Search This Blog

Thursday, February 6, 2020

A $1 Billion Initiative Aims to Bring EV Chargers to Highways and Rural Areas in the US

ChargePoint hopes to fill the gaps that prevent EV adoption.
As reported by Engadget: While TeslaElectrify America and others technically have nationwide EV charging networks, they don't really provide full coverage -- many rural areas are far from any kind of charger infrastructure. ChargePoint believes it can help close that gap, though. It's teaming with NATSO on a $1 billion effort to bring EV chargers to over 4,000 travel centers and truck stops (which NATSO represents) by 2030, with a particular focus on highways and rural areas. This could both spur EV adoption in rural towns and help with long-distance travel for everyone, ChargePoint said.

The two allies hope to make use of both "public and private" cash to support their initiative, including Volkswagen settlement funds.

There's not much mystery as to why ChargePoint and NATSO are willing to spend on EV chargers. ChargePoint could corner an underserved market and reap the rewards if and when electric cars dominate. For NATSO, meanwhile, this may be a matter of survival. Many travel centers and truck stops are built on the assumption drivers are stopping for gas -- they could lose much of their business if people have few good reasons to make pit stops.

Monday, February 3, 2020

What's Behind Tesla's Neural Network For Autopilot And Self-Driving?


As reported by InsideEVsTesla Autopilot may be considered a breakthrough, but there are still concerns about semi-autonomous and fully autonomous vehicle technology. The concerns surrounding Tesla stem from the fact that it already has the tech employed in its cars, albeit in beta form. However, due to this real-world testing, Tesla has recorded over 3 billion miles of Autopilot data for use in its research.

The Autopilot section of Tesla's website explains the technology via several categories, including hardware, neural networks, autonomy algorithms, code foundation, and evaluation infrastructure. We'll be first to admit that we don't have a solid grasp of AI and neural networks, but the information is definitely fascinating. Even more interesting is the latest video, seen below.



The above video is also shown on Tesla's website, along with the section on neural networks. According to Tesla:

"Our per-camera networks analyze raw images to perform semantic segmentation, object detection and monocular depth estimation. Our birds-eye-view networks take video from all cameras to output the road layout, static infrastructure and 3D objects directly in the top-down view. Our networks learn from the most complicated and diverse scenarios in the world, iteratively sourced from our fleet of nearly 1M vehicles in real time. A full build of Autopilot neural networks involves 48 networks that take 70,000 GPU hours to train. Together, they output 1,000 distinct tensors (predictions) at each timestep."


Wednesday, January 29, 2020

US Colleges are Looking to Install Location Tracking Apps on Students’ Phones

Supposedly an easier way to track attendance
As reported by The VergeBarely over a year ago, we pointed out how dystopian it seemed when Chinese schools added “smart uniforms” to track their students’ attendance. But US colleges are already testing out a similar tactic with a location tracking app, which students are now apparently expected to install on their phones.

I say “apparently” because there’s some confusion over whether the schools are actually forcing this on their students. The Kansas City Star reported that at the University of Missouri, new students “won’t be given a choice” of whether to install the SpotterEDU app, which uses Apple’s iBeacons to broadcast a Bluetooth signal that can help the phone figure out whether a student is actually in a room.

But a university spokesperson told Campus Reform on Sunday that only athletes are technically required to use the app, and a new statement from the university on Monday not only claims that it’s “completely optional” for students, but that the app’s being piloted with fewer than 2 percent of the student body.

What the reports do agree on: the app uses local Bluetooth signals, not GPS, so it’s probably not going to be very useful to track students outside of school. “No GPS tracking is enabled, meaning the technology cannot locate the students once they leave class,” reads part of the university’s statement.

SpotterEDU isn’t just used at the University of Missouri, though — it’s being tested at nearly 40 schools, company founder and former college basketball coach Rick Carter told The Washington Post in DecemberThe Post’s story makes it sound remarkably effective, with one Syracuse professor attesting that classes have never been so full, with more than 90 percent attendance. But that same professor attested that an earlier version of the app did have access to GPS coordinates, if only for a student to proactively share their location with a teacher.

And Spotter isn’t the only company marketing this idea to administrators: another startup, Degree Analytics, uses Wi-Fi signals instead of Bluetooth to serve an additional 19 schools, the Post reports. In September, The New York Times wrote about a similar app from a company called FanMaker that provides “loyalty points” to students who stick around to watch college sports games at the stadium instead of skipping out. That app is in use at 40 schools, the Times wrote.

It doesn’t seem like any of these specific systems are particularly invasive, and it currently sounds like (most) students will be able to opt out. But it also sounds like the idea of tracking students’ locations is being quietly normalized, in a way that smacks of surveillance (compare to how some previous pilot programs attempted to track students equipped with RFID-embedded ID cards).

It’s not unthinkable that future apps might tell schools more about students’ behavior, and that it may become harder to say no.


Thursday, January 9, 2020

Intel announces AI breakthroughs at CES


As reported by IoT M2M CouncilIntel CEO Bob Swann (pictured) kicked off this week’s Consumer Electronics Show in Las Vegas by announcing breakthroughs in artificial intelligence (AI) that pave the way for autonomous driving, a new era of mobile computing innovation, and the future of immersive sports and entertainment.

Intel demonstrated all these and more at CES by showing how the company is infusing intelligence across the cloud, network, edge and PC, and driving positive impact for people, business and society.
 
Swan shared updates from its Mobileye business, including a demonstration of its self-driving robo-car navigating traffic in a natural manner. The drive demonstrated Mobileye's approach to deliver safer mobility for all with a combination of artificial intelligence, computer vision, the regulatory science model of RSS (responsibility-sensitive safety) and redundancy through independent sensing systems.
 
Swan also highlighted Intel's work with the American Red Cross and its Missing Maps project to improve disaster preparedness. Using integrated AI acceleration on second-generation Xeon scalable processors, Intel is helping the American Red Cross and its Missing Maps project to build highly accurate maps with bridges and roads for remote regions of the world, which helps emergency responders in the event of a disaster.
 
"At Intel, our ambition is to help customers make the most of technology inflections like AI, 5G and the intelligent edge so that together we can enrich lives and shape the world for decades to come,” said Swan. “As we highlighted today, our drive to infuse intelligence into every aspect of computing can have positive impact at unprecedented scale."
 
Mobile computing was an area of emphasis, as Intel made announcements spanning products, partnerships and platform-level innovations to transform the way people focus, create and engage.
 


Intel executive vice president Gregory Bryant (pictured above) gave a first look and demonstration of the latest Intel Core mobile processors, code-named Tiger Lake. Tiger Lake is designed to bring Intel's people-led vision for mobile computing to lif.
 
With optimizations spanning the CPU, AI accelerators and discrete-level integrated graphics based on the Intel Xe graphics architecture, Tiger Lake should deliver double-digit performance gains, AI performance improvements, a leap in graphics performance and four times the throughput of USB 3 with the integrated Thunderbolt 4. Built on Intel's 10nm+ process, the first Tiger Lake systems are expected to ship this year.
 
 
 
Intel vice president of architecture for graphics and software Lisa Pearce (pictured above) provided insight into the progress on the Intel Xe graphics architecture, which will provide performance gains in Tiger Lake, and previewed Intel's first Xe-based discrete GPU, code-named DG1.

Updates were announced on Intel's Project Athena innovation program, including the first Project Athena-verified Chromebooks. Project Athena-verified designs have been tuned, tested and verified to deliver system-level innovation and benefits spanning battery life, consistent responsiveness, instant wake, application compatibility and more.
 
Intel has verified 25 Project Athena designs to date, and Bryant announced an expanded partnership with Google that has already resulted in the first two Project Athena-verified Chromebooks, the Asus Chromebook Flip (C436) and the Samsung Galaxy Chromebook. Intel expects to verify approximately 50 more designs across Windows and Chrome this year and deliver a target specification for dual-screen PCs.
 
Through deepened co-engineering efforts with OEM partners, Intel helps deliver category-defining devices based on Intel Core processors. This includes dual-screen and fold-able designs such as the Lenovo ThinkPad X1 Fold, which leverages the Intel Core processor with Intel Hybrid Technology (code-named Lakefield) expected to ship midyear, and the Dell Concept Duet. Bryant also previewed the company's latest concept device, a fold-able OLED display form factor, code-named Horseshoe Bend.
 
 
 
The data center is the force that delivers intelligence to businesses around the world and Xeon scalable processors continue to be the foundation of the data center. Intel executive vice president Navin Shenoy (pictured above) announced that third-generation Xeon scalable processors, coming in the first half of 2020, will include DL boost extensions for built-in AI training acceleration, providing up to a 60% increase in training performance over the previous family.

Shenoy highlighted several ways Intel is threading intelligence into data platforms across cloud, network and edge and how this is transforming sports and entertainment. For example, Netflix has used the latest video compression technology, AV1, to enhance its media streaming services and bring content to life across the globe, with up to 60% compression efficiency over the previous compression technology.
 
Intel and Netflix's joint efforts continue with the development of an open-source high-performance encoder (SVT-AV1), optimized on second-generation Xeon scalable processors, that delivers quality and performance gains making it viable for commercial deployment.
 
A claimed first-of-its-kind in computer vision, 3D Athlete Tracking (3Dat) uses AI to enhance the viewing experience with near real-time insights and visualizations. 3Dat uses highly mobile cameras to capture the form and motion of athletes, then applies algorithms optimized with DL boost and powered by Xeon scalable processors to analyse the bio-mechanics of athletes' movements.
 
Shenoy announced that this technology would enhance replays of the 100m and other sprinting events at the Olympic Games in Tokyo later this year.
 
Intel and the sports industry are transforming the sports viewing experience with volumetric video, a progression towards enabling sports viewing without limitations. Intel True View synthesizes the entire volume of a stadium’s field to provide endless angles that allow fans to choose any vantage point and player perspective and stream from their devices.
 
Intel and the NFL showcased the power of streaming volumetric video with a play from the Cleveland Browns versus Arizona Cardinals game. The data produced from the first quarter of an NFL game alone reach beyond 3Tbyte per minute.

Wednesday, January 8, 2020

CES: Under the Mighty Propellers of Hyundai's Flying Taxi

As reported by EngadgetBy law, an automaker must mention "mobility" during any press event. Hyundai knows this but instead of just talking about how their cars can be configured to drive us around in the future, they went full-on flying machine. In a partnership with Uber, they announced the SA-1 air taxi at a press event. Sadly, at said press gathering what we saw was a tiny model. Fortunately, the full-size model was in the booth and we checked it out.

The SA-1 is far larger than anticipated. The four-passenger mock-up vehicle is impressive in its size and thanks to redundant propellers, it seems like a viable air transportation system. While Hyundai is known primarily for its automotive manufacturing, the company seems keen to branch out into aircraft.



The issue is that the company is honing in on Bell and Airbus' turf. Both aircraft companies have air taxis in development. It's going to be tough to beat established players in the market but Hyundai looks like it's going to at least try.



Wednesday, December 11, 2019

Verizon CEO emphasizes C-band for 5G


As reported by FierceWirelessFor its 5G strategy, Verizon has all the spectrum it needs to accomplish its stated near-term goals, but for the nation as a whole, C-band spectrum—3.7-4.2 GHz—will be vitally important to achieve the kinds of economies of scale the industry needs.

That’s according to Verizon CEO and Chairman Hans Vestberg. “When it comes to CBRS, we are already working on that,” he said during UBS’ Global TMT Conference in New York on Tuesday. “We think that’s a good addition,” in the area of capacity.

Vestberg could not comment on the current millimeter wave auction that’s underway at the Federal Communications Commission (FCC) due to the quiet period that all participants much adhere to, but he stressed the importance for the U.S. to have access to mid-band spectrum.

With C-band, what’s important is it’s going to be global spectrum, which means all handsets, equipment and applications will be developed to work on it, so it offers enormous economies of scale. For the U.S., it’s very important to be on that mid-band bandwagon, he said.

It affects the cost of handsets and chips, and it’s one of the few spectrum bands that are available on a global scale. “I would say from that point of view, it’s important,” and as a leader in the market, Verizon is supportive of the C-band coming to market.

Vestberg met with FCC Chairman Ajit Pai last month where they discussed the importance of mid-band spectrum as the U.S. seeks to win the global race to 5G. They also discussed the need for speed and certainty concerning the mid-band spectrum that will be made available, according to an ex parte filing (PDF).

Fixed wireless as 5G use case
For its 5G Home service, Verizon wants to get to a point where the majority of consumers are ordering the equipment online and provisioning the service at home themselves, i.e., through self-installment, so that will cut down on the expense of sending installers to customers’ homes and improve the experience for customers rather than requiring them to be home, say from 9 a.m. to 5 p.m., for a an installer to visit. “It’s a totally new way of thinking from an experience point of view,” Vestberg said.

What it’s waiting for is the next generation of chipset, due in the second half of 2020, that will be the basis for its next CPE because it offers better power output. The target for 5G Home is still to reach 30 million households; there’s no date associated with that but Vestberg reiterated that’s still the goal.

In the markets where it has 5G Home, it’s competing well and its “win share” is good. It’s guaranteeing 300 Mbps, and the usage is “enormous,” according to the CEO. It still has some work to be done on the self-setup; it probably takes a couple of hours and he wants it to be shorter than that. “I want it to be a totally unique experience,” he said.   

Last week, Verizon announced it’s teaming with Amazon Web Services (AWS) on mobile edge compute (MEC), and using Amazon’s new Wavelength cloud platform. The first site in that deal was in Chicago and Bethesda Softworks, a video game publisher, was one of the first customers. It can’t have its games on 4G because there’s too much latency and throughput is too low, but it can publish its games using 5G with MEC.

In 2020, Verizon will open more MEC centers, with Amazon and in Verizon’s own facilities. AWS will be putting infrastructure in Verizon premises, and “it’s going to be many,” Vestberg said.

Verizon has worked to build a virtualized 5G network, and it doesn’t intend to try to be the cloud company. “We have the network, we have the distribution and we have the brand,” and Amazon doesn’t have these in the 5G space, but it has the cloud and developers, so “it’s a win-win for both,” he said. 


Satellite Operators lose battle for private C-Band Auction worth Billions
FCC Chairman Ajit Pai informed Congress Nov. 18 that the agency will run a public auction of C-band spectrum instead of allowing a consortium of satellite operators to sell it directly to 5G wireless operators. 

Satellite operators Intelsat, SES and Telesat, acting as the C-Band Alliance, had been lobbying the Federal Communications Commission for more than a year to approve a private auction of C-band spectrum expected to fetch upwards of $60 billion in proceeds. 

The C-Band Alliance’s proposal received a positive reception at the FCC, where Pai and Commissioner Michael O’Rielly had voiced support for a market-based approach to clearing the spectrum.

But in the past several months, several House members and Sen. John Kennedy (R-La.), chairman of the Senate subcommittee that oversees the FCC, pressed the FCC to run the auction instead. 

In an effort to appease lawmakers opposed to a private auction, the C-band Alliance said its members would contribute billions of dollars to the U.S. treasury and help create a program for rural 5G networks if the FCC allowed its plan to proceed.

The C-Band Alliance’s efforts to placate congressional critics of a private C-Band auction appear to have fallen short. 

In a Nov. 18 letter sent to more than a dozen members of Congress, Pai said that the FCC will publicly auction 280 megahertz of C-band that satellite operators use mainly for television and radio programming so that it can be re-purposed for cellular 5G networks. 

“After much deliberation and a thorough review of the extensive record, I have concluded that the best way to advance these principles is through an auction of 280 megahertz of the C-band conducted by the Federal Communications Commission’s excellent staff,” Pai wrote. “With a quarter century track record of transparent and successful auctions, I am confident that they will conduct a public auction that will afford all parties a fair opportunity to compete for this 5G spectrum, while preserving the availability of the upper 200 megahertz of this band for the continued delivery of programming.

The C-Band Alliance did not outline what recourse it intends to pursue in light of the FCC’s decision, but said “the full cooperation of the satellite operators will be required” to facilitate a spectrum transition.  

This could mean a significant delay in the marketplace (due to court action and logistical transfer and testing of spectrum) preventing Verizon and other 5G wireless carriers from using the critical C-Band spectrum in the near future.

First Commercial Electric Airplane Completes Test Flight


As reported by Engadget: An aircraft designed in 1946 might become the first commercial electric plane, following a short but successful test flight. Vancouver-based Harbour Air's took its "eBeaver" on a ten-minute hop on the Fraser River in Richmond, BC, with CEO and pilot Greg McDougall at the controls. "Today, we made history," he said in a statement. Harbour Air expects the eBeaver to go into commercial service in 2022.

The eBeaver is a highly modified version of de Havilland's legendary DCH-2 Beaver, equipped with a 750 horsepower electric motor from Redmond, Washington-based Magnix. Harbour Air announced its partnership with Magnix earlier this year and said it planned to build the world's first "completely electric commercial seaplane fleet."

So far, electric planes have failed to make much of a dent. Quite a number have been built, tested and even sold, but none are close to being ready to take paying passengers. "This is real," Magnix CEO Roei Ganzarski told Fortune. "This is an airline flying their own aircraft."

Harbour currently has 14 six-passenger DHC-2 Beaver aircraft, many of which are equipped with Pratt & Whitney PT-6A turbine engines that burn about $300 worth of jet A fuel per hour. By contrast, the eBeaver packs enough battery life to fly about 100 miles at a cost of around $10 to $20 worth of electricity.

E-planes have a very limited range compared to ICE-powered models because lithium-ion batteries have less than 5 percent the energy density of gasoline or jet fuel. However, 100 miles is enough for many of the short seaplane hops around Vancouver's lower mainland. The distance between Vancouver and British Columbia capital Victoria (downtown to downtown) is 58 miles and takes about 30 minutes by plane, while the same trip on a ferry can run over four hours including driving time and waiting. (Also, as your author can attest from brutal experience, the flight is a lot less boring.)

Despite the range challenges, electric planes have big advantages over ICE-powered models. That includes lower maintenance and operating costs, no need for fueling infrastructure (other than chargers) and easier boarding on local routes. "We are proving that low-cost, environmentally friendly, commercial electric air travel can be a reality in the very near future," said Ganzarski.