Search This Blog

Friday, January 24, 2014

Google’s Open Automotive Alliance Faces Competition From Apple And Microsoft

AFTER recent rumors that Google was planning to bring a new level of Android connectivity to cars, the company has officially announced that it will be teaming up with Audi, Honda, General Motors, and Hyundai to form an Open Automotive Alliance. This OAA will also include the graphics chipset company, Nvidia, to provide greater integration of the Android platform in today’s cars.
More specific details of this alliance were announced with the 2014 Las Vegas Consumer Electronics Show in full swing, stepping up the competition with Apple’s recent efforts to launch an in-car iOS platform.
With this announcement and others related to the future of the fully connected car, the Vegas Consumer Electronics Show has been a strong reminder of just how intersectional the tech and auto industries are becoming.
Chipset vendors, automakers, and platform companies are teaming up in alliances like Google’s to create new machines that will combine different technologies for a fully integrated, connected driving experience. Google’s Open Automotive Alliance will use a common platform and development model, with the first Android connected cars available later this year.
Yet Google’s not the only company racing to provide smartphone connectivity built into the automotive dashboard.
Hyundai, General Motors, and Audi have already announced that they will also be working with Apple.
Apple’s iOS in the Car project has been in full swing for some time now, providing direct competition for Google’s Android connectivity project.
Microsoft also has a hat in the ring, with versions of its Windows software already embedded into vehicles that you’ll find listed in showrooms and on automotive sites like Carsales.
Manufacturers including Fiat, Nissan, Kia, and Ford all use the Microsoft embedded automotive software. Furthermore, Microsoft launched an alliance of its own back in 1998 called the “Auto PC” initiative, with Citroen, Hyundai, Nissan, and Samsung all on board.
Both Google and Apple must play catch-up with Microsoft’s existing connectivity, but are primed to do so this year.
Apple’s iOS in the Car project will embed its popular Siri system into cars, along with GPS and telephone functions. A number of auto manufacturers have already signed up to participate in this platform, including those mentioned above as well as luxury brands like BMW and Mercedes-Benz.
It’s unlikely that auto manufacturers will have to choose one platform or the other, as they will want to provide the widest range of options to customers.
As a result, Audi, Hyundai, and General Motors are all taking part in both Apple’s and Google’s alliances. Google also stated in its announcement that the Open Automotive Alliance is open to future participants, so more may come on board to provide Android connectivity within their 2014 models and beyond.
In addition to Google’s announcement, the Las Vegas Consumer Electronics Show has showcased other ways in which cars will become more connected in the near future.
GM announced vehicles that will be enabled with the LTE network, and Qualcomm announced a new chipset that will be able to support Bluetooth, LTE, and Wi-Fi within the car. This new chip will also be able to support Android and Blackberry systems.
All of these advancements and the competition between Google, Apple, and Microsoft point to car connectivity becoming a major focal point over the next year. The future car will become a mobile computer, with full hi-tech integration.

Tech Watch: Supreme Court Rebukes Patent Court Judges (again)

As reported by GigaOM: There are many explanations for the sorry state of the U.S. patent system, but one that comes up on a regular basis is the U.S. Court of Appeals for the Federal Circuit. The pro-patent proclivity of the court, which hears every patent appeal in the land, has given it a “rogue” reputation and forced the Supreme Court to reverse its decisions again and again. This week, the Supremes did so once more.

In a 9-0 unanimous decision, the Supreme Court on Wednesday ruled that the appeals court botched things anew, this time over the level of proof that a licensee must bring to challenge a patent. The nuts-and-bolts of the decision turn on procedural issues, but the gist is that it’s now easier for those targeted by patent lawsuits to file a preliminary challenge to accusations of infringement. (You can get a full run-down at Patently-O and SCOTUSblog.)


The procedural issue is important because, under the current system, patent holders — including trolls — can simply file a bare-bones lawsuit without explaining why their target is infringing the patent. Now, under the new Supreme Court ruling, the patent holder will at least have the burden of showing infringement. Congress, meanwhile, is also attempting to tighten up pleading requirements as part of the Innovation Act, a new law expected to pass this spring.

As patent reform advocates point out, this week’s ruling is just one in a series of unanimous reversals of the Federal Circuit’s patent positions. Meanwhile, the patent court’s recalcitrant judges could receive another scolding this spring when the Supreme Court hears a key case over what can and can’t be patented.

Finally, the new boss of the U.S. Patent Office this week published an inaugural blog post that expressed support for Congress’s efforts to reform “abusive patent litigation tactics.”

Proposal for U.S. eLoran Service Gains Ground

As reported by Inside GNSS:  Trying to revive a years-dead federal program is usually the kind of hopeless task that even Sisyphus wouldn’t touch.
But determined supporters of eLoran are gaining ground in their effort to resurrect the cancelled radio-navigation network and, propelled by new worries over GPS jamming, they appear poised push the issue through.
Loran, short for LOng RAnge Navigation, enables ships and aircraft to determine their speed and location using low-frequency signals broadcast from ground stations. The original, and now obsolete, Loran-C system was decommissioned in 2010. At one time, an enhanced system — eLoran — was expected to replace it with signals that, unlike GPS, could reach under ground, under water, and into buildings. eLoran is also far more sophisticated than Loran-C, said David Last, a British expert on positioning, navigation and timing systems (PNT)
“I draw the analogy by saying that Loran in its original form, which a lot of people remember, came from the days of black and white television,” Last told Inside GNSS. “What we’ve got here [with eLoran] is still television, it’s still Loran — but it’s digital. It’s high-definition. It’s color. It’s big screen. It’s all of those things.”
More importantly, the high-powered and nearly unjammable eLoran signals are an excellent backup to GPS signals and not subject to the vagaries of space warfare or asteroids. If something happened to GPS, eLoran could provide, relatively inexpensively, the positioning information needed for navigation and the timing data crucial to the power grid, cell phones, financial networks, and the Internet.
“eLoran is the only cost-effective backup for national needs,” wrote an Institute for Defense Analyses Independent Assessment Team (IAT) led by Brad Parkinson, the first director of the foundational GPS Joint Program Office.
“It is completely interoperable with and independent of GPS,” IDA’s team said in their 2009 report, “with different propagation and failure mechanisms, plus significantly superior robustness to radio frequency interference and jamming. It is a seamless backup, and its use will deter threats to U.S. national and economic security by disrupting (jamming) GPS reception.”
A paper by David Chadwick and Taehwan Kim show just how jam-resistant eLoran is. Their calculations in “An Evaluation of the Effectiveness of Intentional Interference to eLoran” presented at the 2006 MILCOM conference showed the system to be so robust that jammers hidden in suitcases or broadcasting from a plane or a hijacked AM radio station could not prevent the system from operating as a GPS backup for aviation.
It’s no wonder. The signal is approximately 1.3 million times more powerful than the GPS signal, said Dana Goward, president and executive director of the Resilient Navigation and Timing Foundation (RNT) Foundation, which supports adoption of eLoran.
“The cost was so minimal and the backup was such a powerful thing . . . it just made enormous sense,” said an expert familiar with the IAT study, particularly “in terms of deterring somebody from even bothering to jam GPS.”
The IAT’s report convinced the Departments of Transportation and Homeland Security, which threw their support behind the transition to eLoran. The program was ultimately approved by the National Executive Committee for Space-Based PNT — a group co-chaired by Deputy Secretaries from DoT and DoD.
Then, after the government spent some $160 million on upgrades, eLoran was canceled. At some point in the process of preparing its first from-scratch budget, the Obama administration zeroed out the eLoran money and killed the program in 2010 — a decision widely attributed to the Office of Management and Budget.
What exactly happened is a mystery, although numerous sources describe the decision as an attempt to save money in the wake of the Great Recession.
“I think probably the true explanation, we see it in other places as well, is twofold,” said Last. “One is the tragedy of the commons — the argument that, if something is good for everybody in the village, then nobody will want to pay for it. And that certainly applied in Washington, according to everything that I saw. The other is just the whole sense that Loran is ‘that old thing’ going back a long way. It’s very difficult to take something that was known a long time ago and convince people that the current version of it is very different.”
No agency has stepped up since to support the program — perhaps with good reason. Sources suggest that the different departments are afraid to look too interested in eLoran out of fear they will be tasked with paying for the entire program from their ever-shrinking budgets.
Impact of Jamming 
Despite its cancellation and lack of official enthusiasm, support for the program has been growing in the United States — particularly in light of an escalating number of jamming incidents. That support is now breaking out into the open at an agency level.
Last April, at a meeting organized by the U.S. Naval Observatory, the Defense Department’s Chief Information Officer Teresa Takai expressed an interest in eLoran. Takai told the audience at the “Time and Navigation, 1730-2030, from Greenwich to Space” symposium that she would like to look beyond GPS for options to improve the overall PNT capabilities for U.S. and allied warfighters.
The reason is clear, sources told Inside GNSS. The Defense Department is seeing the impact of GPS jamming overseas and is aware of how eLoran can be used to mitigate the problem.
“DoD is looking intently at what is going on in Korea,” one source said — an assessment confirmed by other experts familiar with DoD’s perspective. “The South Koreans were very intent on proceeding with eLoran, the reason being that they had actually experienced North Korean jamming of the GPS signal.”
Jamming by North Korea against its southern neighbor began in 2010. The on-going attacks have escalated to the point that they affected more than 1,000 ships and 250 planes over a 12-month period, according to a paper given in April at the European Navigation Conference (ENC) in Vienna, Austria. The paper’s statistics were presented by Jiwon Seo, an assistant professor in the School of Integrated Technology at Korea’s Yonsei University and Mincheol Kim, a deputy director of the Maritime Safety Facilities Division in the Ministry of Oceans and Fisheries.
The two announced that South Korea planned to add three new eLoran stations and update two old Loran-C facilities to provide eLoran services with a better than 20-meter accuracy throughout the country by the end of 2014. It would also deploy 43 differential eLoran stations for a service to be operation by 2018. With the cooperation of Russia and China, they said, South Korean hopes to expand its eLoran system to cover all of Northeast Asia.
Other countries are taking up eLoran as well. India is planning a system of its own and Russia and Great Britain are working together to make eLoran interoperable with Chayka, the Russian version of the technology. Their goal is to use the compatible signals to improve navigation along hazardous sea routes in the Arctic, according press reports and a presentation the Internavigation Research and Technical Centre in the Russian Federation.
The Netherlands announced in December that it had installed a differential version of eLoran in the port of Rotterdam that used signals from stations in France, Germany and England. The system, which achieves accuracies of less than 5 meters, was developed in part to address the risk of disruption to satellite navigation signals.
Saudi Arabia announced plans some time ago to upgrade its Loran-C system to e-Loran and Iran announced last year that it had a new terrestrial positioning system, though little is know about it. The U.K., easily the most active of the eLoran countries, has been broadcasting eLoran signals 24/7 for nearly three years, said Last.
“About a year ago we introduced it in the Dover Straits,” Last told Inside GNSS. “That’s the part of the English Channel that is the world’s busiest maritime choke point. We get something like 500 ships a day coming through it — and ferries dashing back and forth across it. The whole area is a bit of a nightmare and, of course, it can be very foggy as well.”
Most of the ships traversing the Straits are “almost totally reliant on GPS for their navigation,” he said “and the gap that they are coming through is narrow enough that you could take out the whole of the shipping activity there — take out the GPS activity — using a fairly low-cost jammer on cliff top on either the British or the French side.”
“We took the technology from the U.S. when they dropped the ball, and we spent about three years on it,” Last said. “The outcome of that is that we have turned what was a feasibility study being done by people like the U.S. Coast Guard Academy and developed that into something that we now have a lot of faith in.”
Great Britain is poised to take it one step further. The U.K. announced plans last summer to install seven differential eLoran stations along its South and East Coasts. The differential service is to be operational by this summer.
eLoran in the States? 
With all the eLoran activity overseas, particularly in regions of strategic interest, it makes sense for DoD to support eLoran in the U.S. for its own training purposes, said one source.
The military could use commercial equipment to support their activities, they said, and would not need to expend substantial resources on developing new defense-unique equipment or install receivers in every truck and tank. Given the amount of eLoran activity overseas, such equipment should be readily available, the source suggested.
“There would be interest in eLoran equipment if there was a move to regenerate the eLoran infrastructure, starting I am sure with the suppliers overseas who are getting ready to support Korea and the (other eLoran countries), said another expert who has been following the issue. “It would happen here if the infrastructure were regenerated.”
eLoran supporters are hoping that even limited use of eLoran by the Pentagon will spark commercial interest and, eventually, participation by DoT and DHS. For example, if DoD were to develop a well-recognized signal specification, it would go a long way toward spurring the introduction of commercial eLoran-capable receivers. The money is there in the DoD budget to do these things, argued one source, in part because the amounts needed are so small.
That little bit of spending could make a big difference, though, by proving the practicality of the system, they said.
“You kind of have to coax the civil community,” said the source, who is familiar with the long-running debate over eLoran. The key to getting DoT and DHS onboard, they said, is the “ability to demonstrate it is not a budget buster.”
Private Sector Steps Up
And DoD does appear to be interested.
In November Goward, who as director for Marine Transportation Systems for the Coast Guard served as the United States’ maritime navigation authority for the United States, announced the formation of the RNT Foundation.
The group is proposing that the United States adopt eLoran and is suggesting that a public-private partnership could be established to provide the funding.
“We have had some positive responses to our proposal for a public-private partnership from the Department of Defense and the Department of Transportation,” said Goward, “and, so, we expect to be talking with them in the next couple of weeks. We’ll see where that goes.”
The proposal is similar to one submitted last year by UrsaNav, which is currently doing research at a number of the old Loran-C sites under a cooperative research and development agreement or CRADA with DHS. UrsaNav is a supporter of the RNT Foundation.
“We would want (the government) to contribute the existing infrastructure. ‘As is’ is fine,” Goward said. “And then, depending on the business model that they chose, they could either contribute a certain amount a year or a certain amount up front.”
If the federal government is going to get a direct benefit back from the system, he added, “we would hope they would contribute a little bit to the restoration of the system. If not, I suppose we would raise the funds ourselves and set it up as a purely fee-for-service business. There is a whole spectrum of possibility on how this could be funded.”
Sources confirmed that DoD officials were scheduled to have a preliminary meeting with Foundation representatives in January. The DoT said it had already been involved in meetings.
“The U.S. Department of Transportation and the other agencies have held preliminary discussions and are planning future high-level talks on GPS backup generally,” DoT said in a prepared statement, “but we don’t have a schedule of future talks at this time. USDOT’s role will continue to be to provide input on GPS requirements from the civil sector.”
Sources said that DoD reached out to the DHS in a letter sent as part of setting up the initial meetings. Goward believes DHS will participate in the discussion, althoughInside GNSS was unable to confirm that. DHS, which is often described as being responsible for finding a backup for GPS, did not respond to requests for comment.
Congress Turns Up the Heat
If the three agencies find themselves unable to work out a plan, however, Congress is ready to provide a measure of very pubic incentive.
The House Coast Guard and Maritime Transportation Subcommittee, chaired by Rep Duncan Hunter, R-California, has scheduled a February 4 hearing headlined “Finding Your Way: The Future of Federal Aids to Navigation.” A congressional staffer toldInside GNSS that eLoran may be one of the topics discussed. Indeed, although the full list of speakers was not public as of press time, Goward confirmed that he would be testifying.
Another topic that may come up is the negative findings in a GAO report on GPS vulnerability requested by Sen. Tom Coburn, R-Oklahoma, Sen. Susan Collins, R-Maine, and Rep. Michael McCaul, R-Texas. The three have been following the eLoran issue and asked the Government Accountability Office to examine the risks and potential effects of GPS disruptions. GAO was also tasked with assessing whether DHS and DoT were living up to their responsibility to find mitigations for system problems and develop a backup for GPS.
The report, released last November, faulted DoT and DHS for their “limited progress” — progress hampered, in part, by a disagreement over which agency is responsible for developing a GPS backup system. If that is not enough to inspire action, the U.S. House and Senate have ordered the secretary of defense and the director of national intelligence to work with the National Research Council to study and determine options for responding to “near-term and long-term threats to the national security space systems of the United States.”
The mandate, which was imbedded in the National Defense Authorization Act for 2014 and signed into law at the end of December, gives them until the end of this year to recommend actions to counter the threats including, as an option, building in systemic resilience in some way.
They must describe, as part of the country’s Space Protection Strategy, how the two departments will provide “necessary national security capabilities — through alternative space, airborne, or ground systems — if a foreign actor degrades, denies access to, or destroys United States national security space capabilities” — a clear opportunity to address the need for a backup to GPS.
Dismantling of Loran Stations Continues
What Congress has failed to do, however, is stop the destruction of the old Loran-C sites, which are important to the potential success of a new eLoran network.
The Coast Guard continues to dismantle the old sites, taking down antennas and removing the ground equipment. Two towers were dismantled in December and another is scheduled to be taken down by the end of March, according to the Coast Guard.
Saving the existing sites is essential said Chuck Schue, who follows the developments closely as president and chief executive officer of UrsaNav, a provider of eLoran technology and services. The old sites can be put to use quickly to broadcast signals that could spark development of new receivers and help encourage the market for the service — which could ultimately save the federal government money.
The more the government dismantles the old sites the more it will cost to launch eLoran and the longer it will take, he said.
Dismantling the sites also prolongs the period during which American will be have only one source for timing and navigation data, said Goward. “We should not have so much depending on one source.”
eLoran supporters hope that DoD will ask the Coast Guard, which became part of the Department of Homeland Security after 9/11, to stop taking down the sites until a decision on eLoran can be made.
Although Schue said he believes DHS is participating in the talks, he thinks that officials representing the agency are not from the same part of DHS as the Coast Guard — possibly making it difficult to coordinate on saving the sites.
“It appears that one hand does not know what the other doing.”

Thursday, January 23, 2014

Ford Is Working With MIT, Stanford To Build “Common Sense” Into Self-Driving Cars

Automakers are building research vehicles that
can take in vast amounts of data about their
surroundings in a split second.  Now it's up to data
scientists to figure how cars can use that information.
As reported by GigaOM:  Ford Motor Company is teaming up with the Massachusetts Institute of Technology and Stanford University to research the future brains of its autonomous cars. Projects like Ford’s research vehicles are putting the sensors and computing power into cars that would allow them to read and analyze their surroundings, but these two universities are developing the technology that will allow them to make driving decisions from that data.

“Our goal is to provide the vehicle with common sense,” Ford Research global manager for driver assistance and active safety Greg Stevens said in a statement. “Drivers are good at using the cues around them to predict what will happen next, and they know that what you can’t see is often as important as what you can see. Our goal in working with MIT and Stanford is to bring a similar type of intuition to the vehicle.”

In December, Ford unveiled its latest research vehicle, a Ford Fusion Hybrid equipped with Lidar (laser-radar) rigs, cameras and other sensor arrays, all intended to generate a real-time representation of the world around the car. Such a car can “see” in all directions, allowing it not only to take in far more stimuli than even the most alert driver, but also to react to that information far more quickly. That’s where Stanford and MIT come in.

The Ford Fusion research vehicle from Lidar's point of view
The Ford Fusion research vehicle from Lidar’s point of view

MIT is developing algorithms that will allow an autonomous driving system to predict the future locations of cars, pedestrians and other obstacles. It’s not good enough for a car to merely sense the location of nearby vehicles when it switches lanes or swerves to avoid an accident. It has to know where those vehicles will be a split-second later. Otherwise the car will avoid one accident only to cause another.

That means not only measuring other vehicles’ current speed and trajectory but anticipating how their drivers – or their autonomous vehicle systems – will react to the situation. Basically MIT is trying to create a vehicle brain smart enough to assess risks and outcomes and navigate its course accordingly.

Stanford is doing something a bit different. It’s trying to extend the sensory field of the car by helping it see around obstacles so it can react to dangers the driver can’t immediately see. Stanford and Ford didn’t offer any specifics on just how they would accomplish that feat, by my bet is it has to do with Ford and the automotive industry’s work on inter-vehicle networking.

Cohda Wireless autonomous car

Future autonomous cars won’t just be able to sense their surroundings, they’ll be able to communicate with other vehicles using a secure form of Wi-Fi. For instance, Australian startup Cohda Wireless is developing to vehicle-to-vehicle networking technology that would allow two cars to let each other know they’re approaching one another at a blind intersection.

Ford and other major automakers are working with the University of Michigan and the National Highway Traffic Safety Administration to build vehicle-to-infrastructure grids that would allow cars to tap into highway sensors, giving them a kind of omniscient view of the overall road. With such technology other cars could reveal their intentions before they even take action, making other connected vehicles much more responsive. They could also share their sensor data, so even if only one of the cars far ahead of you is connected to the vehicle grid, that lone vehicle could still tell you what the other cars around it are doing.

While every major automaker is working on autonomous driving technology, Ford has been particularly aggressive. In a recent interview, executive chairman Bill Ford told me how the automaker is trying to use connected vehicle technology to propel the company into a new golden age of automotive innovation.

What Spectrum Crisis? First Airwave Auction In 6 Years Kicks Off With Little Fanfare

The U.S. mobile industry claims it's desperate for new spectrum.
But when the FCC opened it's first mobile broadband airwave
auction in a half a decade, not a single major carrier participated.
As reported by GigaOM:  For the first time in half a decade, the Federal Communications Commission on Wednesday opened up bidding on a new spectrum auction, releasing new airwaves for 3G or 4G data services to potential buyers. But you would hardly know it, judging from the lack of interest across the mobile industry.

Not a single one of the four nationwide providers is participating in what is now known as Auction 96, which is distributing 10 MHz of frequencies in the 1900 MHz PCS band nationwide. That’s not exactly the kind of behavior you’d expect from an industry that insists it faces a spectrum crisis of the direst order.

regional mobile carrier

Granted 10 MHz isn’t much in the grand scheme of things. It’s enough to add incremental 3G capacity or deploy an LTE network half the size of what most major operators have in the field today. But bandwidth is bandwidth. Yet the major carriers have decided to take a pass on these airwaves, looking ahead to the FCC’s big incentive auction next year. If done right, that auction could open large chunks of frequencies in the much more desirable 600 MHz band.

So who is participating in Auction 96? Dish Network is probably the one name you’ll recognize, but a lot of smaller carriers are also submitting bids in hopes to add to their regional holdings. In three rounds, bids now total $221 million. As you would expect, licenses in the big cities are attracting the most interest. The New York City license alone has attracted one quarter of all bids so far, followed by Los Angeles and Chicago.

The auction is just getting started, and it doesn’t run at quite the fast pace as estate sale. The FCC holds three bidding rounds each day until there are no more bids. This auction certainly won’t be the two-month-long process we saw in 2008 for the highly contested 700 MHz airwaves. We’ll probably see this auction conclude in a few weeks if not less.

Battery Modification May Add 27 Cumulative Years Of Life To GPS Satellite Fleet

The GPS IIR/M satellites have been launched between 1997 and
2009.
As reported by NextGov: The Air Force Space Command’s Space and Missile Systems Center and a team of contractors have extended the operational life of 19 GPS satellites in orbit by re-configuring their battery chargers.

Lockheed Martin launched the GPS IIR/M satellites between 1997 and 2009, and the fleet accounts for more than half the 36 GPS birds on orbit with batteries, the “the primary life-limiting component when GPS IIR/IIR-M vehicles are past their design life,” SMC said.

Aerospace Corp., a federally funded research and development center with extensive GPS expertise,  Lockheed and SMC determined that reducing the charge rates during solstice season would add an average of one to two years of life per space vehicle.

Last week, the 2nd Space Operations Squadron, Schriever Air Force Base, Colo., completed the battery charge modification, which will extend the life of each of the GPS IIR/IIM satellites by one to two years, more than 27 years of cumulative life across fleet, SMC said.

The changes represent a savings of hundreds of millions of dollars for the US government.

Wednesday, January 22, 2014

How AI Turns Traffic Lights Into Intelligent Agents

As reported by ReadWrite: Today’s new cars are loaded with sensors and powerful computer processors. That’s the high-tech pathway to turning our vehicles into super-efficient, semi-autonomous—or even self-driving—"transportation devices."

Unfortunately, the roads these clever mobility machines drive on are all too often, well, dumb. You experience the pain of this problem every time you senselessly wait for an extra couple minutes at a red light, when there are no other cars in sight from any direction.

Samah el-Tantawy, a recently minted PhD of Engineering from the University of Toronto, wants to change that.

Inspired by research from her advisor and director of the Toronto Intelligent Transportation Systems Centre, Professor Baher Abdulhai, el-Tantawy devised a system that uses artificial intelligence and game theory that, in a simulated environment, shaved 40% of the time off an average wait at an intersection. She said that could mean 12.5 fewer minutes stuck in your car, if you pass through about 30 intersections on your commute.

Can We Talk?
According to el-Tantawy, many of today’s traffic lights at intersections operate based on pre-programmed repeated cycles that run with little or no input from fluctuations in traffic. Yes, there are sensors in pavements along major arteries, but those inputs into centralized systems might only be able to extend a green light for a few seconds. Like other centralized disconnected top-down systems, there are inherent limitations.

Instead, el-Tantawy’s system—dubbed MARLIN for Multi-agent Reinforcement Learning for Integrated Network (of Adaptive Traffic Signal Controllers)—uses video cameras, other vehicle data inputs (if available), processing power, and routers to analyze how many drivers are zipping through the intersection and how many are simmering with road rage for wasting countless minutes at a red light. With MARLIN, cameras are aimed at all four approaches, and the system is distributed throughout a region rather than just on main streets.

“Our approach is decentralized, where the intelligence or math to assign the greens is done on the fly at each intersection,” she told me.  “The brain sits at each intersection, and calculates the best timing to minimize the number of cars approaching and waiting, and it coordinates those decisions with other lights at other intersections.”

The Shortest Wait Wins
El-Tantawy said no amount of math can perfectly model every situation. There are too many variables. The solution? “Each intersection is connected to the neighboring or adjacent intersection, sending and receiving information about the waiting vehicles,” she said.  Then, “reinforcement learning” comes into play.

Like a child learning to walk by making minute adjustments, each traffic light—or “agent,” as el-Tantawy calls them—makes a decision every second about the best way to keep motorists and pedestrians waiting for as short a period as possible.


“The agents learn, until they converge, with each one getting the best response action to achieve its goals, without negatively affecting the others. We use multi-agent reinforcement learning,” she said. “And it cascades throughout the system. The decisions by agents affect each other, so it’s a game.”

The system has to be simulated in a test environment before being placed in the street, where the learning can continue in real-world conditions. So far, MARLIN has only been used in a test environment—but with great results. That encouraged el-Tantawy and Professor Abdulhai to recently form a start-up traffic tech company to commercialize the system, and get it on as many streets as possible.

It's easy to see how this AI-powered traffic light system could be a major boost to a region's productivity, justifying the anticipated cost of about $20,000 to $40,000 per intersection.

el-Tantawy said her start-up will soon sign up its first municipality to run a field test, but wasn’t ready yet to disclose the location. I hope it’s in my neighborhood.