Time filter

Source Type

Eggenstein Leopoldshafen, United States

Mathematical models can be improved with respect to problem conformity and statistics by incorporating a priori information. Therefore the knowledge needs to be formulated in a mathematical expression being suitable for optimization. The focus is on models which are described by functions. Such models are characteristic diagrams, density functions, step responses, curves and nonparametric representations. Especially the treatment of the following properties and constraints is considered: pointwise and intervalwise constraints, continuity and differentiability as well smoothness constraints, nonnegativity, convexity/concavity, unimodality/antimodality, symmetry, sector constraints, various kinds of monotonicity, pole singularity, and asymptotic constraints. © Oldenbourg Wissenschaftsverlag. Source

Werling M.,BMW AG | Heidingsfeld M.,University of Stuttgart | Reinisch P.,BMW AG | Groll L.,Institute For Angewandte Informatik
At-Automatisierungstechnik | Year: 2014

Backing up a trailer can be a challenge, not only to inexperienced recreational drivers. We therefore develop two feedback controllers as themain components of a driver assistance system which supports the driver with steering inputs. Based on the kinematics of the general one-trailer-system, we first derive an input/output-linearizing control lawthat asymptotically stabilizes a given curvature for the trailer. This enables the driver to directly steer the trailer, e.g. by means of a turning knob, such that the trailer will be automatically prevented from jackknifing. The control task is then modified and solved so that the assistance system can also take over the complete stabilization of given paths. In combination with a path planning algorithm this enables, for instance, automated parallel parking. The complete system is implemented on a rapid prototyping environment and evaluated in real-world scenarios. Source

Werling M.,BMW AG | Gutjahr B.,BMW AG | Galler S.,BMW Peugeot Citroen Electrification | Groll L.,Institute For Angewandte Informatik
At-Automatisierungstechnik | Year: 2015

In many critical traffic situations sole breaking cannot avoid a collision. Especially when pedestrians suddenly enter the driving corridor, the stopping distance is often too long, so that the vehicle cannot come to a full stop fast enough. We therefore propose an optimal control theory based trajectory optimization algorithm as an important component of an anti-collision-avoidance system. By means of a suitable state and time transformation, planning the avoidance trajectory can be formulated as a linear-quadratic optimization problem. According to the theory of Riccati state control design for finite optimization horizons its solution can be computed very efficiently. This renders possible the implementation on automotive control units with cycle times in the range of milliseconds. The proposed algorithm is validated in a pedestrian setup by means of a test vehicle. © 2015 Walter de Gruyter Berlin/Boston. Source

Werling M.,BMW AG | Groll L.,Institute For Angewandte Informatik | Bretthauer G.,Institute For Angewandte Informatik
At-Automatisierungstechnik | Year: 2012

Handling of dynamic street scenarios requires the stabilization of time-critical maneuvers such as obstacle avoidance and merging into the traffic flow. As for autonomous driving, conventional path-based approaches quickly reach their limits here. We therefore propose an input/output linearizing trajectory stabilization concept. It explicitly accounts for the time t during maneuver execution enabling precise driving amongst moving obstacles. © Oldenbourg Wissenschaftsverlag. Source

Discover hidden collaborations