EXCLUSIVE: Innoviz CEO and Co-Founder Omar Keilaf Talks to Futurecar
【Summary】After announcing a new $65 million funding round today, Futurecar spoke with Innoviz CEO and co-founder Omar Keilaf and asked him some questions about the recent events, the company’s future plans, and its high-tech LiDAR technology.
After announcing a new $65 million funding round today, Innoviz Technologies hopes to be the industry leader in LiDAR technology. The company is getting ready for the commercialization and mass adoption of LiDAR technology for self-driving cars. In addition, Innoviz has recently announced partnerships with Delphi and Magna on autonomous driving technology.
Futurecar spoke with Innoviz CEO and co-founder Omar Keilaf and asked him about the recent events, the company's future plans, and its LiDAR technology.
Before founding Innoviz, Keilaf led the system and product deﬁnition efforts at ConsumerPhysics, building the world's ﬁrst handheld molecular sensor for mobile devices. Previous roles include Anobit, a company acquired by Apple.
One of the technologies we asked him about was the use of deep-learning with LiDAR to help self-driving cars make better decisions.
The Innoviz Executive Team: from left, Oren Buskila VP R&D & co-founder, Oren Rosenzweig CBO, Amit Steinberg CTO & co-founder, and Omer Keilaf CEO & co-founder
Q. Can you explain a bit about the "algorithmic layer" Innoviz employs that uses deep learning to turn 3D vision into critical driving insights?
A. LiDAR enables autonomous vehicles with vision capabilities. The more advanced the LIDAR is, the more exact the vision will be. That is the essential base layer to enable autonomous driving.
In order to develop an entire autonomous driving stack, we enable a strong base and on top of our LiDAR technology we have also embedded software capabilities and a deep learning algorithmic layer that allows us to enhance the vehicle's sensing abilities and reach high levels of perception, including computer vision capabilities to enable the vehicle to identify objects and classify them, as well as SLAM capabilities to simultaneously map and identify exact locations in the surroundings without GPS. These added layers will eventually enable the top layer of Artificial Intelligence which will allow the vehicle to make decisions based on all the data collected.
Q. What is 3D scanning as opposed to traditional LiDAR scanning?
A. LiDAR scans the environment and generates 3D mapping of a vehicle's surrounding in real time. That is a major advantage of LiDAR over other sensing solutions such cameras or radar – LiDAR enables depth-perception derived from 3D mapping. So every LiDAR will provide a 3D map. The question is how exact that mapping is and how advanced the technology is to enable superb performance metrics. Regarding traditional LiDAR scanning – see question below regarding Velodyne as this exactly relates to Velodyne's mechanical LiDAR as opposed to our own solid state design.
Q. Do you think deep learning will be a requirement for autonomous driving vehicles in the future?
A. Autonomous vehicles will require multiple technologies to enable a vehicle geared with complete autonomous driving stack capabilities, including deep-learning.
FC. This is one of the challenges for self-driving car engineers — to write software code for every possible scenario a vehicle may encounter while on the road. However, that is not a practical solution. This is where deep-learning comes in. A deep-learning neural network may eventually be used to make instant decisions to safely navigate city streets while avoiding obstacles, including pedestrians.
Q. Other companies are using roof mounted spinning LiDAR to navigate, how does the Innoviz solid-state technology differ?
A. Nearly everyone in the space agrees that LiDAR is essential to helping autonomous vehicles sense their surroundings, because LiDAR offers object detection and scanning performance that cameras and radar sensors cannot match.
The issue with current LiDAR solutions is that most of them still don't quite deliver the performance required for Levels 3 - 5 autonomous driving, and they are also too big and unreliable to be very practical. Not to mention that existing solutions can cost upwards of $75,000 or more per device (He is referring to Velodyne's HDL-64E LiDAR, which is not solid-state).
Our LiDAR technology uses proprietary systems, MEMs (micro-electric-mechanical systems) and detector designs to deliver superior sensing capabilities, even in challenging environments such as bright direct sunlight, varying weather conditions and multi-LiDAR environments.
Meanwhile, our solid-state design is smaller, more reliable and more durable than other solutions. Unlike traditional scanning methods, we have no mechanical moving parts in our design, so we are able to guarantee both safety and reliability. So from a technology perspective, our product is simply superior, but it is also farther along and closer to mass production than any other solutions.
Q. What are the main differences in the InnovizPro and InnovizOne?
A. Both InnovizOne and InnovizPro offer high-definition 3D sensing to provide an accurate and reliable scan of a vehicle's surroundings. Offering an especially wide field of view and extra long detection range in a compact, affordable, solid-state design, InnovizOne is designed to be seamlessly integrated into any vehicle and will deliver automotive grade quality for levels 3 – 5 autonomous driving. InnovizOne is expected to be available in 2019.
Meanwhile, InnovizPro, equipped with our advanced LiDAR technology, is designed to enable auto manufacturers, shared mobility companies, autonomous driving technology companies and Tier 1 suppliers to conduct effective development and testing of autonomous driving technology. InnovizPro will be commercially deployed in 2018.
FC. The InnovizPro was introduced as an alternative to expensive, bulky roof mounted LiDAR systems used in autonomous car testing. The InnovizPro is an attractive option for companies who will just use it for development and testing phase of their self-driving systems.
Q. Can you name any automotive partners that will use Innoviz Lidar, or future plans?
A. We have signed partnership agreements with top Tier 1 companies Delphi Automotive and Magna International to integrate our advanced LiDAR technology into their autonomous driving systems. In addition, we are working with other leading stakeholders in the industry.
Q. Can you elaborate more on the partnership between Innoviz and Delphi? What will Delphi bring to Innoviz?
A.We are working closely with both Delphi and Magna in order to offer the best solution to OEMs and to help the market get to mass commercialization as quickly as possible. Innoviz's proprietary LiDAR sensing solutions will be integrated into Delphi's and Magna's systems to provide automakers with a comprehensive portfolio of autonomous driving technologies. Leveraging these partnership we will be able to scale more quickly.
Q. Can you explain how the object detection capabilities work? Is a separate hardware module required, such as Nvidia's Drive PX platform?
A. As mentioned above, Innoviz is also developing advanced computer vision and software capabilities which are embedded to our LiDAR technology. Through software and sophisticated algorithms we are able to process the 3D map generated by the LiDAR and identify and classify objects in the surroundings. Our SLAM – simultaneous localization and mapping - enable us also to identify exact location of the vehicle without the usage of GPS.
Q. Velodyne is a major player in the LiDAR field here in Silicon Valley, how does Innoviz's technology differ from that from Velodyne?
A. Both InnovizOne and InnovizPro offer high-definition intelligent 3D sensing to provide an accurate and reliable scan of a vehicle's surroundings. The devices will leverage Innoviz's proprietary system, MEMs and detector design to deliver superior performance to existing LiDAR solutions, even in challenging environments such as night-time, bright sunlight and varying weather conditions.
In addition, the company's proprietary solid-state design introduces a device without mechanical moving parts compared to traditional LiDAR, enabling Innoviz to offer a smaller product footprint at mass market price.
FC. Several other companies are working on LiDAR technology along with Innoviz. Another Israeli startup Oryx recently raised $50 million to develop new lidar for autonomous vehicles, In the U.S. Last year, Ford Motor Co. and Chinese Internet giant Baidu invested $150 million into Silicon Valley based LiDAR company Velodyne.
However, with the recent funding round, IP, and partnerships with leading OEM automotive suppliers Delphi and Magna, Innoviz looks to become a major player in LiDAR technology for self-driving cars.
Originally hailing from New Jersey, Eric is a automotive & technology reporter covering the high-tech industry here in Silicon Valley. He has over 15 years of automotive experience and a bachelors degree in computer science. These skills, combined with technical writing and news reporting, allows him to fully understand and identify new and innovative technologies in the auto industry and beyond. He has worked at Uber on self-driving cars and as a technical writer, helping people to understand and work with technology.
GM to Invest $2.2 billion in its Detroit-Hamtramck Factory to Build Only Electric Vehicles
Mercedes Benz Scales Back Production of the Electric EQC Due to Battery Shortage
BMW Invests $25 Million in Silicon Valley-based Software Motor Company
Cruise Debuts the 'Origin', an Autonomous Urban Robotaxi Built for Ride-Hailing
Tesla Announces it Will Use Baidu Maps in its China-Made Vehicles
Uber is Testing a New Feature in California Allowing Some Drivers to Set Their Own Fares
The NHTSA Will Review a Petition to Investigate Sudden Acceleration Complaints on Tesla Vehicles
Volvo to Build an Electric Vehicle Battery Plant in the U.S.
- Next Surge of EVs Coming in 2020, Claims Engadget
- EV Startup NIO to Partner with Intel’s Mobileye on Self-Driving Cars
- Electric Vehicle Startup Rivian Secures $1.3 Billion in New Funding from T. Rowe Price & Others
- GM’s Luxury Brand Cadillac May Become Fully-Electric by 2030, Exec Says
- MIT Teaching Autonomous Cars to See Around Corners
- Tesla Model 3 Prices Hike and Entry-level Variant Gets Extended Range
- Mobileye CEO Claims Robotaxis Are Coming as Soon as 2022
- Kia Motors Reveals its $25 Billion Electrification Strategy for 2020-2025
- Volvo to Build an Electric Vehicle Battery Plant in the U.S.
- Here’s What You Need to Know About Neuron's Electric Semi and Pickup Truck