an example of "neuromorphic electronics," a 20-year-old field founded by Carver Mead, Sarpeshkar's thesis advisor at Caltech. Neuromorphic circuits mimic biological structures found in the nervous system, such as the cochlea, retina and brain cells.
Sarpeshkar's expansion from neuromorphic to cytomorphic electronics is based on his analysis of the equations that govern the dynamics of chemical reactions and the flow of electrons through analog circuits. He has found that those equations, which predict the reaction's (or circuit's) behavior, are astonishingly similar, even in their noise properties.
Cells may be viewed as circuits that use molecules, ions, proteins and DNA instead of electrons and transistors. That analogy suggests that it should be possible to build electronic chips — what Sarpeshkar calls "cellular chemical computers" — that mimic chemical reactions very efficiently and on a very fast timescale.
Netbook Technology News