Search This Blog

Tuesday, February 25, 2014

A Plan to Lower Speed Camera Error Rates By Reversing the Incentives

As reported by The Atlantic CitiesBaltimore's particular speed camera problem first came to light in 2012, when the Baltimore Sun revealed that at least seven of the city's 83 radar cameras, all of them owned and operated by Xerox State and Local Solutions, were prone to issuing fines to drivers who were not exceeding the speed limit. Xerox itself claimed it found only five cameras that didn't work, and shut them down. The city, meanwhile, downplayed the problem even further, claiming the error rate for Xerox speed cameras was "one-quarter of one percent." In short: Nothing to see here!

Xerox's contract with Baltimore ended in 2012, but the deal is making headlines again thanks to a recent audit showing the company's cameras performed worse than even the Sun realized. The big takeaway? That error rate of "one-quarter of one percent"—promoted by city officials!—was actually upwards of 10 percent; 26 percent of issued citations were "questionable." 
The Sun, which first reported on the leaked audit last month, explains the ramifications: "The city issued roughly 700,000 speed camera tickets at $40 each in fiscal year 2012. If 10 percent were wrong, 70,000 would have wrongly been charged $2.8 million." And that's the low-end projection for how much Baltimore and Xerox may have bilked from citizens. 
But wait, it gets worse. The administration has also refused a request from the city council to officially release the audit, conducted by URS Corp. at a cost of $278,000, because doing so would violate a contract with Xerox that prohibits Baltimore from "referring or relating to, or reflecting, each party's internal considerations, discussions, analyses, and/or evaluations of issues raised during the settlement discussions."
Baltimore's speed camera fiasco does at least have something resembling a silver lining. In December 2013, the city announced it would no longer engage in revenue sharing with traffic camera vendors—a practice that Maryland Governor Martin O'Malley has decried as a "bounty" system. While Xerox got a cut of each fine its cameras issued, the decision actually resulted fromproblem cameras owned and operated by Brekford, the company Baltimore brought in to replace Xerox. 
But Maryland legislators aren't content to see Baltimore simply abandon the bounty system (which is supposedly illegal under state law anyway). They want to completely flip the incentives for camera operators: instead of paying companies for each citation they issue, pending legislation would require operators to be fined $1,000 every time they issue a citation in error. "This gives the vendors great incentive to make sure that they have done their homework," says Baltimore County Delegate Jon Cardin, the bill's sponsor.
You'd have to build a fine like that into any vendor contract, which could scare away companies (Xerox, Brekford) that have histories of fleecing drivers. A flat fee to vendors combined with penalties for faulty citations might even mean cities would be unable to find a company to operate speed cameras. Considering that traffic cameras are mostly for revenue generation (despite promising that the Xerox contract would reduce speeding in Baltimore, the annual haul from the cameras increased every year), that probably wouldn't be a bad thing. But insofar as there's a case for speed cameras, eradicating the incentive to wrongly ticket good drivers should clearly be part of it. 

Approaching Navigation Nirvana: GNSS Antenna Module Supports All Satellites

As reported by M2M World NewsSwiss-based u blox, a global leader in cellular and positioning modules and integrated circuits, introduces the CAM-M8Q GPS/GLONASS/BeiDou/QZSS antenna module. 

The module integrates a u-blox M8 satellite receiver IC plus SAW filter, LNA, TCXO, RTC, passives and a pre-tuned GNSS chip antenna in a tiny 9.6 x 14.0 x 1.95 mm package. Just add power for instant stand-alone global positioning!
Combining low power consumption with high-sensitivity, high jamming immunity and concurrent GNSS operation (GPS/GLONASS, GPS/BeiDou, or GLONASS/BeiDou) the surface-mount CAM-M8Q is truly a drop-in solution for reliable and accurate satellite positioning anywhere in the world.
“Our CAM-M8Q is perfect for customers designing highly compact products who want to speed up product development while freeing resources for core activities,” explains Thomas Nigg, VP Product Marketing at u-blox.
CAM-M8Q is ideal for a wide range of applications such as personal locators, handheld navigators, wearable electronics as well as vehicle telematics systems used for emergency call, anti-theft, insurance and road pricing. Consistent omnidirectional antenna performance insures excellent performance regardless of module orientation.
It also has a feature whereby the internal chip antenna can be used as a backup antenna if the design makes use of an external antenna. This is an important feature for applications where there is a risk that the primary external antenna may malfunction or suffer damage, for example in vehicle tracking systems where damage to the external antenna is possible.
The module uses the latest u-blox M8 GNSS receiver chip qualified according to AEC-Q100 and is manufactured in ISO/TS 16949 certified sites. Qualification tests are performed as stipulated in the ISO16750 standard: “Road vehicles – Environmental conditions and testing for electrical and electronic equipment”.
The CAM M8Q is form-factor compatible to predecessor modules UC530 and UC530M, allowing the upgrade of existing designs with minimal effort.
Now, if we can get ublox to add in an integrated eLORAN receiver, and MEMS (micro-mechanical sensors that include an accelerometer and gyroscope for short term inertial navigation) - we'd have a nearly perfect tracking product - potentially with jamming detection.  They already have the MEMS interface available in a different module (the 3D ADR), so they are very close to reaching navigation nirvana.

Monday, February 24, 2014

LEAP: Microsoft Tries To Re-Invent GPS With Cloud-Based Processing

As reported by The RegisterMicrosoft has released a tool that it hopes will make it possible to add location-sensing abilities to many more devices.

The Microsoft Cloud-Offloaded GPS Experimental Kit is part of an effort to gather location data from GPS satellites without using much power. That's important because, as Microsoft explains, few applications drain a mobile device's battery more rapidly than contacting satellites. Redmond therefore wonders what might happen if devices “log just enough raw GPS signal for postprocessing” in the cloud, instead of doing everything on the mobile device itself.

Microsoft’s even cooked up a cloud service, dubbed Low Energy Assisted Positioning (LEAP), to do that work. In this paper, the researchers explain it works as follows:  “The GPS receiver only needs to be on for less than a second to collect the submillisecond level propagation delay for each satellite's signal.

With a reference to a nearby object, such as a cell tower, the LEAP server can infer the rest of the information necessary to perform GPS position calculation.”

Microsoft's also created hardware to help the experiment along. Detailed here and dubbed CLEO, the device comprises a GPS receiver, battery and SD card slot. “Compared to more than 30 seconds of heavy signal processing on standalone GPS receivers, we can achieve three orders of magnitude lower energy consumption per location tagging,” Microsoft says.

CLEO devices cannot go online: users must gather their SD cards to upload data into a PC. But there's nothing to stop a CLEO being attached to something moving, where it will presumably last longer than conventional location-tracking devices thanks to its skimpy data-generating design.

At a time when plenty of apps offer real-time data acquisition and processing, this experiment may not seem particularly sophisticated. But with a bit more polish applied, Microsoft could conceivably find itself with a very handy addition to mobile device operating systems. And with a little lateral thinking it could find itself with a very interesting device and service combination to fuel its aim of becoming a devices and services company.

One small wrinkle suggests itself: Google got itself into an awful lot of trouble by geolocating WiFi modems. Material about this project mentions using known pieces of infrastructure like cell towers to assist the location process, but neglects to say how Redmond came by or intends to source that data. Surely Microsoft won't go where even Google now fears to tread?

Saturday, February 22, 2014

New, Inexpensive Solar Materials Boost Promise Of Hydrogen Fuel

As reported by the University of Wisconsin-Madison News: Generating electricity is not the only way to turn sunlight into energy we can use on demand. The sun can also drive reactions to create chemical fuels, such as hydrogen, that can in turn power cars, trucks and trains.

The trouble with solar fuel production is the cost of producing the sun-capturing semiconductors and the catalysts to generate fuel. The most efficient materials are far too expensive to produce fuel at a price that can compete with gasoline.
"In order to make commercially viable devices for solar fuel production, the material and the processing costs should be reduced significantly while achieving a high solar-to-fuel conversion efficiency," says Kyoung-Shin Choi, a chemistry professor at the University of Wisconsin-Madison.
Photo: Kyoung-Shin Choi
Kyoung-Shin Choi
In a study published last week in the journal Science, Choi and postdoctoral researcher Tae Woo Kim combined cheap, oxide-based materials to split water into hydrogen and oxygen gases using solar energy with a solar-to-hydrogen conversion efficiency of 1.7 percent, the highest reported for any oxide-based photoelectrode system.
Choi created solar cells from bismuth vanadate using electrodeposition — the same process employed to make gold-plated jewelry or surface-coat car bodies — to boost the compound's surface area to a remarkable 32 square meters for each gram.
"Without fancy equipment, high temperature or high pressure, we made a nanoporous semiconductor of very tiny particles that have a high surface area," says Choi, whose work is supported by the National Science Foundation. "More surface area means more contact area with water, and, therefore, more efficient water splitting."
Bismuth vanadate needs a hand in speeding the reaction that produces fuel, and that's where the paired catalysts come in.
While there are many research groups working on the development of photoelectric semiconductors, and many working on the development of water-splitting catalysts, according to Choi, the semiconductor-catalyst junction gets relatively little attention.
"The problem is, in the end you have to put them together," she says. "Even if you have the best semiconductor in the world and the best catalyst in the world, their overall efficiency can be limited by the semiconductor-catalyst interface."
“Without fancy equipment, high temperature or high pressure, we made a nanoporous semiconductor of very tiny particles that have a high surface area.”
Kyoung-Shin Choi
Choi and Kim exploited a pair of cheap and somewhat flawed catalysts — iron oxide and nickel oxide — by stacking them on the bismuth vanadate to take advantage of their relative strengths.
"Since no one catalyst can make a good interface with both the semiconductor and the water that is our reactant, we choose to split that work into two parts," Choi says. "The iron oxide makes a good junction with bismuth vanadate, and the nickel oxide makes a good catalytic interface with water. So we use them together."
The dual-layer catalyst design enabled simultaneous optimization of semiconductor-catalyst junction and catalyst-water junction.
"Combining this cheap catalyst duo with our nanoporous high surface area semiconductor electrode resulted in the construction of an inexpensive all oxide-based photoelectrode system with a record high efficiency," Choi says.
She expects the basic work done to prove the efficiency enhancement by nanoporous bismuth vanadate electrode and dual catalyst layers will provide labs around the world with fodder for leaps forward.
"Other researchers studying different types of semiconductors or different types of catalysts can start to use this approach to identify which combinations of materials can be even more efficient," says Choi, whose lab is already tweaking their design. "Which some engineering, the efficiency we achieved could be further improved very fast."

Friday, February 21, 2014

The Future of Photography: Drones

As reported by The AtlanticDrones are being used to film ski and snowboarding events at the Winter Olympics in Sochi, as you may have noticed. But the use of unmanned aerial vehicles for sports photography is far from a passing gimmick. In fact, you should expect more and more athletic events to be filmed by drone.

Remo Masina isn't involved in shooting the Olympics, but he's already familiar with drones' finesse for filming winter sports: He uses them to film skiers for commercials. The drones are quieter and cheaper than a manned helicopter (though they can still cost up to $40,000), he told the Associated Press, and they allow the filmmaker to get much closer to his subject.
Drones are also more flexible than cable-suspended camera systems, which are also present at the Winter Olympics. While live transmission is tricky—it requires an extra transmitter, which weighs on the drone—Masina says he's flown such devices at over 40 mph while delivering live, high-definition video.

Masina says that drones are "for sure" the future of sports broadcasting. UK company HeliPOV, which provides drones on which to mount existing cameras, has some great examples of just how cool the resulting video can be:


It won't take long for drones to be fully integrated into broadcasting: One company is already prototyping drones that automatically follow a person from a few feet away—perfect, they say, for downhill skiing. And drones are also making appearances at events smaller than the Olympics.

There are limitations: In many countries, drone regulations are still lagging behind the times, and it might not be clear to a broadcaster that they can be used legally. Then there are concerns about crashes. But with the risks low and potential benefits high, it could be that sports photography will be one of the first uses of drones to go mainstream.

Expect similar kinds of photography drones to be used for near-real-time GIS street level mapping, and for aerial maps typically provided by Earth satellite systems:  Drone generated street views for both indoor, and outdoor applications are not far away.

Delta 4 Rocket Boosts GPS Navigation Satellite Into Space

As reported by CBS NewsA towering United Launch Alliance Atlas 5 rocket thundered to life and climbed into space Thursday evening, boosting an upgraded Global Positioning System navigation satellite into orbit.

Under a clear, moonless sky, the 205-foot-tall rocket's hydrogen-fueled RS-68 main engine throttled up at 8:59 p.m. EST, followed five seconds later by ignition of two strap-on solid-fuel boosters.
Trailing a brilliant plume of fiery exhaust visible for miles around, the Delta 4 quickly vaulted away from launch complex 37 at the Cape Canaveral Air Force Station, accelerating to the east with 1.2 million pounds of thrust. 
The launching was delayed 19 minutes because of concern about higher-than-allowable levels of solar radiation. But the readings dropped back within limits before the launch window closed, and mission managers cleared the rocket for flight. 
The climb out went smoothly, and the solid-fuel boosters burned out and fell away about a minute and 40 seconds after liftoff. The first stage followed suit 2 1/2 minutes later and the rocket continued toward space under the power of a single hydrogen-fueled Pratt & Whitney Rocketdyne RL10B-2 engine.
The second stage engine shut down as expected about 12 minutes and 14 seconds after launch, putting the rocket and its GPS payload into an initial orbit with a high point of around 215 miles and a low point of just 100 miles or so.
Two subsequent engine firings were planned to boost the satellite into the desired 12,500-mile-high circular orbit tilted 55 degrees to the equator.
The Boeing-built Global Positioning System 2F-5 satellite was expected to be released from the Delta 4 second stage three hours and 33 minutes into the mission. 
The satellite is the first of three GPS navigation beacons scheduled for launch through July to replace older spacecraft in the constellation. The 2F-5 satellite is the fifth of 12 planned "block 2" spacecraft that make up the core of the current GPS fleet.
Equipped with ultra-accurate atomic clocks, Global Positioning System satellites circle the globe in six orbital planes. Using an active constellation of more than two dozen satellites, at least four GPS spacecraft are visible in the sky from any point on the Earth's surface, transmitting location and timing signals that allow compact receivers to compute position, altitude and velocity. 
An encrypted band gives military users position accuracy to within a few feet while an unencrypted public channel provides slightly less precise data to a wide variety of devices, from smart phones to automobile-mounted GPS mapping units.
The block 2F satellites are more accurate than earlier spacecraft and feature new channels to support commercial and civil aviation, more easily upgradeable flight computers and beefed up anti-jamming hardware.
The satellite being replaced by the GPS 2F-5 spacecraft was launched in 1997 and is well past its design life.
"The satellite we are replacing is over 16 years old and its design life was 7.5 years," Col. William Cooley, Global Positioning Systems director, told reporters before launch.
"Sometimes we joke those are getting old enough to vote and some are old enough to drink, and they're well past their design life. The oldest is 23 years. We've gotten remarkable performance out of them, but they are aging."

Thursday, February 20, 2014

Why Your Car Won’t Get Remote Software Updates Anytime Soon

As reported by MIT Technology Review: When Toyota recalled over two million cars last week because of flaws with antilock braking systems and other problems, the fix was simple—a few software updates.

The implementation of that fix is far from simple. Every one of those cars has to be taken into a dealership to have the new software installed, an expensive process that can take months. Cars that haven’t been fixed could, in some cases, suddenly stall and crash.

There is an alternative—the same sort of remote software updates used for PCs and smart phones. Indeed, one automaker, Tesla Motors, already provides what it calls “over-the-air updates,” which allowed it to execute a recent software fix without requiring anybody to bring in their cars (see “Tesla Motors’ Over-the-Air Repairs Are the Way Forward”).

Increasingly, many cars have wireless connections, for infotainment and communications; and some automakers already use wireless connections to add software to their cars at the factory. Even so, it will take some time for major automakers to implement over-the-air updates, both because they’re worried about security and because they might face resistance from dealers.

Software is rapidly taking over not only the entertainment console in cars, but also basic functions such as steering, braking, and acceleration, as more cars come with features such as adaptive cruise control and automated parallel parking. This can make it easier to diagnose and fix problems, but it also increases the risk for software bugs or even malicious attacks that might cause serious injury.

Software-focused recalls are increasingly common. In Toyota’s case, its software updates were designed to fix a computer problem that could disable antilock braking and other safety systems, or cause the hybrid propulsion system on its Prius vehicles to shut down. Last month, GM recalled 370,000 trucks at risk of overheating and catching fire, a problem it’s solving with a dealer-installed software patch.

In addition to potentially improving safety by delivering fixes faster, remote updates could save automakers money. “It’s very expensive to update the software when you have to bring it into the dealership,” says Andre Weimerskirch, a research scientist at the University of Michigan Transportation Research Institute.
Greg Schroeder, a senior research engineer at the Center for Automotive Research, says there’s “considerable interest” in the auto industry in developing remote updates. But he says most automakers are still in the early stage of testing the technology.

Toyota, in fact, says it recognizes that remote updates are technically possible, and could be convenient. But despite the inconvenience of the current approach, it’s decided not to do them for now because of “security challenges.”

These challenges are not small. Researchers have shown that existing wireless connections can allow them to hack into cars and take control of car locks and brakes; and this summer hackers demonstrated how to take over a car and steer it, slam on the brakes, or tell drivers that a nearly empty gas tank is full. Charlie Miller, a computer security expert for Twitter, is one of the hackers who took control of two cars this summer to uncover vulnerabilities. He says that remote updates will add a new target for hackers. But he downplays the risk, noting that no malicious hackers have taken over cars, in part because there’s no economic reason to do so. And he says remote updating systems can be made secure—“It’s possible to screw it up. But it’s certainly possible to do it right,” he says.

There is another reason automakers are moving slowly. Unlike Tesla, most automakers depend on independent dealers to sell their cars, and dealers have good reason to oppose automatic updates that would take them out of the loop. They get much of their revenue from servicing cars, says Joachim Taiber, a research professor in automotive engineering at Clemson University. And having drivers come into the dealership also helps dealers sell more cars.  He says Tesla may have been more willing to do automatic updates in part because it doesn't work with independent dealers.

Even if the change is slow, Miller says, remote software updates for cars are inevitable. As the amount of software in a car—and the potential for bugs—increases, remote updates “are going to have to happen,” he says. With the current approach of bringing cars into dealerships, “It can be months before software gets updated. It might never get updated,” he says. “That leaves a lot of cars in a vulnerable state.”