Breakthroughs and trends in the world of technology.
Reef Relief
Corralling a Coral Killer
With its brilliant colors and venomous spikes, the crown-of-thorns starfish, Acanthaster planci, resembles an evil creature from science fiction. The echinoderms cling to reefs and munch on fast-growing coral polyps in the waters of the Indo-Pacific region, where in small numbers they help keep the ecosystem in balance. Sometimes, however, their population explodes and they start gobbling every type of coral they can wrap their arms around. In Australia’s Great Barrier Reef, which has experienced a catastrophic coral die-off in recent years, the starfish is considered an invasive species—up there with bleaching and cyclone damage as a leading cause of destruction. Researchers at the Queensland University of Technology in Brisbane have come up with a cheap and simple underwater killer drone to cull the culprits. Operated by remote tablet, the RangerBot can move in all directions, helping it navigate unstable seas in every condition. It also can remain submerged three times as long as a human diver and take readings on coral bleaching, water quality, pest species, pollution, and siltation. Most important, however, the yellow submarine can scan a reef and autonomously identify a crown-of-thorns starfish with 99.4 percent accuracy. It then extends an arm and injects the invaders with vinegar or bile salts—substances that are toxic to the starfish but benign to coral. The killer ‘bot, funded by Google, is set for widespread release this year in the hope of eventually bringing a reef back to life.- Jennifer Pocock
©Getty Images
Solar Energy
Amped Up
You know how adding a supercharger can increase a car’s engine performance? Well materials scientists at the University of California–Los Angeles have developed a supercharging method to make commercial solar cells 20 percent more efficient. Led by Yang Yang, a professor of materials science, the UCLA team coated a CIGS cell—a device made from a mixture of copper, indium, gallium, and selenide—with a thin layer of perovskite, an inexpensive compound of lead and iodine. The base CIGS layer, which is about 2 microns thick, converts 18.7 percent of the sun’s energy into electricity. The addition of a single-micron layer of perovskite boosted the cell’s conversion rate to 22.4 percent. Yang’s team designed a nanoscale interface that joins the two layers and gives the device higher voltage, increasing its power output. A big benefit of the technology, Yang says, is that the perovskite coating could be easily and inexpensively incorporated into the existing manufacturing process. Yang is also confident the team’s design can be tweaked to approach 30 percent conversion efficiency, which truly would supercharge solar technology. – Thomas K. Grose
©Oszie Tarula/UCLA
Biochemistry
Hydrogen Plants
Scientists have long been intrigued by the notion of manufacturing hydrogen by mimicking the process plants use to produce energy from sunlight. Hydrogen is abundant in nature, renewable, and releases no greenhouse gases, so it has huge potential as a form of green energy—but current production methods are expensive and dirty. Photosynthesis splits the water plants absorb into oxygen and hydrogen, but it’s inefficient. That’s because the process produces only enough energy for a plant’s survival—only around 1 percent to 2 percent of what it could convert and store. Artificial photosynthesis exists but can’t be scaled up for industrial uses, as it relies on costly and toxic materials. But a team of chemists at the University of Cambridge has come up with a semiartificial method, a combination of man-made and natural components that greatly increases energy. It uses hydrogenase, an enzyme found in algae, to accomplish this. Over millennia, evolution switched off hydrogenase because plants don’t need it to survive. But the enzyme allows the hydrogen to split from oxygen in water and flow out as a gas. The Cambridge team built an electrochemical cell using a light-collecting biochemistry called photosystem II that provided the voltage necessary to activate the hydrogenase. The team says its discovery should spark more research into using artificial photosynthesis to produce cheap, clean fuel. – T. G.
©Katarzyna Sokól
Physics
Noodling Around
Hold a strand of spaghetti at both ends, and bend it until it breaks. Chances are you won’t end up with two pieces, but three or more. That’s because when uncooked pasta starts to bend, it does so at its natural center. When it finally breaks, there’s a “snap back” effect caused by vibrations within the noodle as it tries to return to its original shape, causing additional pieces to break. Is there a way to break spaghetti into just two pieces? Yes, says a team of MIT math researchers investigating the pasta puzzle. They found that if you strongly twist the stick while simultaneously bending it, it will break in two. To figure out why, they built a machine with two clamps to hold the pasta at each end. One rotated the strand by various degrees while the other slowly slid toward the other clamp to bend it. They tested hundreds of pieces of spaghetti and found that if it is twisted to almost 360 degrees, then slowly bent, it snaps into two pieces. The mathematical model the researchers devised found that twisting releases a counterwave that neutralizes the snap-back effect. The MIT experiment was more than a fun exercise in spaghetti engineering: It could help researchers develop ways to avoid fractures in other rod-shaped multifiber structures, including engineered nanotubes. As far as having use in the kitchen, let’s hope not. Everyone knows that breaking spaghetti before cooking it is a culinary faux pas. – T. G.
©MIT
Industrial Robotics
Flipped Off
Are fast-food cooks at risk of losing their jobs? Miso Robotics’s Flippy, an AI-enabled short-order cook at CaliBurger, in Pasadena, California, grilled up to 300 burgers an hour. That performance won it a call up to the majors when Dodger Stadium contracted Flippy to cook fried chicken and tater tots. Its mechanical arm comes outfitted with various tools, including a spatula, tongs, and a gripper. It knows when to flip and serve thanks to lasers, a thermal scanner, and camera. Miso claims that Flippy is designed to work with humans, not replace them. Flippy may not need a humanlike hand, but many robots do. Human hands are among the most complex designs in nature. Most robotic hands need many cables and motors to work and are thus expensive—one on the market now has 20 motors. Wired reports that researchers at the University of Pisa and the Italian Institute of Technology have come up with a cheaper device called SoftHand 2 that uses just one cable woven through a number of pulleys in each finger, and needs only two motors. SoftHand 2 may not be as dexterous as costlier machines, but it’s still capable of making quite a few maneuvers. It can, for instance, unscrew a jar. Maybe one day it will open ketchup bottles for Flippy. – T. G.
©Miso Robotics
Autism Research
Emotion App
Roughly 1 in 59 American children has autism, a developmental disorder often characterized by social and communication problems due to an inability to pick up on emotional cues conveyed in facial expressions. But researchers at Stanford University may have a novel solution: an app therapy using Google’s discontinued computing eyewear, Google Glass. The device has a tiny camera to record a wearer’s field of view, and a small screen and speaker to provide audio and visual information. As the child interacts with people, the app identifies and names the emotions they express through the speaker or on the screen. The trial of the so-called Superpower Glass app lasted an average of 10 weeks and included 14 participants, who ranged in age from 3 to 17. The researchers developed three ways to use the app: free play and two games designed to help the children recognize and better understand eight core expressions. After the study ended, 12 of the 14 children were better able to make eye contact with others. And six had large enough changes to move down one step in the severity of their autism classification. – T. G.
©Stanford Medicine
Espionage Technology
Aural Mysteries
Since late 2016, more than three dozen U.S. diplomats and family members in Cuba and China have suffered mysterious brain injuries: They hear loud and strange noises in their heads and suffer headaches, hearing loss, and vertigo. There were fears they were targeted by a sonic weapon. Though that was ruled out, the ailment’s cause remains undetermined. Now, the New York Times reports, some doctors think they were targeted by a microwave weapon. Researchers say that the so-called microwave auditory effect caused by high-intensity beams can cause similar symptoms. However, Kenneth Foster, a professor of bioengineering at the University of Pennsylvania and an expert on microwave phenomena, wrote in Scientific American: “I find the theory wildly implausible.” Foster said that it would be nearly impossible to build a powerful enough microwave weapon that could precisely target victims from a distance. Moreover, “the microwaves would have to be so intense they would actually burn the subject.” None of the victims was burned. Foster, however, speculated that listening devices using ultrasound could be the culprit. Ultrasound, often used in detectors and sensors, has been known to give some people weird aural sensations. Foster thinks it’s possible that the diplomats were sickened by ultrasound-based spy technology that someone used to eavesdrop but not to purposely harm anyone. – T. G.
©Getty Images
Imaging Technology
Optical Illustrations
Any soldier would jump at the chance to see around corners or behind walls. Now researchers are rapidly developing computer-vision technologies that will enable them to do so, according to Quanta magazine. Beyond military applications, they could also have other uses in autonomous cars, robotic vision, and space exploration. The technology took off six years ago after Antonio Torralba, an MIT computer-vision professor, realized that the faint “shadows” on his hotel room wall were actually upside-down images of the patio outside. The window was acting like a pinhole camera, passing light rays that created the image. Torralba and his colleague Bill Freeman later wrote a paper on “non-line-of-sight imaging,” in which they explained that windows, houseplants, and other everyday objects regularly create these subtle images. In fact, the world is awash with them—but the human eye usually can’t detect them. Computer algorithms, however, can. Researchers have built programs that can detect and analyze slight changes in light and color that can create images of objects not in the line of sight—including objects shrouded in fog and the first few pages of a closed book. Such imaging tricks are already in use. NASA, for instance, has a project underway to collect remote images inside lunar caves. – T. G.
©Getty Images
Construction Engineering
Corrosive Reaction
Last August, the 51-year-old Morandi Bridge in Genoa, Italy, collapsed, killing 43 people. Engineering experts believe the cause was corrosion. It was constructed from steel-reinforced concrete, a superstrong building material. But water can wreak havoc if cracks open up, allowing it to seep in and degrade the steel. The Economist, however, notes that new technologies are becoming available to inspect concrete before cracks become dangerous, including scanning drones, electronic sensors, and laser scanners that produce pinpoint 3-D images. One possible new solution was reported just days after the Morandi gave way. A team of materials engineers from Texas A&M University used new imaging tools at the Argonne National Laboratory’s Advanced Photon Source (APS) to take 3-D images of microscopic cracks as they opened up in metal damaged by hydrogen. Based on the images, the team identified metal-strengthening microscopic structures that reduce vulnerability to water damage. All metals consist of microscopic grains; in nickel superalloys, fractures caused by hydrogen follow boundaries between the grains. The APS beam allowed the researchers to watch the grain orientations and boundaries as the cracks progressed. This indicates that the APS tools could be used to image metal components and predict if they’re susceptible to failure. Ultimately, however, knowing which grain boundaries are resistant to cracking could help engineers design stronger metals. – T. G.
©Michele Ferraris
Water Treatment
Urban Wellsprings
Remember the water cycle, where rain seeps into the ground and keeps aquifers filled? Concrete stymies that process in urban areas, where stormwater picks up a host of contaminants, including pesticides, toxic metals, and fuels and oils. Cities typically send runoff into rivers and streams as fast as possible. But civil and environmental engineers at the University of California–Berkeley are developing a way to safely purify and direct stormwater into aquifers, where it can be stored until needed. They’ve created a mineral-coated “engineered sand” that reacts with and destroys organic pollutants. Passive, low-cost, and noninvasive, the material is a mix of plain sand and two types of manganese that react to form manganese oxide, a harmless mineral that binds to and breaks down toxic organic chemicals to safe levels. In the developing world, a serious problem with groundwater is arsenic contamination. Arsenic poisoning kills an estimated 43,000 people a year. Treatments to remove arsenic from water tend to be expensive, so few have been deployed. But researchers at Pakistan’s University of Agriculture Faisalabad have developed an inexpensive filter made from watermelon rinds, according to Scidev.net. The rinds are treated with xanthate salts, which attract and bind arsenic, to make a filter that can remove 95 percent of the poison from water. – T. G.
©University of California–Berkeley
Emergency Management
Practice Drill
The town is fake but Disaster City—a 52-acre training center operated by Texas A&M University—responds to an all-too-real need: better emergency response. Its collapsible buildings, wrecked cars and trains, and lots of rubble were designed to help urban search-and-rescue crews learn how to safely extract survivors from buildings and infrastructure destroyed by natural or man-made catastrophes. The Federal Emergency Management Agency, which oversees national disaster relief, recently awarded a five-year, $5.9 million contract to the Texas A&M Engineering Extension Service (TEEX), an arm of the university, to develop a curriculum of training exercises for all 5,800 members of the National Urban Search and Rescue Response System. TEEX also will maintain an online training portal. The Dallas Morning News reports that a big part of the training will focus on rescuing city dwellers from flooding and fast water—dangers that Hurricane Harvey’s historic rains unleashed on Houston last year. – T. G.
©TEEX
Medical Imaging
Spot Check
Lung cancer is the leading cause of cancer deaths in the United States in part because it’s often detected quite late, when survival rates are just 17 percent. Earlier detection could help improve those odds, but the minuscule specks of cancerous tissue picked up by CT scans aren’t easy to see. Radiologists typically spot them only 65 percent of the time. But computer-vision researchers at the University of Central Florida have developed an artificial-intelligence algorithm to analyze CT scans of lungs for early signs of tumors, and the accuracy rate is 95 percent. The algorithm works much in the same way as facial-recognition software, which learns by scanning thousands of images. The UCF team used more than 1,000 CT scans to train the software to look for small but telltale patterns in the images that indicate the presence of tumors. Additionally, the researchers taught the AI system to ignore healthy tissue, nerves, and other masses picked up in the scans. One of the team’s members is now working to further train the algorithm to tell the difference between benign and cancerous tumors. Another member is investigating whether the same system can be used to find or predict brain disorders. – T. G.
©Bo Lusk, TNC
Data Analytics
Crowd Control
Law-enforcement authorities who must keep control of large crowds may one day be able to predict how people will distribute themselves and whether their mood could turn ugly. Crowd interactions are hard to quantify mathematically because behavior patterns are so complex. So physicists at Cornell University applied concepts from density-functional theory (DFT), a Nobel Prize-winning method developed to study large collections of quantum mechanically interacting electrons. To test the theory, the team created a model system using Drosophila melanogaster, or walking fruit flies. They came up with formulas to quantify how much the flies liked different locations in an environment, a vexation factor, and whether they enjoyed being crowded together, a frustration factor. By keeping track of a single fly in new environments, the researchers were able to accurately predict how a crowd of flies would distribute themselves. And by tracking the frustration factor, they could anticipate mood shifts. The physicists are confident the model also will work with congregations of humans. It should be possible, for instance, to figure out from live video data how a crowd might configure itself, and to measure density fluctuations using a smartphone app to determine if tempers were flaring. – T. G.
©Getty Images