AEye introduces “iDAR” perception system for autonomous vehicles; fusion of LiDAR, computer vision and AI
AEye, a startup with investors including Kleiner Perkins Caufield & Byers, Airbus Ventures and Intel Capital (earlier post), announced the AE100, its first robotic perception system for autonomous vehicle, ADAS, and mobility markets.
The AE100 is a solid state, cost-optimized system based on AEye’s iDAR (Intelligent Detection and Ranging) technology. iDAR integrates three core components: the first agile MOEMS (Micro-optical Mechanical) LiDAR, pre-fused and bore-sighted with a low-light HD video camera; distributed artificial intelligence; and software-definable and extensible hardware. AEye says that this combination enables iDAR to adapt dynamically to the real-time demands of automated and autonomous vehicles.
Leveraging a more powerful, yet safer, 1550-nm laser, AEye’s iDAR-powered AE100 system can always interrogate 100% of the scene, while typical fixed pattern LiDAR systems are only capable of interrogating 6% of any scene, due to the large vertical gaps inherent in raster or spinning systems. AEye says that the embedded intelligence in the AE100 systems offers features including:
Coverage: The AE100 can use dynamic patterns for mapping the environment, as it is not tied to one fixed mode. In evaluating any given scene, AE100’s software definable scanning delivers more than 10x the 3D resolution over legacy systems.
Speed: The AE100 is three times faster than current LiDAR systems, and does not miss any objects between scans while identifying and solving any temporal anomalies. This reduces scan gaps, resulting in more than 25 feet of faster response distance at average highway speeds—more than two car lengths.
Range: The AE100 extends effective range at comparable resolution by 7-10x over currently deployed LiDAR systems.
AEye’s iDAR technology mimics how a human’s visual cortex focuses on and evaluates potential driving hazards. Using a distributed architecture, iDAR critically and dynamically assesses general surroundings, while applying differentiated focus to track targets and objects of interest, while always critically assessing general surroundings. That enables accessible direct detection for every pixel and voxel in each frame.
AEye says that this scalable, software-definable approach enables iDAR to deliver higher accuracy, longer range, and more intelligent information to optimize path planning software, thereby enabling improved autonomous vehicle safety and performance at a reduced cost.
iDAR’s true-color LiDAR (TCL) instantaneously overlays 2D real-world color on 3D data, adding computer vision intelligence to 3D point clouds. By enabling absolute color and distance segmentation, and co-location with no registration processing, TCL enables the quick, accurate interpretation of signage, emergency warning lights, brake versus reverse lights, and other scenarios that have historically been tricky for legacy LiDAR-based systems to navigate.
The scorecard of performance characteristics used to judge current LiDAR-only systems must be rewritten from basic serial data collection to intelligent and value-added information creation. Up until now, basic capabilities such as resolution, speed, and range have been the only metrics tracked. iDAR brings new capabilities that redefine robotic perception performance, such as target acquisition time and blob revisit rates, where iDAR delivers a 3000x improvement.
The result is a system that has the capability to very quickly or instantaneously calculate things such as 3D multivariable velocity and acceleration, true color point clouds, and augmented single centimeter level localization for radically improved control, innovation and performance of path-planning software.—John Stockton, vice president of product at AEye
Built on the iDAR technology platform, the AE100 is designed specifically for the automotive and mobility markets and is the first of a family of products to support autonomous vehicles, ADAS and general mobility. It will have a limited release in mid- 2018, and a larger commercial release later in Q3.
Given the adaptive nature of iDAR, specifications will be software-definable for the first time by frame or region of interest or pixels/voxels within specific frames. This gives OEMs and Tier 1s ultimate control of perception.
Customizable general performance specifications include up to 200Hz frame rates, less than 60μs object or blob revisit rates, software-definable resolution of .09 degrees H/V, and a maximum range of 300-400 meters. In addition, iDAR’s seamless integration of 2D and 3D data allows for virtual “plug-and-play” extension of all computer vision algorithms into a 3D point cloud with no added latency.
In conjunction with our OEM and Tier 1 partners, we have developed a product that addresses the complex system requirements, high standards and performance demands of our customers. A key objective was to design a solid state modular platform that is software definable in order to increase reliability and optimize the cost. We created an easy transition from first generation spinning LiDAR hardware that allows path planning software teams to plug and play the AE100 as they replace their legacy systems. It requires no software changes, and enables them to bring in the more advanced features over time.—Barry Behnken, AEye’s VP of Engineering
The AE100 will be available beginning this summer to OEMs and Tier 1s launching autonomous vehicle initiatives. The company’s iDAR Development Partner Program is currently available to partners interested in integrating iDAR into their products. At CES in Las Vegas this week, AEye will be offering private demonstrations and in-car on-road experiences by appointment.