Situation understanding is a key factor for future advanced driver assistance systems (ADAS) to handle complex traffic situations. An important part is the prediction of future actions of other traffic participants. Tackling this problem the paper at hand focuses on the prediction of lane change maneuvers on highways. A novel approach is presented that first separates the prediction into a situation based and a movement based approach and fuses them afterwards. A situation based probability resting on inter-vehicle relations is derived, which enables very early reasoning about the current traffic situation and gives prior knowledge about possible lane change maneuvers. Then observations of the vehicles lateral movement inside the lane are used to perform a probabilistic multi-class classification with a Support Vector Machine (SVM). Both probabilities are combined to enhance the movement based result using the situation related probability as prior knowledge from the current driving situation. The approach is tested on a dataset recorded on a fixed-base driving simulator. Considering only the situation based information early prediction of a feasible lane change is possible. Furthermore the evaluation affirms the improvement of the prediction in case the approach is extended to incorporate both probabilities. Finally the combined approach is tested on a dataset recorded in highway traffic scenarios.

Due to a large number of integrated advanced driver assistance systems (ADAS) the driver nowadays can hand over the driving task to the vehicle in specific, monotone driving scenarios. Short reaction times and the constant awareness of the computer reduces the number of accidents and thus increases safety. Currently available ADAS still need to be constantly monitored by the driver in case a situation appears that cannot be handled properly by the system.

In order to develop automated driving systems an approach to generate trajectories is of essential importance. The trajectory has to satisfy two fundamental requirements. On the one hand collision free driving of the vehicle has to be ensured and on the other hand the limited driving dynamics has to be complied. The real-time capable trajectory optimization including a complex vehicle dynamics model causes high computational costs. In this contribution the trajectory optimization is based on the kinodynamics of a mass point. The road limits and obstacles are included in the optimization problem by a potential field.

The paper at hand proposes an efficient trajectory planning approach for automated vehicles. The concept of potential field based online trajectory optimization is enhanced by a spline-based interpolation strategy, valid for normal driving functions. The resulting benefits of the developed Timed Elastic Spline (TES) approach concern improvements in computational efficiency and faster convergence and thus effect the computation time. An optimization algorithm is applied to generate the optimal trajectory considering the objectives of collision avoidance and comfort. The results show the performance of the developed algorithm, which is designed to solve a broad range of traffic scenarios. Additionally measurements indicate that the algorithm is suitable for real time application.

A key aspect for the development of automated driving functions is the planning of a predictive trajectory in dependence of the current traffic situation. The complexity of the utilized trajected planning approach is bounded by the available computation time that has to be in accordance with the dynamics of the surrounding environment. This work is concerned with an approach to trajectory planning for normal driving situations, which is based on interpolation of trajectory points between optimized states. This measure has the objective to improve the performance particularly with respect to the computational burden. By means of a simulation environment, different interpolation strategies are investigated and a maneuver-based comparison with respect to computation time and optimality in terms of a defined cost function is performed.

The paper at hand proposes a real-time capable approach to combined trajectory planning and control. One single prediction model is used to plan a feasible trajectory and to perform lateral guidance of the vehicle at the same time. Nonlinear model predictive control (NMPC) methods are applied to solve the optimal control problem, which incorporates environmental constraints leading to a model predictive planning and control approach (MPPC). Experiments are conducted utilizing a rapid prototyping system. The analysis shows the versatile application range of the developed algorithm, such as challenging emergency evasive maneuvers as well as automated steering as an approach to lateral guidance of the vehicle in general.

A novel approach for combined trajectory planning and control is presented in this contribution. The developed method integrates optimal control theory and trajectory planning and leads to an active safety system, which is applicable for automated driving and can avoid collisions in critical situations. The Combined Planning and Control (CPC) algorithm extends the Timed Elastic Band (TEB) approach by a suitable vehicle dynamics model, which facilitates stable vehicle guidance. The problem of trajectory generation is reformulated such that a nonlinear model predictive control (NMPC) method can be applied. The analysis of different traffic scenarios shows the functionality of the developed concept.

Many accidents lead back to the driver. It is assumed to be beneficial that driver assistance systems and/or automated driving functions ease tension off the driver or release him from the task of driving in order to reduce traffic fatalities. The development of driver assistance systems started with ABS (Anti-Lock Braking System) in 1978, TC (Traction Control) in 1987 and ESC (Electronic Stability Control) in 1995. These systems are known for their contribution on the reduction of traffic fatalities over the last years. But still far too many cars crash every day. The aforementioned systems do not take environmental information into account, which limits the collision avoidance potential. A system that senses its environment by radar and/or camera sensors is the AEB (Autonomous Emergency Braking) which is supposed to further reduce the amount of accidents in the future as the take rate increases. Although assisting the driver in emergency situations by braking is beneficial in some situations, it also has its drawbacks. The problem is that emergency braking can only be initiated if the obstacle is detected for sure. Nowadays sensors can only provide limited safe ranges. This means that at typical highway speeds the collision with a stationary obstacle can only be mitigated by braking. A swerving maneuver is a viable alternative to braking or the combination of both if the required space is not occupied by other obstacles. The emergency steering maneuver offers the advantage that the last-point-to-steer is even closer to the obstacle than the last-point-to-brake with growing vehicle speed, which yields a higher collision avoidance potential. The disadvantage is that the maneuver is far more difficult than full braking. A subject study [1] in a driving simulator has proven that the average skilled driver is not capable of steering the vehicle properly around obstacles in the majority of occasions. This is why an advanced driver assistance system (ADAS) is useful that supports the driver by steering torque overlay and, as an option, by braking interventions. The improvement on the collision avoidance behaviour of the driver has been proven by the authors in the aforementioned study [1].

Advanced Driver Assistance Systems (ADAS) that support drivers during emergency maneuvers have been considered in research and development projects before. It is commonplace that the two problems trajectory planning and vehicle control are solved in two different steps. The paper at hand proposes a real time capable method that solves the planning and control problem in one step. This is achieved by using a scheme similar to model predictive control combined with an environmental model to avoid the necessity of a precalculated reference signal, i.e. a trajectory. The real time capability is achieved by a rough discretization of the inputs, which yields the possibility to predict the plant's state trajectory for all possible combinations of discretized input values. Compared to an iterative minimization the computational burden is lower and can be determined deterministically. This contribution contains collision avoidance maneuvers in different traffic situations evaluated in simulation and a fixed base driving simulator.

This paper is concerned with the planning of optimal trajectories for vehicle collision avoidance with a Timed Elastic Band (TEB) framework. The avoidance trajectory is represented by a TEB which is optimized with respect to multiple partially conflicting objectives. The resulting trajectory constitutes the optimal compromise between a mere braking and a lane change maneuver that avoids the collision with the smoothest feasible path. The approach is applicable to general critical traffic situations as the TEB considers the constraints imposed by the vehicle dynamics, road boundaries, static obstacles and moving vehicles.

This contribution is concerned with an emergency steering assist and its evaluation through subject testing in a driving simulator. In an emergency traffic situation where a rear end collision is imminent a swerving maneuver is often too difficult for most drivers. Therefore an assistance system is usefull that supports the driver by steering torque overlay. The paper presents the algorithm used for the assistance. The system was prototypically implemented in a driving simulator and testet with subjects to evaluate the benefits and challenges. The results show that the collision avoidance behaviour of the driver can be improved by the emergency steering assist.

A control systems concept for an emergency evading assistance system is proposed in this paper. The control systems concept consists of two parts. The first one calculates steering angles for the emergency maneuver while the second part tries to teach the driver these steering angles. The first part will be investigated in more detail while for the second one a cascaded controller is used. The control system from [1] is extended to different friction coefficients and velocities. In common the friction coefficient is not known. Therefore a simple method is proposed that provides the friction coefficient approximately from the first part of the maneuver.

Aktive und passive Sicherheit moderner Automobile befinden sich bereits auf einem sehr hohen Niveau. Dennoch sind auch heute noch bei den bewaehrten und eingefuehrten Systemen, wie etwa dem Gurtstraffer, noch weitere Verbesserungen moeglich. Zukuenftige Sicherheitskonzepte sehen einen flexiblen, der Situation und den Fahrzeuginsassen angepassten und praeventiven Schutz vor. Im Beispiel wird dazu auf einen reversiblen mechatronischen Gurtstraffer eingegangen. Dazu wird auch ein Simulationsmodell mit seinen Teilmodellen Insassenmodell, Sensormodell und Aktormodell vorgestellt. Zur Darstellung des Potenzials eines reversiblen mechatronischen Gurtstraffers werden zwei Simulationen vorgestellt: eine Einfahrt in eine langgezogene Linkskurve (Gurtstraffung) mit anschliessender Geradeausfahrt (Gurtloesung) und eine Kurvenfahrt mit Schleudervorgang (Gurtstraffung) und anschliessendem Abkommen von der Strasse mit Fahrzeugueberschlag.