CV for Traffic Scene Understanding
Daniel Göhring’s current research focusses on vehicle control, navigation, sensor fusion, object tracking, and computer vision. In particular he is interested in computer vision approaches for traffic scene understanding, including vehicle classification and pedestrian recognition, to combine this data with data from other sensors and to finally achieve an accurate prediction of other traffic member’s trajectories. On the behavioral side he is interested in applications for swarm intelligence, to achieve a highly adaptive traffic behavior.
Xiuyan Guo is currently applying SLAM (Simultaneous Localization and Mapping), specifically speaking FastSLAM, to address autonomous vehicle’s localization and mapping problem. Autonomous vehicles are mobile robots as well. But they are distinguished from ordinary mobile robots in safety aspect. Therefore, they need much higher accuracy in localization and mapping. He is working to improve the accuracy.
Reactive Path Planning based on Swarm Behavior Methods
In his current research Simon Rotter focuses on applications of swarm approaches for path planning in the autonomous car. In his work he is taking advantage of other cars’ positions, velocities and trajectories. This data is used for reactive path planning based on biologically inspired swarm behavior methods.
Localisation by Employing Stereo Cameras
Robert Spangenberg’s scientific interest lies in the usage of stereo cameras to improve the localisation of autonomous vehicles. A sparse set of landmarks is used together with odometry as input to the localization algorithm. The base for the landmark detection is fast and robust stereo matching based on Semi-Global Matching.
Complex Traffic Scenarios
As a PhD candidate Steffen Heinrich works for the R&D department of a German OEM in the field of automated driving. He is advised by Prof. Raúl Rojas at the Freie Universität Berlin. The main focus of his work is on motion planning algorithms in complex traffic scenarios.
Within his research Andreas Hartmannsgruber in “Autonomous Driving” he is mainly focused on the motion planning of autonomous vehicles. This includes the high-level behavior planning as well as the low-level path and velocity planning (trajectory planning). Therein his focus is to adapt swarming behavior from biological and social science to the motion planning in crowded traffic. According to his research, he also considers the behavior recognition of the surrounding traffic participants.
Daniel Lammering’s field of research covers vehicle E/E-architectures and functional safety aspects related to autonomous driving. Therefore he considers new redundancy concepts and generic approaches for the system- and software-architecture. For the intercommunication, future bus systems, like automotive Ethernet, have to be taken into account. Moreover new methods of measurement to determine the probability of errors from the driving algorithms must be derived. One of his goals is a feasible technical safety concept for autonomous driving.