Innoviz Co-Founder and CEO Omer Keilaf kicked off 2022 with an informative LiDAR webinar. During this event, he took an in-depth look at LiDAR design requirements for safe Level 3 driving where drivers can safely turn their attention away from driving tasks. This conversation was important amid urgent calls to standardize the technology and language used to describe autonomous driving.
Keilaf drew information from years of Innoviz’s experience collaborating with leading automotive original equipment manufacturers (OEMs) and Tier-1 suppliers to educate audiences. Along with this background knowledge, he presented the report he penned that defines key industry terms and LiDAR specifications that are not well known to general audiences.
This presentation offered insight into what carmakers require of their LiDAR systems to make their automated vehicles safe for Level 3 highway driving. Some of these topics included the SAE autonomy levels, driving decision requirements, LiDAR, camera and radar perception requirements, and LiDAR sensor requirements.
Innoviz applied these principles to its designs for the new, high-performance InnovizTwo, which Keilaf introduced in the webinar, and said, “It’s really time that the automotive space started doing some standardization for safety in this AV market.”
Continue reading to get some of the inside scoop on LiDAR technology.
Highlights from the LiDAR webinar:
Long-Range Object Detection for High Speeds
One of the greatest challenges for automated driving is the vehicle’s ability to detect small, extremely low reflectivity objects, like a tire or pothole, at long distances while driving at highway speeds.
This type of braking event requires long-range object detection and the highest possible resolution and frame rate to ensure objects and obstacles can be detected and avoided at high speeds. In the webinar, Keilaf explained the difficulty carmakers have in selecting LiDAR sensors that are safe for highway driving since more requirements are posed on the sensors.
Innoviz kept this in mind when developing the InnovizTwo LiDAR technology, to provide a solution for carmakers who need reliable LiDAR technology for speeds upwards of 60 mph.
Automated driving is a no-brainer for highway applications. The road environment remains fairly consistent, and human drivers tend to lose focus when they are driving straight for a long period of time, or are in bumper-to-bumper traffic. Sudden stops are also harder on highways, especially in the rain or icy road conditions. LiDAR has the potential to support humans in identifying small obstacles on the road and reduce highway collisions caused by inattention, misbehavior, or distraction, which is where InnovizTwo comes into play.
View from Innoviz LiDAR on a highway
Model Designs Influencing LiDAR Mounting Locations
Consumers are likely most familiar with Innoviz’s competitors’ LiDAR, typically seen mounted on top of cars, but this is only one approach to how LiDAR is used. In this presentation, Keilaf urges OEMs to consider all possible mounting locations for all models in which a LiDAR will be installed, to ensure an unobstructed field of view. He also discusses how a model’s design is the best indicator for LiDAR designs and placement.
In the near future, LiDAR technology will be more discreetly placed within the rearview mirror, behind the grille or at the rear window or tailgate of a vehicle. Similarly to how cameras are positioned on vehicles, LiDARs will be smaller and more creatively integrated into cars to not impact design. Full 360-degree coverage for safety does not need to impact vehicle aesthetics!
Environmental Conditions Influencing LiDAR Performance
Along with a model’s design, the Level 3 Highway Standards report describes how weather, lighting, road conditions, and blockage of the sensor can also reduce system performance, resulting in a longer braking time. In this webinar, Keilaf explains how LiDAR must be designed to take into account dynamic conditions while ensuring optimal performance.
Innoviz tests product capability and durability in a variety of scenarios, including rain, dirt, dust, and high heat. All of Innoviz’s products are automotive-grade, meaning they’ve been tested like any other car part and perform as designed in all conditions.
Drops of rain on camera lens vs. Innoviz LiDAR sensor
“The faster the car drives autonomously, the more time it needs to react, to see further away, and predict and understand the scene better in order to drive safely,” said Keilaf in the webinar.
To better understand how to meet LiDAR performance requirements, Keilaf and his team gathered and combined the LiDAR data on automotive requirements from OEMs that they obtained with some of the great technology used for InnovizOne. Using these resources, Innoviz was able to release InnovizTwo for safe and comfortable highway driving.
The Final Takeaway
To end the webinar, Keilaf gave viewers a final takeaway from the report. He stressed the importance of autonomous vehicle manufacturers carefully evaluating the claims of LiDAR vendors, as safe and reliable autonomous system performance requires many trade-offs.
With the understanding of key LiDAR terminology, comes the ability to directly compare performance and better define what it means for a highly-automated vehicle to be functionally safe.
Learn more about the InnovizTwo.