Messages

16

EEE / Fibers that can hear and sing

« on: May 10, 2018, 01:10:48 PM »

For centuries, "man-made fibers" meant the raw stuff of clothes and ropes; in the information age, it's come to mean the filaments of glass that carry data in communications networks. But to Yoel Fink, an associate professor of materials science and principal investigator at MIT's Research Lab of Electronics, the threads used in textiles and even optical fibers are much too passive. For the past decade, his lab has been working to develop fibers with ever more sophisticated properties, to  enable fabrics that can interact with their environment.

In the August issue of Nature Materials, Fink and his collaborators announce a new milestone on the path to functional fibers: fibers that can detect and produce sound. Applications could include clothes that are themselves sensitive microphones, for capturing speech or monitoring bodily functions, and tiny filaments that could measure blood flow in capillaries or pressure in the brain. The paper, whose authors also include Shunji Egusa, a former postdoc in Fink's lab, and current lab members Noémie Chocat and Zheng Wang, appeared on Nature Materials' website on July 11, and the work it describes was supported by MIT's Institute for Soldier Nanotechnologies, the National Science Foundation and the U.S. Defense Department's Defense Advanced Research Projects Agency.

Ordinary optical fibers are made from a "preform," a large cylinder of a single material that is heated up, drawn out, and then cooled. The fibers developed in Fink's lab, by contrast, derive their functionality from the elaborate geometrical arrangement of several different materials, which must survive the heating and drawing process intact.

The right stuff

The heart of the new acoustic fibers is a plastic commonly used in microphones. By playing with the plastic's fluorine content, the researchers were able to ensure that its molecules remain lopsided — with fluorine atoms lined up on one side and hydrogen atoms on the other — even during heating and drawing. The asymmetry of the molecules is what makes the plastic "piezoelectric," meaning that it changes shape when an electric field is applied to it.

In a conventional piezoelectric microphone, the electric field is generated by metal electrodes. But in a fiber microphone, the drawing process would cause metal electrodes to lose their shape. So the researchers instead used a conducting plastic that contains graphite, the material found in pencil lead. When heated, the conducting plastic maintains a higher viscosity — it yields a thicker fluid — than a metal would.

Not only did this prevent the mixing of materials, but, crucially, it also made for fibers with a regular thickness. After the fiber has been drawn, the researchers need to align all the piezoelectric molecules in the same direction. That requires the application of a powerful electric field — 20 times as powerful as the fields that cause lightning during a thunderstorm. Anywhere the fiber is too narrow, the field would generate a tiny lightning bolt, which could destroy the material around it.

Sound results

Despite the delicate balance required by the manufacturing process, the researchers were able to build functioning fibers in the lab. "You can actually hear them, these fibers," says Chocat, a graduate student in the materials science department. "If you connected them to a power supply and applied a sinusoidal current" — an alternating current whose period is very regular — "then it would vibrate. And if you make it vibrate at audible frequencies and put it close to your ear, you could actually hear different notes or sounds coming out of it." For their Nature Materials paper, however, the researchers measured the fiber's acoustic properties more rigorously. Since water conducts sound better than air, they placed it in a water tank opposite a standard acoustic transducer, a device that could alternately emit sound waves detected by the fiber and detect sound waves emitted by the fiber.

In addition to wearable microphones and biological sensors, applications of the fibers could include loose nets that monitor the flow of water in the ocean and large-area sonar imaging systems with much higher resolutions: A fabric woven from acoustic fibers would provide the equivalent of millions of tiny acoustic sensors.

Zheng, a research scientist in Fink's lab, also points out that the same mechanism that allows piezoelectric devices to translate electricity into motion can work in reverse. "Imagine a thread that can generate electricity when stretched," he says.

Ultimately, however, the researchers hope to combine the properties of their experimental fibers in a single fiber. Strong vibrations, for instance, could vary the optical properties of a reflecting fiber, enabling fabrics to communicate optically.

Max Shtein, an assistant professor in the University of Michigan's materials science department, points out that other labs have built piezoelectric fibers by first drawing out a strand of a single material and then adding other materials to it, much the way manufacturers currently wrap insulating plastic around copper wire. "Yoel has the advantage of being able to extrude kilometers of this stuff at one shot," Shtein says. "It's a very scalable technique." But for applications that require relatively short strands of fiber, such as sensors inserted into capillaries, Shtein say, "scalability is not that relevant."

But whether or not the Fink lab's technique proves, in all cases, the most practical way to make acoustic fibers, "I'm impressed by the complexity of the structures they can make," Shtein says. "They're incredibly virtuosic at that technique."

17

EEE / Tuning in to a new hearing mechanism

« on: May 10, 2018, 01:10:15 PM »

More than 30 million Americans suffer from hearing loss, and about 6 million wear hearing aids. While those devices can boost the intensity of sounds coming into the ear, they are often ineffective in loud environments such as restaurants, where you need to pick out the voice of your dining companion from background noise.

To do that, you need to be able to distinguish sounds with subtle differences. The human ear is exquisitely adapted for that task, but the underlying mechanism responsible for this selectivity has remained unclear. Now, new findings from MIT researchers reveal an entirely new mechanism by which the human ear sorts sounds, a discovery that could lead to improved, next-generation assistive hearing devices.

“We’ve incorporated into hearing aids everything we know about how sounds are sorted, but they’re still not very effective in problematic environments such as restaurants, or anywhere there are competing speakers,” says Dennis Freeman, MIT professor of electrical engineering, who is leading the research team. “If we knew how the ear sorts sounds, we could build an apparatus that sorts them the same way.”

In a 2007 Proceedings of the National Academy of Sciences paper, Freeman and his associates A.J. Aranyosi and lead author Roozbeh Ghaffari showed that the tiny, gel-like tectorial membrane, located in the inner ear, coordinates with the basilar membrane to fine-tune the ear’s ability to distinguish sounds. Last month, they reported in Nature Communications that a mutation in one of the proteins of the tectorial membrane interferes with that process.

Sound waves

It has been known for more than 50 years that sound waves entering the ear travel along the spiral-shaped, fluid-filled cochlea in the inner ear. Hair cells lining the ribbon-like basilar membrane in the cochlea translate those sound waves into electrical impulses that are sent to the brain. As sound waves travel along the basilar membrane, they “break” at different points, much as ocean waves break on the beach. The break location helps the ear to sort sounds of different frequencies.

Until recently, the role of the tectorial membrane in this process was not well understood.

In their 2007 paper, Freeman and Ghaffari showed that the tectorial membrane carries waves that move from side to side, while up-and-down waves travel along the basilar membrane. Together, the two membranes can work to activate enough hair cells so that individual sounds are detected, but not so many that sounds can’t be distinguished from each other.

Made of a special gel-like material not found elsewhere in the body, the entire tectorial membrane could fit inside a one-inch segment of human hair. The tectorial membrane consists of three specialized proteins, making them the ideal targets of genetic studies of hearing.

One of those proteins is called beta-tectorin (encoded by the TectB gene), which was the focus of Ghaffari, Aranyosi and Freeman’s recent Nature Communications paper. The researchers collaborated with biologist Guy Richardson of the University of Sussex and found that in mice with the TectB gene missing, sound waves did not travel as fast or as far along the tectorial membrane as waves in normal tectorial membranes. When the tectorial membrane is not functioning properly in these mice, sounds stimulate a smaller number of hair cells, making the ear less sensitive and overly selective.

Until the recent MIT studies on the tectorial membrane, researchers trying to come up with a model to explain the membrane’s role didn’t have a good way to test their theories, says Karl Grosh, professor of mechanical and biomedical engineering at the University of Michigan. “This is a very nice piece of work that starts to bring together the modeling and experimental results in a way that is very satisfying,” he says.

Mammalian hearing systems are extremely similar across species, which leads the MIT researchers to believe that their findings in mice are applicable to human hearing as well.

New designs

Most hearing aids consist of a microphone that receives sound waves from the environment, and a loudspeaker that amplifies them and sends them into the middle and inner ear. Over the decades, refinements have been made to the basic design, but no one has been able to overcome a fundamental problem: Instead of selectively amplifying one person’s voice, all sounds are amplified, including background noise.

Freeman believes that by incorporating the interactions between the tectorial membrane and basilar membrane traveling waves, this new model could improve our understanding of hearing mechanisms and lead to hearing aids with enhanced signal processing. Such a device could help tune in to a specific range of frequencies, for example, those of the person’s voice that you want to listen to. Only those sounds would be amplified.

Freeman, who has hearing loss from working in a noisy factory as a teenager and side effects of a medicine he was given for rheumatic fever, worked on hearing-aid designs 25 years ago. However, he was discouraged by the fact that most new ideas for hearing-aid design did not offer significant improvements. He decided to conduct basic research in this area, hoping that understanding the ear better would naturally lead to new approaches to hearing-aid design.

“We’re really trying to figure out the algorithm by which sounds are sorted, because if we could figure that out, we could put it into a machine,” says Freeman, who is a member of MIT’s Research Laboratory of Electronics and the Harvard-MIT Division of Health Sciences and Technology. His group’s recent tectorial membrane research was funded by the National Institutes of Health.

18

EEE / New way to grow microwires

« on: May 10, 2018, 01:09:32 PM »

Microwires made of silicon — tiny wires with a thickness comparable to a human hair — have a wide range of possible uses, including the production of solar cells that can harvest much more sunlight for a given amount of material than a conventional solar cell made from a thin wafer of silicon crystal. Now researchers from MIT and Penn State have found a way of producing such wires in quantity in a highly controlled way that could be scaled up to an industrial-scale process, potentially leading to practical commercial applications.

Other ways of making such wires are already known, and prototypes of solar cells made from them have been produced by several researchers. But these methods have serious limitations, says Tonio Buonassisi, MIT professor of mechanical engineering and a co-author of a paper on the new work that was recently published online in the journal Small, and will soon appear in the print edition. Most require several extra manufacturing steps, provide little control over the exact sizes and spacing of the wires, and only work on flat surfaces. By contrast, the new process is simple yet allows precise control over the wire dimensions and spacing, and could theoretically be done on any kind of curved, 3-D surface.

Microwires are thought to be capable of reaching efficiencies close to those of conventional solar cells in converting sunlight to electricity, but because the wires are so tiny they would do so using only a small fraction of the amount of expensive silicon needed for the conventional cells, thus potentially achieving major reductions in cost.

In addition to microwires’ potential use in solar cells, other researchers have proposed ways such microscopic wires could be used to build new kinds of transistors and integrated circuits, as well as electrodes for advanced batteries and certain kinds of environmental monitoring devices. For any of these ideas to be practical, however, there must be an efficient, scalable manufacturing method.

The new method involves heating and intentionally contaminating the surface of a silicon wafer with copper, which diffuses into the silicon. Then, when the silicon slowly cools, the copper diffuses out to form droplets on the surface. Then, when it is placed in an atmosphere of silicon tetrachloride gas, silicon microwires begin to grow outward wherever there is a copper droplet on the surface. Silicon in the gas dissolves into these copper droplets, and then after reaching a sufficient concentration begins to precipitate out at the bottom of the droplet, onto the silicon surface below. This buildup of silicon gradually elongates to form microwires each only about 10 to 20 micrometers (millionths of a meter) across, growing up from the surface. The whole process can be carried out repeatedly on an industrial manufacturing scale, Buonassisi says, or even could potentially be adapted to a continuous process.

The spacing of the wires is controlled by textures created on the surface — tiny dimples can form centers for the copper droplets — but the size of the wires is controlled by the temperatures used for the diffusion stage of the process. Thus, unlike in other production methods, the size and spacing of the wires can be controlled independently of each other, Buonassisi says.

The work done so far is just a proof of principle, he says, and more work remains to be done to find the best combinations of temperature profiles, copper concentrations and surface patterning for various applications, since the process allows for orders-of-magnitude differences in the size of the wires. For example, it remains to be determined what thickness and spacing of wires produces the most efficient solar cells. But this work demonstrates a potential for a kind of solar cell based on such wires that could significantly lower costs, both by allowing the use of lower grades of silicon (that is, less-highly refined), since the process of wire growth helps to purify the material, and by using much smaller amounts of it, since the tiny wires are made up of just a tiny fraction of the amount needed for conventional silicon crystal wafers. “This is still in a very early stage,” Buonassisi says, because in deciding on a configuration for such a solar cell “there are so many things to optimize.”

Michael Kelzenberg, a postdoctoral scholar at the California Institute of Technology who has spent the last five years doing research on silicon microwires, says that while others have used the copper-droplet technique for growing microwires, “What's really new here is the method of producing those liquid metal droplets.” While others have had to place the droplets of molten copper on the silicon plate, requiring extra processing steps, “Buonassisi and his colleagues have shown that metal can be diffused into the growth substrate beforehand, and through careful heating and cooling, the metal droplets will actually form on their own — with the correct position and size.”

Kelzenberg adds that his research group has recently demonstrated that silicon microwire solar cells can equal the efficiency of today’s typical commercial solar cells. “I think the greatest challenge remaining is to show that this technique is more cost-effective or otherwise beneficial than other catalyst metal production methods,” he says. But overall, he says, some version of silicon microwire technology “has the potential to enable dramatic cost reductions” of solar panels.

19

EEE / Turning windows into powerplants

« on: May 10, 2018, 01:07:44 PM »

If a new development from labs at MIT pans out as expected, someday the entire surface area of a building’s windows could be used to generate electricity — without interfering with the ability to see through them.

The key technology is a photovoltaic cell based on organic molecules, which harnesses the energy of infrared light while allowing visible light to pass through. Coated onto a pane of standard window glass, it could provide power for lights and other devices, and would lower installation costs by taking advantage of existing window structures.

These days, anywhere from half to two-thirds of the cost of a traditional, thin-film solar-power system comes from those installation costs, and up to half of the cost of the panels themselves is for the glass and structural parts, said Vladimir Bulović, professor of electrical engineering in the Department of Electrical Engineering and Computer Science. But the transparent photovoltaic system he developed with Richard Lunt, a postdoctoral researcher in the Research Laboratory of Electronics, could eliminate many of those associated costs, they say.

A paper by Bulović and Lunt describing their new system has been published online in the journal Applied Physics Letters, and will appear in a forthcoming issue of the print edition.

Previous attempts to create transparent solar cells have either had extremely low efficiency (less than 1 percent of incoming solar radiation is converted to electricity), or have blocked too much light to be practical for use in windows. But the MIT researchers were able to find a specific chemical formulation for their cells that, when combined with partially infrared-reflective coatings, gives both high visible-light transparency and much better efficiency than earlier versions — comparable to that of non-transparent organic photovoltaic cells.

In a new building, or one where windows are being replaced anyway, adding the transparent solar cell material to the glass would be a relatively small incremental cost, since the cost of the glass, frames and installation would all be the same with or without the solar component, the researchers say, although it is too early in the process to be able to estimate actual costs. And with modern double-pane windows, the photovoltaic material could be coated on one of the inner surfaces, where it would be completely protected from weather or window washing. Only wiring connections to the window and a voltage controller would be needed to complete the system in a home.

In addition, much of the cost of existing solar panels comes from the glass substrate that the cells are placed on, and from the handling of that glass in the factory. Again, much of that cost would not apply if the process were made part of an existing window-manufacturing operation. Overall, Bulović says, “a large fraction of the cost could be eliminated” compared to today’s solar installations.

This will not be the ultimate solution to all the nation’s energy needs, Bulović says, but rather it is part of “a family of solutions” for producing power without greenhouse-gas emissions. “It’s attractive, because it can be added to things already being deployed,” rather than requiring land and infrastructure for a whole new system.

Fine-tuning the cells

The work is still at a very early stage, Bulović cautions. So far, they have achieved an efficiency of 1.7 percent in the prototype solar cells, but they expect that with further development they should be able to reach 12 percent, making it comparable to existing commercial solar panels. “It will be a challenge to get there,” Lunt says, “but it’s a question of excitonic engineering,” requiring optimization of the composition and configuration of the photovoltaic materials.

The researchers expect that after further development in the lab followed by work on manufacturability, the technology could become a practical commercial product within a decade. In addition to being suitable for coating directly on glass in the manufacture of new windows, the material might also be coated onto flexible material that could then be rolled onto existing windows, Lunt says.

Using the window surfaces of existing buildings could provide much more surface area for solar power than traditional solar panels, Bulović says. In mornings and evenings, with the sun low in the sky, the sides of big-city buildings are brightly illuminated, he says, and that vertical “footprint” of potential light-harvesting area could produce a significant amount of power.

20

EEE / Speeding swarms of sensor robots

« on: May 10, 2018, 01:07:00 PM »

SHARE
COMMENT
Concerns about the spread of radiation from damaged Japanese nuclear reactors — even as scientists are still trying to assess the consequences of the year-old Deepwater Horizon oil spill — have provided a painful reminder of just how important environmental monitoring can be. But collecting data on large expanses of land and sea can require massive deployments of resources.

At the Institute of Electrical and Electronics Engineers’ International Conference on Robotics and Automation in May, MIT researchers will present a new algorithm enabling sensor-laden robots to focus on the parts of their environments that change most frequently, without losing track of the regions that change more slowly. At the same conference, they’ll present a second paper describing a test run of the algorithm on underwater sensors that researchers at the University of Southern California (USC) are using to study algae blooms.

The work of Daniela Rus, a professor of computer science and electrical engineering, and postdocs Mac Schwager and Stephen Smith (now an assistant professor at the University of Waterloo in Ontario), the algorithm is designed for robots that will be monitoring an environment for long periods of time, tracing the same routes over and over. It assumes that the data of interest — temperature, the concentration of chemicals, the presence of organisms — fluctuate at different rates in different parts of the environment. In ocean regions with strong currents, for instance, chemical concentrations might change more rapidly than they do in more sheltered areas.

Floor it

In its current version, the algorithm assumes that researchers already have a mathematical model of the rates at which conditions change in different parts of the environment. The algorithm simply determines how the robots should adjust their velocities as they trace their routes. For instance, given particular rates of change along a route, would it make more sense to make one pass in an hour, slowing down considerably in areas of frequent change, or to make four or five passes, collecting less detailed data but taking more regular samples?

“From a practical point of view, it seems like an easy problem,” says Calin Belta, an assistant professor of mechanical engineering, systems engineering and bioinformatics at Boston University, who was not involved in the research. But it turns out to be a monstrously complex calculation. “It’s very hard to come up with a mathematical proof that you can really optimize the acquired knowledge,” he adds.

The MIT researchers draw an analogy with dust accumulating on a floor — dust that’s cleared whenever a sensor passes nearby. Because environmental change occurs at different rates in different areas, the dust piles up unevenly. The researchers were able to show that, with their algorithm, the height of the piles of dust would never exceed some limit: Only so much change could occur in any area before the sensor would measure it.

Ups and downs

Although the MIT researchers’ algorithm is designed to control robots’ velocity, the first robots on which it was tested don’t actually have velocity controllers. USC researchers have been studying harmful algae blooms using commercial robotic sensors designed by the Massachusetts company Webb Research. Because the sensors are intended to monitor ocean environments for weeks on end, they have to use power very sparingly, so they have no moving parts. Each sensor is shaped like an airplane, with an inflatable bladder on its nose. When the bladder fills, the sensor rises to the surface of the ocean; as the bladder empties, the sensor glides downward.

The more rapidly the bladder fills and empties, the steeper the sensor’s trajectory up and down, and the longer it takes to traverse a given distance — so it’s possible to concentrate the sensor’s attention in a particular location. Working with colleagues in the USC computer science department, the MIT team developed an interface that allows ocean researchers to specify regions of interest by drawing polygons around them on a digital map and indicating their priority with a numerical rating. The new algorithm then determines a trajectory for the sensor that will maximize the amount of data it collects in high-priority regions, without neglecting lower-priority regions.

At the moment, the algorithm depends on either some antecedent estimate of rates of change for an environment or researchers’ prioritization of regions. But in principle, a robotic sensor should be able to deduce rates of change from its own measurements, and the MIT researchers are currently working to modify the algorithm so that it can revise its own computations in light of new evidence. “That’s going to be a hard problem as well,” Belta says. “But they have the right background, and they’re strong, so I think they might be able to do it.”

The researchers also envision that the algorithm could prove useful for fleets of robots performing tasks other than environmental monitoring, such as tending produce, or — in a more literal application of the vacuuming-dust metaphor — cleaning up environmental hazards, such as oil leaking from underwater wells.

21

EEE / Why Your GPS Receiver Isn’t Bigger Than a Breadbox

« on: May 09, 2018, 11:16:22 AM »

As I drive through the vineyard-covered hills of San Luis Obispo, Calif., the tiny Global Positioning System receiver in my phone works with Google Maps to alert me to upcoming turns. The app reassures me that I’ll arrive at my destination on time, in spite of a short delay for construction.

How different this trip would have been in the pre-GPS era, when the obscured road sign at one intersection would likely have sent me off track. I have a weak sense of direction, and getting lost—or worrying about getting lost—was a stressful part of my life for a long time.

This GPS-guided journey is taking me to Bradford W. Parkinson, the person who made GPS technology—a tool we now take for granted—come together. Parkinson is being awarded the 2018 IEEE Medal of Honor for leading the development of GPS and pushing its early applications.

22

EEE / Virtual Synchronous Machines: A unified interface for grid integration

« on: May 09, 2018, 11:14:13 AM »

Power systems are going through a paradigm change from centralized generation to distributed generation and further on to smart grids. More and more renewable-energy sources, electric vehicles, energy storage systems, and so forth are being connected to power systems through power electronic converters. Moreover, the majority of loads are expected to connect to the grid through power electronic converters as well. This article shows that these converters, either on the supply side or on the load side, can all be controlled to behave like virtual synchronous machines (VSMs) and possess the dynamics of synchronous machines, providing a unified interface for smart grid integration. Synchroconverter technology and its developments are the focus of this article because the mathematical model of synchronous machines is embedded in the controller of synchronverters to provide close imitation.

23

EEE / Wide-Bandgap Power Devices: Adoption Gathers Momentum

« on: May 09, 2018, 11:12:45 AM »

Abstract:
As power electronics continues to extend into renewable energy markets, smart grids, smart homes, transportation electrification, electric and hybrid electric vehicles (EV/HEVs), and other emerging industrial and medical applications, more designers are attracted to the wide-bandgap (WBG) power devices, e.g., silicon carbide (SiC) and gallium nitride (GaN). In the past few years, engineers have spent time understanding the virtues of these emerging power devices and their drawbacks, e.g., reliability, cost, and availability. Today, many analysts believe that engineers are transitioning from education mode to implementation mode. According to research firm Yole D?veloppement's technology and market analyst Hong Lin, "We are gradually going from the customer awareness and education stage to the customer trial and adoption stage. And this is especially true for SiC transistors."

24

EEE / Film Capacitors in Power Applications: Choices and Particular Characteristics Ne

« on: May 09, 2018, 11:11:58 AM »

The diverse collection of capacitor types has not changed much over recent years, but applications certainly have. In this article, we look at how capacitors are used in power electronics and compare the available technologies. Film capacitors are showing their advantages in upcoming applications such as electric vehicles, alternative energy power conversion, and inverters in drives. However, aluminum (Al) electrolytics are still important when energy storage density is the main requirement.

25

Faculty Sections / Control of Power Converters in AC Microgrids

« on: May 09, 2018, 11:03:11 AM »

The enabling of ac microgrids in distribution networks allows delivering distributed power and providing grid support services during regular operation of the grid, as well as powering isolated islands in case of faults and contingencies, thus increasing the performance and reliability of the electrical system. The high penetration of distributed generators, linked to the grid through highly controllable power processors based on power electronics, together with the incorporation of electrical energy storage systems, communication technologies, and controllable loads, opens new horizons to the effective expansion of microgrid applications integrated into electrical power systems. This paper carries out an overview about microgrid structures and control techniques at different hierarchical levels. At the power converter level, a detailed analysis of the main operation modes and control structures for power converters belonging to microgrids is carried out, focusing mainly on grid-forming, grid-feeding, and grid-supporting configurations. This analysis is extended as well toward the hierarchical control scheme of microgrids, which, based on the primary, secondary, and tertiary control layer division, is devoted to minimize the operation cost, coordinating support services, meanwhile maximizing the reliability and the controllability of microgrids. Finally, the main grid services that microgrids can offer to the main network, as well as the future trends in the development of their operation and control for the next future, are presented and discussed.

26

EEE / Understanding STAR-DELTA Starter !

« on: April 20, 2017, 04:41:16 PM »

27

EEE / Brushless DC Motor, How it works ?

« on: April 20, 2017, 04:37:05 PM »

28

EEE / DC Motor, How it works?

« on: April 20, 2017, 04:36:14 PM »

29

EEE / Re: EarthDay2017

« on: April 20, 2017, 04:24:18 PM »

Thank you sir for your reply

30

EEE / Facebook Aims to Remake Telecom With Millimeter Waves and Tether-tennas

« on: April 20, 2017, 03:59:04 PM »

A world of millimeter-wave networks, laid out by computer, crisscrossing cities and into the stratosphere, where cell phone towers can be easily replaced by tethered autonomous copters—that’s the telecommunications infrastructure of the future. So says Facebook’s Yael Maguire, head of the company’s Connectivity Lab.

Speaking at Facebook’s F8 developer conference in San Jose today, Maguire said that Facebook is aiming to bring down the cost of connecting by an order of magnitude, working to develop the building blocks of more flexible and extensible networks.

A big part of that strategy, he said, is millimeter wave communications. Near term, he says, are millimeter wave networks connecting cities, a simpler route to high capacity communications than extending fiber optic networks to every building.  The company’s project here is called Terragraph; it is, Maguire said, “designed to be consumer level in pricing and deployed by mobile operators.”

Facebook’s real secret sauce in city-level millimeter wave communications, Maguire explained, is its use of computer vision and artificial intelligence to analyze a city and determine possible paths for communications. It has also, he said, developed new technology for rerouting these networks when communications are interrupted, and is currently testing this technology in San Jose.

Further out are Facebook’s efforts to use its Aquila aircraft in the stratosphere to allow long-distance communications. In recent tests, Facebook engineers clocked 36 Gbps of data transfer over 13 km using a stationary millimeter-wave link and 16 Gbps of data transfer from the ground to a Cessna aircraft; the Cessna acted as a stand-in for the futuristic Aquila. Maguire admitted it could take 10 years before this communications technology becomes practical. (Facebook also achievedand 80 Gbps using an optical link to the Cessna.)

Nearer term for connectivity in rural areas, he said, is reducing the cell-phone tower to what Maguire calls its absolute essence: the Tether-tenna. He described this as a cable tethered to ground-based power and fiber optic networks, held up by an autonomous ultralight helicopter. There are a few things to work out, he said, “it’s a high voltage system, it has to survive high winds and lightning.” In tests, he reported, the Tether-tenna has run for 24 hours; the company expects to extend that to months.

Maguire also said the company is particularly interested in using its technology to create instant infrastructure for use in a crisis; long term, Aquila will be able to be fill that niche, but he sees Tether-tenna as also being useful—and just a few years away from implementation.