Technology breakthrough reduces costs of driverless cars

In a move that could potentially bring down the potential cost of future driverless cars, Velodyne has unveiled its Velodyne Puck (VLP-16) 3D LiDAR (Light Detection and Ranging) system at an introductory price of US$7,999.

It is expected to dramatically expand the use of 3D LiDAR technology sensors in robotics, unmanned aerial vehicles and factory automation applications. “There’s a great deal of interest in this market generally and in the Velodyne Puck class of product specifically. What’s new isn’t simply recording data with the sensor but using real-time, 3D LiDAR sensors to actually help avoid collisions – in the air, to stay clear of other UAVs, and likewise on the ground to find a suitable landing site when faced with the unknown in a military hot zone or an area in the wilderness,” said Wolfgang Juchmann, Velodyne’s Director of Sales and Marketing at the unveiling of the system.

The unit’s low power consumption (<10W), light weight (about 830 grams), compact footprint (Ø100mm x 65mm), and dual return option make it ideal for UAVs. Velodyne’s LiDAR Puck supports 16 channels, ~300,000 points/sec, a 360° horizontal field of view and a 30° vertical field of view, with ±15° up and down.

Velodyne’s expertise with 3D laser distance measurement started when its founders Dave and Bruce Hall entered the Defense Advanced Research Projects Agency (DARPA) – sponsored 2005 Grand Challenge race across the Mojave Desert for unmanned vehicles. Team DAD (Digital Audio Drive), travelled 6.2 miles in the first event and 25 miles in the second. The team then developed the laser-based system that laid the foundation for Velodyne’s current products. The first Velodyne LiDAR scanner was about 30 inches in diameter and weighed close to 100 lbs. By focusing on the commercialization of the LiDAR scanner instead of competing in subsequent challenge events Velodyne was able to dramatically reduce the sensor’s size and weight while improving performance. Velodyne’s HDL-32E LiDAR sensor will ride atop the RoboSimian a headless, ape-like robot at the 2015 DARPA Robotics Challenge. Created by NASA’s Jet Propulsion Laboratory, RoboSimian uses Velodyne’s spinning LiDAR sensor as a key element of its perception system. The sensor, which is capable of rotating a full 360° up to 20 times per second, enables the robot to look between 10° up and 30° down.

Automotive Industries (AI) asked Juchmann what impact the introduction of the Puck range will have on the market. Juchmann: The Puck will help car manufacturers move forward with highly automated vehicles. The Puck’s significantly smaller size will enable OEMs to strategically place the VLP-16 semi-hidden within the periphery of the vehicle. With the 10 times reduction in unit cost (compared with the original HDL-64E), the Puck will permit automotive R&D teams to experiment with multiple VLP-16s without placing an additional burden on budgets.

AI: How did Velodyne manage to slash prices so dramatically? Juchmann: Design for high-volume manufacturing and automation during the manufacturing process helped with this aggressive offering, while still providing 16 independent channels that help to capture cars, bikes, pedestrian in 3D to avoid collisions. We do believe it’s a potential game-changer. Competitive LiDAR providers typically offer only single-channel or quadruple-channel LiDAR devices, with less than a 360° horizontal field of view. The race is on for them to either add additional channels at the same price, which will not be an easy task, or to further reduce their pricing to differentiate these limited channel systems from Velodyne’s 16-channel Puck.

AI: Tell us about the automotive OEMs and Tier I suppliers that are customers of your company’s LiDAR division. Juchmann: While Google and its self-driving car may be the most famous customer of ours in the “automotive” space – using Velodyne’s HDL-64E sensor for all three generations of its R&D vehicles – we are proud to say that virtually every automotive OEM and Tier 1 supplier has at least one of our sensors in its R&D department. Ford Motor Company has mounted four of the HDL-32Es on its Ford Fusion Hybrid R&D vehicles, which have been displayed at previous automotive and mobility shows. GM’s EN-V2 sports a spinning HDL-32E, providing 3D, surround views that enable the EN-V2 to avoid collisions. Tier 1 supplier Bosch has an R&D vehicle equipped with an HDL-64E, and the company recently announced that it will also offer LiDAR sensors by 2020.

AI: What has been Velodyne’s relationship with Google, and how do you see it developing further? Juchmann: Velodyne’s relationship with Google has been a very productive and visionary one. Google – itself a Silicon Valley company and former start-up – has shown the world that self-driving cars are becoming a reality. Google’s approach to technology in this project has always been to prioritize function over design. The company’s engineers realized early on that Velodyne’s 3D LiDAR sensors are the most functional solutions on the market. Google has not shied away from placing the sensors in a prominent position, on the roof of three generations of vehicles (Prius, Lexus and the latest generation without gas pedal and steering wheel), since that is clearly the most functional position for a surround view.

AI: Will 2015 be the “Year of the Robot” for Velodyne? Juchmann: It does seem that 2015 is poised to be the Year of the Robot. With the introduction of the Puck, Velodyne has opened up new markets for 3D, real-time LiDAR scanners. One of those markets is factory automation, as Amazon and other warehouse giants have implemented in order to move quantities of goods around in real time. The path in which warehouse robots move can be pre-programmed. If, however, something unforeseen happens, the robot needs to be smart enough to recognize a difficult situation by analyzing the environment using 3D LiDAR vision and then make a decision on how to deal with objects that intrude in the pre-planned path. Likewise, several home improvement retailers have announced that they will use robots to greet customers and help them find items by guiding them through the store. In both situations, collision avoidance is paramount. And in a department store teeming with customers, there are plenty of “obstacles” that can’t be on some pre-planned route.

AI: How will the RoboSimian project further propel Velodyne’s LiDAR division into future technological breakthroughs? Juchmann: The RoboSimian project, created by NASA/JPL, is part of one of the DARPA Grand Challenges that the federal government sponsors periodically to nurture new technology in emerging application areas – in this case, deploying machines to go where people cannot or should not. As it happens, Velodyne’s LiDAR division is the outgrowth of one such competition, so we’re especially excited that DARPA has decided to push the world of “humanoid robots” to the next level with this new competition –and that our 3D, real-time LiDAR sensors will play such a crucial role in executing some very complex tasks during the challenge.