NVIDIA to Install Computers in Cars to Teach Them How to Drive

As reported by ITWorld: As thousands of dashcam videos on YouTube vividly demonstrate, drivers see the craziest things. Be it an angry bear, a low-flying aircraft or even a guy riding a shopping cart on the freeway, the videos make for entertaining viewing but also illustrate a problem facing developers of self-driving cars: how can you program a computer to make sense of all this?

On Tuesday, chip maker Nvidia introduced a $10,000 computer that it says will allow cars to learn the right and wrong reactions to different situations, essentially figuring out what to do from experience rather than a rigid set of pre-defined situations.

“Driving is not about detecting, driving is a learned behavior,” said Jen Hsun Huang, CEO of Nvidia, during a presentation at the company’s GTC 2015 conference in San Jose.

The Drive PX is based on two of the company’s Tegra X1 processors and will crunch video from up to 12 cameras. Over time it should learn, for example, to slow down for dogs but not slam on the brakes for a piece of newspaper blowing across the road.

Today’s commercial autonomous systems are largely related to detecting when cars stray from their lanes or preventing collisions. Several fully self-driving cars have been developed as part of research projects, but they rely on highly detailed maps and are generally restricted to operating in controlled environments.

A DARPA project already proved the learning technology on a lower level, said Huang. A small autonomous robot was fed with 225,000 images of a backyard. When it started out, the robot ran straight into an obstacle, but after analyzing the images, it managed to successfully scoot around the yard without hitting objects, figuring out for itself how to get around.

The Drive PX is intended to be used by auto makers in research and development projects and is unlikely to mean self-driving cars are coming anytime soon. But if it works as promoted, it could help advance their arrival.

One proponent of autonomous driving, Tesla Motors CEO Elon Musk, said the most difficult part of realizing the technology was at speeds between 10- and 50 miles per hour.

“It’s fairly easy to deal with things that are sub five or 10 miles per hour, you just make sure it hits nothing” said Musk, who was speaking alongside Huang at the event. “From 10 to 50 miles per hour in complex suburban environments, that’s when you can get a lot of unexpected things happening. Once you’re above 50 miles per hour, it gets easier again.”

An additional element of Drive PX will ensure that actions learned in one car are shared with others.

Nvidia didn’t say which auto makers would be using the platform, which will be available from May, but did say that it’s already receiving inquiries from car companies about the technology.

New Technology May Double Radio Frequency Data Capacity

As reported by Columbia Engineering: A team of Columbia Engineering researchers has invented a technology—full-duplex radio integrated circuits (ICs)—that can be implemented in nanoscale CMOS to enable simultaneous transmission and reception at the same frequency in a wireless radio. Up to now, this has been thought to be impossible: transmitters and receivers either work at different times or at the same time but at different frequencies. The Columbia team, led by Electrical Engineering Associate Professor Harish Krishnaswamy, is the first to demonstrate an IC that can accomplish this. The researchers presented their work at the International Solid-State Circuits Conference (ISSCC) in San Francisco on February 25.

“This is a game-changer,” says Krishnaswamy, director of the Columbia high-Speed and Mm-wave IC (CoSMIC) Lab. “By leveraging our new technology, networks can effectively double the frequency spectrum resources available for devices like smartphones and tablets.”

In the era of Big Data, the current frequency spectrum crisis is one of the biggest challenges researchers are grappling with and it is clear that today's wireless networks will not be able to support tomorrow's data deluge. Today's standards, such as 4G/LTE, already support 40 different frequency bands, and there is no space left at radio frequencies for future expansion. At the same time, the grand challenge of the next-generation 5G network is to increase the data capacity by 1,000 times.

So the ability to have a transmitter and receiver re-use the same frequency has the potential to immediately double the data capacity of today's networks. Krishnaswamy notes that other research groups and startup companies have demonstrated the theoretical feasibility of simultaneous transmission and reception at the same frequency, but no one has yet been able to build tiny nanoscale ICs with this capability.

“Our work is the first to demonstrate an IC that can receive and transmit simultaneously,” he says. “Doing this in an IC is critical if we are to have widespread impact and bring this functionality to handheld devices such as cellular handsets, mobile devices such as tablets for WiFi, and in cellular and WiFi base stations to support full duplex communications.”

The biggest challenge the team faced with full duplex was canceling the transmitter's echo. Imagine that you are trying to listen to someone whisper from far away while at the same time someone else is yelling while standing next to you. If you can cancel the echo of the person yelling, you can hear the other person whispering.

“If everyone could do this, everyone could talk and listen at the same time, and conversations would take half the amount of time and resources as they take right now,” explains Jin Zhou, Krishnaswamy’s PhD student and the paper’s lead author. “Transmitter echo or ‘self-interference’ cancellation has been a fundamental challenge, especially when performed in a tiny nanoscale IC, and we have found a way to solve that challenge.”

Krishnaswamy and Zhou plan next to test a number of full-duplex nodes to understand what the gains are at the network level. “We are working closely with Electrical Engineering Associate Professor Gil Zussman and his PhD student Jelena Marasevic, who are network theory experts here at Columbia Engineering,” Krishnaswamy adds. “It will be very exciting if we are indeed able to deliver the promised performance gains.”


This work was funded by the DARPA RF-FPGA program.

Could Human-Driven Cars Become Illegal?

As reported by Huffington Post: Self-driving cars might be a novelty today. But in the not-too-distant future, they could become common.

Eventually, autonomous cars might prove to be so much safer than human drivers that you won't even be allowed to take the wheel anymore, Tesla co-founder and CEO Elon Musk said on Tuesday.

"People may outlaw driving cars because it's too dangerous," Musk told NVidia CEO Jen-Hsun Huang at the company's GPU Technology Conference in San Jose, California, according to CNBC. "You can't have a person driving a two-ton death machine."

Musk later clarified on Twitter that he doesn't support outlawing human-driven cars -- only that he could envision it happening in the future. 

In any case, it would be a while before human drivers are completely replaced. Musk said there are 2 billion cars on the road, and automakers can make 100 million vehicles per year. That means it would take at least 20 years to replace every car with an autonomous one.

While Musk has in the past called artificial intelligence "our biggest existential threat," and compared it to "summoning a demon," he said on Tuesday that autonomous cars won't be that demon.

"That's sort of like a narrow form of AI," Musk said, according to The Verge. "It would be like an elevator. They used to have elevator operators, and then we developed some simple circuitry to have elevators just automatically come to the floor that you're at ... the car is going to be just like that."

SpaceX Boosting Output, On-Track for 13 Rocket Launches This Year

As reported by Reuters: Space Exploration Technologies, or SpaceX, is rapidly increasing production of the engines that power its Falcon 9 rocket and expects to meet its target of 13 launches and two test flights this year, President Gwynne Shotwell told Reuters.

SpaceX, the technology upstart founded by entrepreneur Elon Musk, is stepping up hiring of engineers and other workers to help boost production, including many from other sectors such as the automotive industry and the military, company officials said.

This year, the company expects to produce at least 180 engines, with that number set to increase to 240 next year, and 400 in 2017, Shotwell told Reuters in an interview late last week.

Shotwell said increasing production put the company on track to complete 13 launches this year. It fell short of its targets last year due to a number of factors.

"Certainly from a manufacturing perspective, we should be able to meet those targets," said Shotwell, who is due to testify before the House Armed Service Committee on Tuesday about a drive to end U.S. reliance on a Russian-built engine that powers one of two rockets used by SpaceX rival United Launch Alliance (ULA).

The Air Force expects to certify SpaceX by June to launch some military and intelligence satellites using its Falcon 9 rockets. Currently, those satellites can only be launched by ULA, a joint venture of Lockheed Martin Corp and Boeing Co, the two largest U.S. arms makers.

SpaceX has shaken up the satellite industry in recent years, winning a variety of launch contracts from commercial firms, as well as NASA, and putting pressure on ULA to lower its costs. But skeptics say the jury is still out on whether SpaceX can keep up with rising demand and growing backlog.

SpaceX has already launched three times this year and is gearing up for a fourth launch on March 21, followed by a cargo resupply mission for NASA in early April.

The company also has a prototype crew capsule at Cape Canaveral for a test flight to prove that a spaceship carrying astronauts could safely abort a mission if a rocket blew up on the launch pad, she said.

SpaceX plans a second test flight this year for NASA, aimed showing its ability safely land astronauts if a launch was aborted during flight.

Shotwell said the company was also making "great progress" on its 27-engine Falcon Heavy rocket, and planned to test it later this year at a refurbished space shuttle launch pad at Kennedy Space Center in Florida. 

NASA's Testing its 18-Engine Electric Plane Concept

As reported by Engadget: NASA's set to test a wing concept it says "may herald (the) future" of electric planes, but it almost looks like a joke -- it has one-third the wing area of a normal aircraft and 18 electric motors. However, the space agency is dead serious about the LEAPTech wing, a joint partnership with two private aerospace companies. It consists of a 31-foot, carbon composite span with tiny motors powered by lithium iron phosphate batteries. After successful testing at slower speeds, NASA will "fly" a wing section aboard a specially-equipped truck at speeds up to 70mph. Eventually, the wing will be mounted to a commercial Tecnam P2006T aircraft and flown by test pilots.

So, what's up with the crazy LEAPTech wing? According to inventor Joby Aviation, the thrust from all the motors and props increases the air velocity over the wing uniformly, drastically boosting lift. Each motor is independently controlled by a computer, allowing engineers to tailor speeds for optimal performance. All of that allows for a much smaller wing with reduced drag, which in turn delivers higher efficiency, faster speeds, a smoother ride and a lower noise signature. At the same time, a LEAPTech aircraft takes off and lands at the same speeds and distances as a normal plane.

The concept is part of NASA's plan to transition aircraft to electric propulsion within the next ten years. NASA said the technology "has the potential to achieve transformational capabilities in the near-term for (private) aircraft, as well as for transport aircraft in the longer-term." That said, electric planes suffer from the same range issues as electric cars, and NASA's wing doesn't look like it would fly at all without power. 

The space agency will no doubt have to thoroughly prove the LEAPTech wing concept before sending up test pilots.

An Autonomous Car Is Going Cross-Country for the First Time

As reported by Wired: Lots of people decide, at one point or another, to drive across the US. College kids. Beat poets. Truckers. In American folklore, it doesn’t get much more romantic than cruising down the highway, learning about life (or, you know, hauling shipping pallets). Now that trip is being taken on by a new kind of driver, one that won’t appreciate natural beauty or the (temporary) joy that comes from a gas station chili dog: a robot.

On March 22, an autonomous car will set out from the Golden Gate Bridge toward New York for a 3,500-mile drive that, if all goes according to plan, will push robo-cars much closer to reality. Audi’s taken its self-driving car from Silicon Valley to Las Vegas, Google’s racked up more than 700,000 autonomous miles, and Volvo’s preparing to put regular people in its robot-controlled vehicles. But this will be one of the most ambitious tests yet for a technology that promises to change just about everything, and it’s being done not by Google or Audi or Nissan, but by a company many people have never heard of: Delphi.

“It’s time to put our vehicle to the ultimate test by broadening the range of driving conditions,” says Delphi CTO Jeff Owens.

Delphi doesn’t build cars; it builds the stuff that goes into cars. It’s a key supplier to the auto industry, and has been for almost as long as there’s been an auto industry. It’s got a solid record of innovation, too. It built the first electric starter in 1911, the first in-dash car radio in 1936, and the first integrated radio-navi system in 1994.

Now it’s built a self-driving car, based on a 2014 Audi SQ5 (chosen simply because it’s cool. No, really.). The car looks like any other SQ5 (but for the stickers), but it’s packed with sensors and computers Delphi developed to replace humans: A camera in the windshield looks for lane lines, road signs, and traffic lights. Delphi installed a midrange radar, with a range of about 80 meters, on each corner. There’s another at the front and a sixth on the rear. That’s in addition to the long-range radars on the front and back, which look 180 meters ahead and behind.

This isn’t Delphi’s bid to start selling vehicles directly to consumers. It’s in the business of developing things automakers don’t want to (or can’t) develop themselves, and the rise of autonomous driving is a fertile field of opportunities. This market, including active safety features (which do things like keep you in your lane, adjust your speed on the highway, and brake before you hit that cyclist you didn’t see) is growing 35 percent every year. It made Delphi $1.4 billion in 2014, a number the company wants to grow by 50 percent year over year.

Building your own autonomous car is a good way to develop the hardware (radar and LIDaR) and software (the algorithms that make driving decisions) automakers will need. “What we expect to do is be able to create better sensors and more sensors, and then the software algorithms as well, which the [automakers] will need as they take more steps along that journey to automated driving,” says Owens.

So why the road trip? It’s about collecting data. Delphi says it’s covered hundreds of miles in the past year or so around Silicon Valley and Las Vegas, both on the highway and on city streets. Going from California to New York provides terabytes of information on how the sensor suite detects the world around it, and how the car drives. With that data, it can continue to improve its technology, tweaking software and hardware alike to make the car’s driving more reliable.

Delphi plans to make the trip in eight days, driving at most eight hours a day. The leisurely pace will keep everyone fresh, Owens says. Besides, the car will not be breaking the speed limit—just because Google does it doesn’t make it okay to speed—so some extra time is necessary. Sticking to a southern route and driving while the sun is up means better weather and conditions for the car’s sensors. When it’s not on the highway, one of the humans inside will take the wheel.

As far as skill goes, Owens says, “virtually anything you would do on the highway, the car will be capable of doing as well.” That means maintaining a steady speed and safe distance from other cars, and passing slower vehicles. If it gets cut off or a couch falls off a pickup truck right into its path, the car will do the smart thing: brake like hell, and move to the left or right if it’s safe.

If all goes well, the rolling catalog of automotive expertise will arrive in the Big Apple on the eve of the New York auto show, showing the public, and automakers, what the future holds.

Second Galileo FOC Satellite Reaches Corrected Orbit

As reported by Inside GNSS: The European Space Agency (ESA) announced March 13, 2015 that second Galileo full operational capability (FOC) satellite launched into the wrong orbit last August has now entered its corrected target orbit, which will allow detailed testing to assess the performance of its navigation payload.

Launched with another FOC spacecraft, its initial elongated orbit saw it travelling as high as 25,900 kilometers above Earth and down to a low point of 13,713 kilometers, confusing its onboard Earth sensor used to point satellite’s navigation antennas toward the ground.

A recovery plan was devised between ESA’s Galileo team, flight dynamics specialists at ESA’s ESOC operations center and France’s CNES space agency, as well as satellite operator SpaceOpal and satellite manufacturer OHB.

This involved gradually raising the lowest point of the satellites’ orbits more than 3,500 kilometer while also making them more circular.

The second FOC satellite — and fifth operational Galileo spacecraft, counting four in-orbit validation (IOV) — entered its corrected orbit at the end of November 2014. Both its navigation and search and rescue payloads were switched on the following month to begin testing.

Now the sixth satellite has reached the same orbit, too.

This latest salvage operation began in mid-January and concluded six weeks later, with some 14 separate maneuvers performed in total.

Its corrected position is effectively a mirror image of the fifth satellite’s, placing the pair on opposite sides of the planet. The exposure of the two to the harmful Van Allen Belt radiation has been greatly reduced, helping to ensure future reliability.

Significantly, the corrected orbit means they will overfly the same location on the ground every 20 days. This compares with a standard Galileo repeat pattern of every 10 days, helping to synchronize their ground tracks with the rest of the constellation.

The test results from Galileo 5 proved positive, with the same test campaign for the sixth satellite due to begin shortly, overseen by ESA’s Redu centre in Belgium. A 20 meter­­–diameter antenna will study the strength and shape of the navigation signals at high resolution.

“I am very proud of what our teams at ESA and industry have achieved,” says Marco Falcone, head of Galileo system office. “Our intention was to recover this mission from the very early days after the wrong orbit injection. This is what we are made for at ESA.”

The decision whether to use the two satellites for navigation and search-and-rescue purposes will be ultimately taken by the European Commission, as the system owner, based on the in-orbit test results and the system’s ability to provide navigation data from the improved orbits.

The next pair of satellites is due for launch on March 27.