Laser Scanning
-
Uncertainty Estimation of LiDAR Scene Semantic Segmentation (DFG i.c.sens)Despite the capability of advanced deep learning models to accurately assign semantic labels to LiDAR point clouds, there is a notable lack of methods for uncertainty quantification. However, the estimation of uncertainty is essential for assessing the reliability of any prediction, particularly for safety-critical systems such as autonomous vehicles that rely on real data, including LiDAR point clouds. These systems need not only to perceive their surroundings but also to quantify uncertainty to avoid over-reliance on potentially erroneous predictions. Two primary types of uncertainty are generally distinguished: epistemic and aleatoric. Epistemic uncertainty, which arises from the model itself, reflects the reliability of a model’s predictions, whereas aleatoric uncertainty stems from characteristics inherent in the data.Led by: apl. Prof. Claus BrennerTeam:Year: 2022Funding: DFG Graduiertenkolleg i.c.sens
-
Collaborative acquisition of predictive mapsSelf-driving cars and robots that run autonomously over long periods of time need high precision and up-to-date models of the environment. Natural environments contain dynamic objects and change over time. Since a permanent observation of “everything” is impossible and there will always be a first time visit of the changed area, a map that takes into account the possibility of change is needed.Team:Year: 2017Funding: DFG-Graduiertenkolleg i.c.sens
![]()
![]()
Robotics
-
Landmark-based localizationWithin the project, new approaches are developed for a highly accurate localization of vehicles relative to their environment. Furthermore, it is analyzed how detailed descriptions of the environment can be used for interpreting the scenery, for example for active driver assistance systemsTeam:Year: 2017
![]()
![]()
-
Collaborative acquisition of predictive mapsSelf-driving cars and robots that run autonomously over long periods of time need high precision and up-to-date models of the environment. Natural environments contain dynamic objects and change over time. Since a permanent observation of “everything” is impossible and there will always be a first time visit of the changed area, a map that takes into account the possibility of change is needed.Team:Year: 2017Funding: DFG-Graduiertenkolleg i.c.sens
![]()
![]()
