Mar 24 2020
1:30 pm - 3:00 pm
Track Names: AUTONOMOUS TECHNOLOGIES, Tuesday 1:30 - 3:00
Session Date: Mar 24 2020 1:30 pm - 3:00 pm
Solving the Data Dilemma
The automotive industry is looking to evolve from ADAS to autonomous systems, which means massive quantities of data feeds, and new requirements for computational capability and power consumption. Fortunately, sensor capability and reliability have dramatically increased. In this session, AEye will discuss how advanced, LiDAR-enabled sensor systems can reduce data streams without having to make compromises on resolution and field-of-view requirements necessary for early detection and classification. How should these systems be architected, and how can camera deficiencies be covered by LiDAR, especially if the LiDAR is cued by those deficiencies? This session will cover these topics, as well as how these systems can filter out the most important aspects of a given scene, enabling the vehicle’s perception system to target only the salient data needed to build the environment model and safely navigate roadways.
Autonomous 3D Lidar for Underground & Indoor Mapping
3D Lidar Mapping in unstructured and GPS denied areas poses significant technological challenges. Mining and construction companies have been searching for a way quickly map and catalogue their assets. Without a GPS signal or pre-established reference points, it is difficult geo-reference 3D Lidar data into one cohesive point cloud. Considerations must also be made for physically moving the Lidar sensor around the environment to create the scan. In an underground mine, or new construction building, sending people in can be dangerous and costly. Velodyne, in combination with Paracosm, and Exyn Technologies have developed a breakthrough autonomous drone solution for underground and indoor mapping. The solution utilizes a Velodyne Puck sensor mounted to a drone and flown through the area to generate a 3D map. Exyn Technologies has implemented several new software technologies including SLAM and sense-and-avoid in order to navigate the craft autonomously. While the UAV is navigating, it is also collecting data which can be post-processed into a high definition 3D Lidar point cloud map. Using autonomous algorithms, the drone continues flying until it has completed the map or is running low on battery. The post-processing and asset mapping is performed by industry leader Paracosm. The raw lidar data is pushed through Paracosm’s post-processing software in order to create map and identify areas of interest. The data is converted into easily recognizable file types and can be integrated into other workflows such as Building Information Management (BIM) programs. Frank Amir and Nader will present and answer any questions about the various systems involved.
Q&A and panel discussions with session presenters
There will be a 15 minute Q&A /panel discussion with the presenters of the Autonomous Technologies session.