DYNAMIC CONDITION-BASED POINT CLOUD GENERATION FOR AUTONOMOUS DRIVING

A method includes determining, using a light detection and ranging (LiDAR) system, a first point cloud representing an environment of a vehicle, determining, based on the first point cloud, an autonomous driving scenario of the vehicle, and determining one or more conditions associated with the autonomous driving scenario. The method also includes determining, based on the autonomous driving scenario and the one or more conditions, one or more extrinsic parameters for the LiDAR system, the one or more extrinsic parameters for the LiDAR system representing a geometric relationship between the LiDAR system and the vehicle, and re-configuring the LiDAR system based on the one or more extrinsic parameters. The method further includes determining, using the re-configured LiDAR system, a second point cloud representing an environment of the vehicle, and autonomously operating, using the second point cloud, the vehicle.