Lidar based 3D Tracking and State Estimation of Dynamic Objects

Generally these values are under constrained and rely 3D point cloud data obtained using Lidar sensor has been on high co-variance for doing trajectory prediction. Prediction very crucial in 3D localization and mapping of static objects is currently one of the hardest problem in autonomous within a scene. For Autonomous vehicles, Lidar point cloud vehicles due to lack of state information like position, velocity, is fused with camera to help in detecting objects like cars acceleration, yaw, yaw rate to correctly model oncoming and pedestrians. However, it hasn't been used to determine vehicle future trajectory as these cannot be determined accurately the dynamic states like velocity, yaw, yaw rate, etc of nonego using just visual camera data or constant velocity objects. The rich positional data obtained using Lidar can be models.