In a recent round table on the 360 Blog, a pair of Velodyne visionaries share their hopes for self-driving cars in our everyday lives. Mike Jellen, President and CCO, wants to leave home and “step into a car that’s set up like an office.” With his hands free from the steering wheel, Mike can turn his attention to preparing for the day and conducting more business.
Meanwhile, Mircea Gradu, Senior Vice President of Validation, wants his autonomous car to function like a valet service for practical and fun purposes. But what about when his valet car needs maintenance? “I can send it to the shop on its own,” says Mircea. “Then a loaner valet car shows up to take me to work. When my car is serviced, it drives itself back home.” But Mircea’s dream self-driving car isn’t all about work. “The autonomous valet can take me to a restaurant or museum with friends… [and] it can help me with one-way trips for downhill mountain biking, down-river kayaking, and down-wind kitesurfing.”
Learn about Mike Jellen’s 3D management style for the autonomous revolution!
“We think the biggest unsolved problem for autonomous driving at highway speeds is avoiding road debris,” writes Velodyne founder and CEO, David Hall, in his December 2017 article published on LinkedIn and the 360 Blog. As Hall explains, “most road debris is shredded truck tire—all black material on a dark surface.” For self-driving cars to reach speeds necessary for highway travel, sensors must quickly identify distant obstacles, from fallen trees to tires. Hall elaborates that an “autonomous car needs to see further out, with denser point clouds, and higher laser repetitions.”
So, Hall developed the new VLS-128, which is Velodyne’s most advanced sensor ever and 70 percent smaller than the model it replaces, the HDL-64. With double the laser channels, triple the beam density, and double the zoom resolution, this new sensor offers a 300-meter range. “I believe this will solve the high-speed roadway recognition problem,” writes Hall. “[And] lead the autonomous revolution to highway speeds.”
Like to go fast? Check out the annual autonomous event self-racing cars!
Imagine a helicopter with advanced lidar sensors, hovering over a tropical forest, collecting data for a 3D map. Or a surveyor wearing a backpack, equipped with a curious lidar tower, walking through giant sequoia groves for a volumetric study on the world’s largest trees. Or preservationists, using mobile lidar
scanners to create a digital representation of the Great Wall of China. Hopefully, imagining these scenarios is easy—because they’re already happening, thanks to the adventurous exploits of Green Valley International.
In a recent post on the 360 Blog, Green Valley founder and CEO, Dr. Qinghua Guo, discussed lidar’s value to environmental management. “Our company motto is map the world in 3D,” says Dr. Guo, explaining their LiBackpack is “perfect for under-canopy mapping,” while the LiAir provides “detailed information on the top of canopy.” The resulting maps provide foresters and ecologists with more accurate representations of forest density, tree height, and timber volume. About the Great Wall, Dr. Guo says, “The possibilities from that scan are exciting. You can see individual bricks and even measure them with a millimeter of precision.”
Watch this cool demonstration video about Green Valley’s LiBackpack!
“I envision the road trip of the future to be more like a train ride in a car,” says Wolfgang Juchmann, Vice President of Business Development for AutonomouStuff. In a recent article on the 360 Blog, Juchman imagines how the efforts of his company—which integrates various software and hardware technologies into automotive perception kits—might transform road trips of the future. Chief among his visions is that travelers will enjoy the ride while the cars do the driving. “If there’s beautiful scenery, I can take the wheel and drive myself. But, for the boring parts or in traffic, I will let the car drive while I relax, read, or communicate via text or phone.”
Several hypothetical road trip vehicles come from Honda Motors’ concept car series, the Great Journey. Resembling German Playmobil toys crossed with Disney’s Cars film series, one concept is called the Desert Train. This four-part vehicle includes a tractor-like drive unit, a boxy living unit, a water tank, and gear trailer—complete with concert-quality speakers. Add autonomy to the mix, and it’s not hard to image future road-trip vehicles becoming what Velodyne’s Mike Jellen calls “monorail-like viewing platforms.” In such vehicles, road-trippers could sit at a table and enjoy the view while eating dinner or playing cards.
Beginning in the early 17th century, countless maps depicted California as an island floating just off-shore from North America. The reasons for the error, which persisted for 50 years, relates to how map-making has evolved. During the Age of Exploration, maps were beautiful hand-drawn affairs based upon descriptions by European explorers plus (hush-hush) discreet copying of previous maps, which tended to compound errors. As navigational technology advanced, more precise observations of latitude and longitude became possible, making maps increasingly accurate. During the 20th century, aerial photography was developed for ground mapping, military reconnaissance, and eventually remote sensing.
Hence the science of photogrammetry was born, as described in a recent 360 post, which estimates ground features from photographs. Photogrammetry has advantages and disadvantages. During mid-day sun, it produces startling full-color images that are familiar to surveyor’s eyes. This tends to work fine for measuring shapes and distances from the top down, but vertical position accuracy—say a building height from street level—can be imprecise. Here’s where lidar comes in, providing three-dimensional data at all angles. lidar can be used day or night. And the data streams are much smaller and easier to process, providing real-time positional information—including exact building height or object dimensions. This suggests, someday soon, aerial lidar surveyors may comb the skies as part of a new evolution in mapping precision.
As 2017 comes to an end, the National Safety Council estimates that 2016 may have been the deadliest year on U.S. roadways in nearly a decade. The council’s data shows that “as many as 40,000 people died in motor vehicle crashes in 2016.” That’s a 6 percent increase in crash deaths since 2015 and a 14 percent increase since 2014, which marks “the most dramatic two-year escalation in 53 years.” [i] For a 2008 report, researchers for the National Highway Traffic Safety Administration studied over 5,000 accident scenes and interviewed participants over a 2.5-year period to determine crash causation. They found that 41 percent of crashes were due to human recognition errors, such as inattention and distraction, while 34 percent were attributable to human decision errors, such as aggressive driving or speeding.[ii]
As these statistics show, human error among drivers is a major public safety issue. In recent years, a combination of government legislation and automaker innovation has led to the introduction of Advanced Driving Assistance Systems. ADAS include features like lane departure warnings, collision avoidance systems, and back-up cameras. The goal of ADAS is to mitigate, to the extent possible, the hazards associated with human driver error.
Meanwhile, one of the primary goals of the autonomous revolution, which seeks to create fully self-driving cars, is to eliminate human decision and recognition errors from driving situations. Fully autonomous, level 5 vehicles would be controlled by an artificially intelligent computer, using a combination of sensing systems, such as lidar, radar, and cameras that see and sense the roadways during day and night. Autonomous systems are designed to drive cars non-aggressively at safe speeds while not succumbing to human failings such as cell phone distractions or drowsy inattention. And while such technological gifts may sound like fantasy, deployment may be closer than most people realize. As Velodyne’s Director of Automotive Sales and Marketing, John Eggert, says, “Within a decade, high definition lidar will be a standard component on automobiles, and no more exotic than airbags, antilock brakes, or cameras once were.”
Early in 2018, Velodyne’s 360 Blog will publish a more in-depth look at the safety potential lidar can offer to the world, both through autonomous driving and other applications. In the meantime, please enjoy the holidays and drive safely and attentively wherever you go.
[i] “2016 Motor Vehicle Deaths Estimated to be Highest in Nine Years” National Safety Council, accessed December 5th, 2017.
[ii] “National Motor Vehicle Crash Causation Survey: Report to Congress.” U.S. Department of Transportation, National Highway Traffic Safety Administration. July 2008.