TWIE 145: Real-Time Brain Simulation

This Week in Engineering - Solar dish desalination; bike-powered charger; gentle touch sensors; Grasshopper hovers; nineteenth century recording; and real-time brain simulation.

blog comments powered by Disqus
Recent Video
TWIE 164: Lunar Solar Power
TWIE 162: Sonic Solar Cells
TWIE 161: Dark Matter Miss
TWIE 160: Self-Healing Metals
Transcript For This Video

Solar dish desalination
To help improve solar power¦s efficiency and affordability, IBM, together with Airlight Energy, have developed the High Concentration Photo Voltaic Thermal system, which uses a parabolic array of mirrors to concentrate sunlight onto a number of receivers. With that much energy, the receiver chips must be actively cooled, but instead of wasting the heat, the system uses it to desalinate water. Salt water pumped to the receiver reaches 90 degrees celsius, and passes through a porous membrane distillation system. A large system could desalinate 30-40 L of drinkable water per square meter of receiver per day, and researchers believe the system could be built at about one-third the cost of comparable photovoltaic systems.

Bike-powered charger
Siva Cycle has created the Atom, a lightweight generator that mounts on a bicycle wheel for charging USB devices like smartphones. The device converts a small amount of the bike¦s kinetic energy into highly variable electric current, which is then smoothed into the 5 V, 500 mA DC expected by USB devices. At about ten ounces, the device will not substantially affect a bike¦s balance, and it can charge a phone or other device about as quickly as plugging it into a computer. Charging phones with bikes? Don¦t stop there -- how about a hand crank? It¦s like a nineteenth century Magneto, but with Fruit Ninjas.

Gentle touch sensors
Scientists from the Georgia Institute of Technology have created a flexible array of pressure sensors, with a sensitivity comparable to skin. The team created transistors made with zinc oxide, which is not only a semiconductor, but also piezoelectric, meaning it generates current when bent. Each transistor held about 1500 zinc oxide nanowires, all integrated into a thin, transparent flexible polymer sheet. The transistor array could determine where pressure is concentrated. The research could someday give robots and prosthetics tactile sensitivity approaching human skin. Researchers, keep it up! I won¦t be satisfied until my personal robot can find the contact lense I just dropped on plush carpet.

Grasshopper hovers
Private spaceflight company SpaceX has successfully tested a prototype of their Grasshopper reusable rocket, which, in a one-minute flight, hovered at an altitude of 250 m before returning back to the launch pad and landing. The ten-story tall Grasshopper uses one Merlin engine -- the first stage of the Falcon 9 rocket which has flown cargo missions to the International Space Station. The Grasshopper program is developing technology for a fully-reusable rocket to make spaceflight cheaper and more efficient.

Nineteenth century recording
In 1885, inventor of the telephone Alexander Graham Bell made a recording of his voice on a crude disk made of wax and paperboard, which has been preserved in the Smithsonian Museum. Now, researchers have taken these disks, too fragile to play, and read the sound waves with a high-resolution digital scanner first developed for particle physics experiments. The recording can help validate other recordings of the telephone¦s inventor, and also help scholars better understand 19th-century speech. Amazing! The recording quality from that long ago is still less annoying than state-of-the-art autotune.

Real-time brain simulation
The (Brains in Silicon÷ laboratory of Stanford University has recently created Neurogrid, a computer system that simulates the activity of one million neurons in real time. Unlike traditional computing, Neurogrid uses analog signals for computations, in addition to the more standard digital communication signals. This is not the first brain simulation -- IBM¦s Blue Gene/Q Sequoia supercomputer simulated 530 billion neurons -- but it couldn¦t keep up with real time, clocking in 1500 ti