Christoph Winter

Mr. Christoph Winter

Deutsches Zentrum für Luft- und Raumfahrt, Weßling, Germany
orcid

Research Focus & Metrics

Academic Career Length

31Years (1996-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications

1 June 2015
Verified
Conference Paper
Closed access
PRPeter Ritzer
Peter Ritzer
CWChristoph Winter
Christoph Winter
JBJonathan Brembeck
Jonathan Brembeck
This work describes an advanced path following control strategy enabling overactuated robotic vehicles like the ROboMObil (ROMO) [1] to automatically follow predefined paths while all states of the vehicle's planar motion are controlled. This strategy is useful for autonomous vehicles which are guided along online generated paths including severe driving maneuvers caused by e.g. obstacle avoidance. The proposed approach combines path following, i.e. tracking a plane curve without a priori time parameterization of a trajectory, with feedback based vehicle dynamics stabilization. A path interpolation method is introduced which allows to perform the path following task employing a trajectory tracking controller. Furthermore a tracking controller based on I/O linearization and quadratic programming based control allocation is proposed which allows employing the vehicle's overactuation in an optimal manner. The work concludes by a simulative evaluation of the controller performance.

26 December 2016
Verified
Conference Paper
Closed access
CWChristoph Winter
Christoph Winter
PRPeter Ritzer
Peter Ritzer
JBJonathan Brembeck
Jonathan Brembeck
This paper describes a real-time capable online path planning on roads and its experimental investigation for the highly maneuverable robotic electric vehicle research platform ROboMObil. The path planning algorithm is based on an efficiently solvable and compact optimization problem and contributes to the autonomous driving of centralized controlled vehicles. The necessary development from a global offline problem formulation towards an online receding horizon method is shown, which is capable of taking environmental changes into account. The online planned path together with a generated velocity profile serves as an input to the ROboMObil's geometric path following control, allowing for automated driving. Finally, a test drive shows the results of implementing the presented algorithms on this research vehicle and investigates the energy saving capabilities of the proposed path planning methodology.

1 June 2014
Verified
Conference Paper
Closed access
JBJonathan Brembeck
Jonathan Brembeck
CWChristoph Winter
Christoph Winter
In this paper an energy optimal path planning and velocity profile generation for our highly maneuverable Robotic Electric Vehicle research platform ROboMObil is presented. The ROMO [1] is a development of the German Aerospace Center's Robotics and Mechatronics Center to cope with several research topics, like energy efficient, autonomous or remote controlled driving for future (electro-) mobility applications. The main task of the proposed algorithms is to calculate an energy optimal trajectory in a real-time capable way. It is designed to incorporate data from actual traffic situations (e.g. oncoming traffic) or changed conditions (e.g. snowy conditions). The resulting trajectory is then fed forward to a lower level time independent path following control [2] that calculates the motion demands for our energy optimal control allocation. This in turn distributes the demand to the actuators of the over-actuated vehicle. We show a numerical reliable way to formulate the energy optimal path planning optimization objective, which is able to provide a consistent replanning feature considering the actual vehicle states. Besides this, different types of optimization methods are evaluated for their real-time capabilities. The velocity profile will be calculated afterwards and the generation of the profile is also enabled to handle dynamic replanning. Finally, we show several experimental results, using a virtual road definition and tests on a commercial real-time platform.

8 March 2023 in Applied Sciences
Verified
Journal Article
Open access
PDF file
KAKenan Ahmic
Kenan Ahmic
JUJohannes Ultsch
Johannes Ultsch
Jonathan BrembeckJonathan Brembeck
Jonathan Brembeck
CWChristoph Winter
Christoph Winter
Reinforcement learning-based controllers for safety-critical applications, such as autonomous driving, are typically trained in simulation, where a vehicle model is provided during the learning process. However, an inaccurate parameterization of the vehicle model used for training heavily influences the performance of the reinforcement learning agent during execution. This inaccuracy is either caused by changes due to environmental influences or by falsely estimated vehicle parameters. In this work, we present our approach of combining dynamics randomization with reinforcement learning to overcome this issue for a path-following control task of an autonomous and over-actuated robotic vehicle. We train three independent agents, where each agent experiences randomization for a different vehicle dynamics parameter, i.e., the mass, the yaw inertia, and the road-tire friction. We randomize the parameters uniformly within predefined ranges to enable the agents to learn an equally robust control behavior for all possible parameter values. Finally, in a simulation study, we compare the performance of the agents trained with dynamics randomization to the performance of an agent trained with the nominal parameter values. Simulation results demonstrate that the former agents obtain a higher level of robustness against model uncertainties and varying environmental conditions than the latter agent trained with nominal vehicle parameter values.

1 June 2016
Verified
Conference Paper
Closed access
PRPeter Ritzer
Peter Ritzer
CWChristoph Winter
Christoph Winter
JBJonathan Brembeck
Jonathan Brembeck
This work describes the development and experimental validation of a geometric path following control strategy with demand supervision applied to an over-actuated robotic vehicle, the ROboMObil [1]. The proposed method enables the ROboMObil to automatically follow paths while the driver is free to control the velocity along the path. Beside the longitudinal degree of freedom, two lateral degrees of freedom can be controlled relative to the path. If this demand interface were provided without supervision, the driver may potentially overwrite the path following control in a manner such that the vehicle limits are violated and the vehicle becomes unstable. To avoid such critical situations a demand supervisor is introduced into the path following framework. The work concludes by a simulative demonstration of the supervised control system and an experimental validation of the presented approach implemented in the ROboMObil.