LiDAR
Latest
MIT builds an early warning system for sailors
MIT believes that it's developed an early warning system for rogue waves that could save the lives of sailors. Currently, predicting when a ten foot high wave is going to crash across a ship takes a room full of supercomputers. That's not helpful for the scrappy fishing boats you see on Discovery Channel shows like Deadliest Food Grabbers and Gruff Seamen at Sea. That's where this new method from Themis Sapsis and Will Cousins comes in, which spot impending trouble and alert crews to batten down the hatches.
Ford wants you to know it's serious about self-driving cars
CES doesn't officially start until tomorrow, but already Ford has talked up its new project with DJI to launch drones from moving vehicles, as well as impending updates to its Sync 3 connected car platform. The automotive giant is also using the biggest tech show of the year to reaffirm its commitment to self-driving vehicles, announcing that it plans to add 20 more Ford Fusion Hybrid cars to its autonomous armada in 2016. According to Ford, tripling the number test vehicles in its garage would give it the biggest fleet of any outfit working on self-driving technology.
See London through the eyes of a self-driving car
When a driverless car glides down the street, avoiding pedestrians and stopping at traffic lights, what exactly does it see? Is the machine's perspective of the world so different to you and me? To give us an inside look, ScanLAB Projects strapped a laser scanner to a Honda CR-V and cruised through the streets of London. Although the car was being driven by a human, the LIDAR (light detection and ranging) equipment performed similarly to how it would in a driverless car. Like radar or sonar, this involves firing a laser in every direction and then measuring the time it takes to reflect back off nearby objects. These timings are then collected, analyzed and collected again to give the car a real-time picture of its surroundings. ScanLAB Projects is a London design company that uses the technology for art and visualisation purposes; they've already scanned museums, an underground railway line and the arctic circle. Its latest video -- produced for the New York Times Magazine -- combines the LIDAR recordings from the drive for a beautiful, eery look at the city.
Self-driving cars can be fooled by fake signals
You'd think that self-driving cars would be most vulnerable to remote hacks, but the biggest danger may come from someone nearby with a handful of cheap electronics. Security researcher Jonathan Petit has determined that you can fool LIDAR (the laser ranging common on autonomous vehicles) by sending "echoes" of fake cars and other objects through laser pulses. All you need is a low-power laser, a basic computing device (an Arduino kit or Raspberry Pi is enough) and the right timing -- you don't even need good aim. Petit managed to spoof objects from as far as 330 feet away in his proof-of-concept attack, and he notes that it's possible to present multiple copies of these imaginary objects or make them move. In other words, it'd only take one prankster to make a self-driving car swerve or stop to avoid a non-existent threat.
How laser scans and VR are preserving London's hidden 'Mail Rail'
Deep in the heart of London, buried beneath 70 feet of soil and concrete, lies a hidden underground railway. For almost 80 years, the UK's "Mail Rail" transported letters and parcels between the capital's main post offices and a few overground train stations, where they could then be delivered across the country. It was a unique way to avoid street congestion, but by 2003 the line had become uneconomical to run. The decision was made to shut it down and it's laid dormant ever since, invisible to the public.
Caltech wants to equip phones with built-in 3D scanners
Wouldn't it be great if we can use phones to 3D scan every day objects for printing? A research team from Caltech has designed a cheap and tiny camera chip that could make that happen. Even better, the team claims it can take 3D scans so precise, it could lead to "replica[s] accurate to within microns of the original object[s]." The silicon chip called nanophotonic coherent imager (NCI) is less than a millimeter square, so it can fit within actual phones or cameras without the need for extra hardware. It can capture an object's height, depth and width using tiny LIDARs that immerse an object in laser beams. The light that bounces off the object is then analyzed to determine its size and other details.
Google has some new friends working on its self-driving car
Apparently following the adage about keeping your friends close and (potential) enemies closer, the Automotive News World Congress in Detroit hosted Elon Musk last night and followed up today with Chris Urmson of Google. He's the director of its self-driving car project, and while there are no big-name car manufacturers lined up as partners yet, he did call out seven particular industry friends providing various parts for the latest prototypes. That includes LG (battery), Bosch (LIDAR), ZF Lenksysteme (steering gear), RCO, Continental and FRIMO, while Roush is helping to actually put it together here in the Detroit area. There are also NVIDIA chips inside, which Urmson said run mostly Linux software "with some other stuff happening." He confirmed that Google has talked to the usual list of big names (GM, Ford, Toyota, Daimler, etc.), and hopes cars are ready to go by 2020.
This little laser could be the key to inexpensive self-driving cars
To listen to the Googles and the Elon Musks of the world is to believe that one day soon, we'll be tooting down in the highway in cars that control themselves. Those cars will need eyes, though, and that's where a company called Velodyne Acoustics comes in -- it whipped up the Puck, a scaled down version of the laser scanning tech that helps Google's self-driving cars see (which they also make) that only costs about a tenth of the price. Not up on the market rates for portable LiDAR arrays? Let's put it in perspective: the units Google use cost around $75,000. This new model? Just a shade under $8,000. That old cliché about getting what you pay for still holds true here, though -- this cheapo LiDAR array only has 16 lasers with which it scans the world, down from the 64 seen in higher-end units. As a result we're looking at a pretty staggering dip in the sheer number of environmental data points, but the unit's smaller size and price tag just might mean big things for car companies trying to craft the first truly affordable wave of autonomous cars.
Ford's Fusion Hybrid research car will explore our driverless future
Ford is convinced that autonomous cars are the future, and it's putting its money where its mouth is by unveiling a Fusion Hybrid variant built solely for researching driverless-car technology. The experimental machine boosts existing driver-assistance tech with four LIDAR sensors that create a 3D map of the vehicle's surroundings; while the data will only improve driver awareness in the short term, it should eventually help the car navigate on its own. Ford isn't planning a production version, so don't look for a robotic Fusion at your local dealership anytime soon. However, the project should be invaluable for translating autonomous-driving concepts into practical realities.
NASA's humanoid Valkyrie robot could lead the way to Mars (video)
NASA's Robonaut is only just getting its legs, but it already has some fresh (and arguably superior) competition. The agency's Johnson Space Center has given IEEE Spectrum a sneak peek at Valkyrie, a DARPA Robotics Challenge entry built to take on dangerous tasks that would normally require a human presence. In addition to sporting more joints than many of its typical two-legged rivals, the sometimes autonomous machine is chock-full of cameras, LIDAR and sonar that help it quickly move around. It can operate under planetary gravity loads that Robonaut can't touch, and even an inexperienced user can keep it running; the modular design lets astronauts swap out batteries and replacement limbs within minutes. Valkyrie isn't guaranteed to enter service, as it will have to win the support of both DARPA and NASA officials before it progresses much further. If it passes muster, however, it could be vital to NASA's long-term plans for Mars -- robots such as Valkyrie will likely prepare any future Martian landing site and assist humans once they arrive.
Nokia Here collection vehicle v2.0 ride-along (video)
What do the Lumia 920, Surface Pro, Velodyne LIDAR, NovAtel GPS and Jetta TDI wagon all have in common? They're all essential parts of the next generation Nokia Here collection vehicle. When Here was announced last fall, we checked out v1.0 of the car which featured an elaborate but clunky set of sensors and cameras. That's also when Nokia announced it was acquiring Earthmine, a Berkeley-based 3D-mapping company. Nine months later, we're seeing the results of this collaboration with v2.0 of the collection vehicle. It's a simpler and more advanced setup which provides much improved image quality. Nokia recently invited us to take ride in a next generation Here car -- check out the gallery below then hit the break for our video and breakdown of the technology.
LIDAR system uses lasers to detect clear air turbulence before it hits
Fact: turbulence is no fun at all. Aside from the brief moments of bone-deep terror it can cause passengers, clear air turbulence (CAT) can also prove costly for airlines in terms of damages. To help pilots deal with difficult-to-spot areas where CAT is likely to occur, the German Aerospace Center (DLR) is spearheading a new joint European project known as DELICAT. Essentially, the DELICAT program calls for planes to be equipped with a light detection and ranging instrument (LIDAR) that emits short-wave ultraviolet lasers. Radiation then bounces off oxygen and nitrogen particles in the air, indicating fluctuations in air density that signal the presence of CAT pockets. Until the end of August, the DLR will run test flights in a specially modified Cessna Citation plane to both show off LIDAR's capabilities and to give researchers invaluable data on CAT patterns. Though the system is still in its infancy, the folks at the DLR hope that the technology will one day become a standard part of commercial air travel.
LIDAR reveals ancient city remnants beneath Cambodian jungle
LIDAR scanning has recently become cost-effective enough for archaeologists to use on large historical sites, and they're taking full advantage. A helicopter jaunt last year has revealed a massive urban site below the jungles near Angkor Wat in Cambodia that likely housed thousands of people. New canals, temples and other man-made structures were discovered during a two-day scan, which emitted up to 200,000 laser pulses per second and would have taken years if done by traditional excavation methods. The technique can scope out features as small as a footprint, and is also being used in cities around the Egyptian pyramids and other archaeologically interesting regions -- marking another way that Indy-style archeologists are becoming obsolete.
Nokia Here collection vehicle hands-on (video)
After our interview with Peter Skillman last week we were given a brief tour of the Nokia Here collection vehicle, a bright blue Volkswagen Jetta Wagon specially outfitted with a smorgasbord of sensors. The main attraction is a retractable roof-mounted mast that collapses behind an aerodynamic fairing for stowage. From top to bottom, this mast features a military-grade GPS antenna, a 360-degree panorama camera in a white pod (which captures the spherical views you'll see in Nokia Here), a Velodyne LIDAR unit in a spinning silver cylinder (for 3D mapping), high-resolution signage cameras in a black box (for forwards and backward automatic feature extraction) and finally a wheel encoder to measure distance / velocity. Sadly the car was locked so we were unable to check out what's inside or go for a ride, but we're hoping to remedy this at some point in the future. Until then check out the gallery below and hit the break for our hands-on video.%Gallery-170826%
DARPA Innovation House project wants teams to take imaging data, see the big picture
Where are the bad guys? The military has eyes and ears everywhere these days, including drones large and tiny, satellites, radar imaging, LIDAR, infrared, thermal and even the enemy's own cellphones. The problem is how to take all that imaging and create a single picture of the environment. To that end, DARPA and George Mason University in Arlington have created the first Innovation House Project, which will put eight teams together for eight weeks in a "crucible-style" living environment to try to invent new ways of crunching the diverse sensor info. The military's research arm wants those units to think way off-piste "without fear of failure" to dream up solutions, and will have access to specialists and mentors from the military and academia. Unlike DARPA's usual challenges which have a grand prize, all teams accepted to the project will receive $30,000 in funding, but groups who go on to survive a four week cut will get an additional $20K. Proposals will be accepted up to July 31 (with no academic credentials needed), and the competition will begin in earnest on September 17. DARPA will get a license of any software created, allowing teams to hold the rights -- and hopes to continue the concept down the road, with new themes for team-based research on a tight deadline. So, if you're a data, imaging or "geospatial" whiz -- and don't mind being locked in a house and put under the brainstorming gun by DARPA -- check the PR for all the details.
US Navy Fire Scouts will automatically spot pirates, give 30 seconds to comply
War. What is it good for? Well, if new use of technology by the US Navy has anything to do with it, finding Pirates for a start. By upgrading its existing Fire Scouts with new 3D laser imaging tech, it's hoped that the drones will be able to recognize the small ships used by these unscrupulous seafarers. The system, known as LIDAR (Light Detection And Ranging, also known as LADAR) uses millions of laser pulses reflected off an object to create the three-dimensional image, which could then referenced against known pirate ships from a database. Ultimately, human operators will make the final call, to avoid any ED-209 style mis-understandings. That said, if you're taking the dingy out past the Californian breakwaters this summer, you might want to keep the stars and stripes in clear view, as that's where the Navy will be running its initial trials.
Microsoft Kinect used to map asteroids, glaciers, other scary things
Ken Mankoff is a PhD student at the University of California, Santa Cruz, where he studies ice and ocean interactions. He also counts himself among a growing legion of environmental scientists who have begun using Microsoft's Kinect to create detailed, 3D maps of caves, glaciers and even asteroids. As Wired reports, the Kinect has garnered something of a cult following within the scientific community, especially among those who, until now, have relied upon comparatively more expensive and complicated technologies to gather detailed 3D data. The approach du jour for most researchers is something known as Light Detection and Ranging (LIDAR) -- a laser-based technology capable of creating precise maps over relatively large areas. The Kinect, by contrast, can only see up to 16 feet in front of itself, but at just $120, it's significantly cheaper than the average LIDAR system, which can run for anywhere between $10,000 and $200,000. It's also surprisingly accurate, capable of capturing up to 9 million data points per second. Mankoff, for one, has already used the device to map a small cavern underneath a glacier in Norway, while Marco Tedesco, a hydrologist at the City College of New York, is looking to attach a Kinect to a remote-controlled helicopter, in the hopes of measuring so-called meltwater lakes found on glaciers during the summer. Then there's Naor Movshovitz, also a PhD student at UC Santa Cruz, who's more interested in using the Kinect and its image processing software to figure out how asteroids behave when broken up by a projectile. There are limitations, of course, since the device still has trouble performing amidst severe environmental conditions, though its supporters seem confident they'll find a solution. Read more at the source link below.
Kinect quadrocopter gets a new mission: 3D mapping (video)
In the future, our flying robot overlords won't just navigate terrain autonomously, they'll also report back to base with detailed 3D maps of everything they've seen -- or at least that's what this homebuilt UAV does in a video released this week. In a nutshell, MIT's combined its room-mapping Roomba with the Kinect quadrocopter radar developed at UC Berkeley, resulting in a flying contraption sure to be the envy of topographers everywhere. We're not sure that the world's robot incumbents will be too happy, though -- perhaps MIT should invest in some laser protection next.
DARPA's Urban Photonic Sandtable Display enables 3D battlefield planning without goofy glasses
You probably point and laugh at your friends when they have big, bulky 3D glasses perched on their noses in theaters. That kind of tomfoolery just won't do amongst the military brass, who frown at the slightest hint of snickering in the operations room. This new 3D system, called the Urban Photonic Sandtable Display (UPSD), should help. It's a DARPA project, a fully holographic table (no glasses required) that can be scaled up to six feet diagonally and allows visual depth of up to 12-inches. The technology comes courtesy of Zebra Imaging, which earlier wowed us with some insane 3D printouts, and the data will come from LIDAR systems like this ROAMS bot. No word on when the system will be deployed to the field, but it should allow grizzled commanders and uppity businessmen to find unobtanium deposits, even if they happen to be located right under a big 'ol tree.
Neato XV-11 robot vacuum gets its very own open source LIDAR hack
There's nothing like a little bounty to light a fire under a group of open source fanatics, is there? We saw this principle applied recently when Adafruit offered up cold, hard cash for an Open Source Kinect driver, and now one enterprising reader over at robotbox.net has gone and hacked the LIDAR unit on a Neato XV-11 robot vacuum -- and won $401 for the effort. What's this mean to you? Well, the gentleman (who goes by the nom de hack Hash79) can now read data sent from the optical ranging hardware on the vacuum to a PC. There has been a pretty enthusiastic group of hackers surrounding the device for a while now and now with a little hard work (and a $399 autonomous robot vacuum) you too can have a 360 degree scanning LIDAR with one degree accuracy and a 10Hz refresh rate. Pretty sweet, right? Video after the break.
