To guide an automated vehicle safely through complex traffic, knowledge about the future evolution of the driving situation has to be considered. The contribution at hand proposes an approach for automated driving in structured environments. An environment representation for trajectory planning is presented that enables predictive driving by interconnecting trajectory prediction for the surrounding traffic and planning of an according ego trajectory. A maneuver-based approach with an efficient trajectory model is carried out to enable an accurate and fast estimation of future motions of other vehicles. The results are directly considered in the dynamic environment representation utilized for trajectory planning. In the trajectory optimization process, the required safety distance is hence taken into account uniquely for each future time instance. The evaluation of the trajectory prediction approach shows very good performance on a simulated as well as on a dataset recorded by a test vehicle. On account of the predictive character of the developed environment potential field, simulation experiments demonstrate the feasibility and effectiveness of the proposed method.

The contribution at hand combines a sampling-based trajectory planning approach and a model predictive trajectory tracking controller to a collision avoidance system. The planner generates candidate trajectories by the suitable selection of breakpoints which are connected by a spline interpolation. A procedure is presented to systematically select sample states to perform a collision avoidance maneuver in case of an emergency situation. The vehicle is controlled to the optimal trajectory of the planner by comparison with a model predictive trajectory set. This is determined by the prediction of a detailed nonlinear vehicle model for constant input variables. A suitable objective function with a time weighting is utilized to evaluate the individual trajectories and to select the optimal input variables.

Zusammenfassung Der vorliegende Beitrag untersucht die Erkennung benachbarter Fahrstreifen auf Grundlage von Kamerabildern. Hierbei wird sowohl die Anzahl befahrbarer Fahrstreifen als auch deren Verlauf innerhalb eines festgelegten Bereichs vor dem Fahrzeug bildbasiert geschätzt. Die Erkennung erfolgt durch Convolutional Neural Networks. Der Beitrag bewertet die Güte der Schätzung und Genauigkeit der Fahrstreifenerkennung. Zur Interpretation der Entscheidungen werden die durch das neuronale Netz gelernten Merkmale und Zwischenrepräsentationen visualisiert.

Automated driving is a key technology for the future of transportation. There are several motivations to develop automated vehicles. First and foremost, it promises to reduce the number of traffic accidents. Figure 1 shows the accidents recorded by the German police over the past years ([1]) ranging back to 1960.

The paper at hand proposes an environment model for trajectory planning in structured environments. It is composed of a static and a dynamic environment model. The generated static potential field takes restrictions imposed by the static environment into account. The dynamic environment model is based on the physical interpretation of the required safety distance. By the use of an advanced obstacle trajectory prediction method, the safety distance is calculated in accordance to the predicted situation. As the safety distance affects the dynamic potential field, information provided by the obstacle trajectory prediction is directly considered in the ego vehicle trajectory planning process. On account of the predictive character of the developed environment potential field, simulation experiments demonstrate the feasibility and effectiveness of the proposed method.

The paper at hand proposes a real-time capable approach to trajectory planning. An online sampling strategy is chosen, exploiting the structure of the surrounding environment. Lateral states are sampled from state space, whereas longitudinal states are generated via sampling from the action space. The combination yields breakpoints, which are then used to generate a candidate trajectory via spline interpolation. A bi-level candidate evaluation strategy is presented assessing comfort and human-like driving as well as a post-check of collision avoidance with accurate geometric modeling. The result is a reactive feedback motion planner, which shows promising results with respect to on-road automated driving.