No abstract

No abstract

Behavior and motion planning play an important role in automated driving. Traditionally, behavior planners instruct local motion planners with predefined behaviors. Due to the high scene complexity in urban environments, unpredictable situations may occur in which behavior planners fail to match predefined behavior templates. Recently, generalpurpose planners have been introduced, combining behavior and local motion planning. These general-purpose planners allow behavior-aware motion planning given a single reward function. However, two challenges arise: First, this function has to map a complex feature space into rewards. Second, the reward function has to be manually tuned by an expert. Manually tuning this reward function becomes a tedious task. In this paper, we propose an approach that relies on human driving demonstrations to automatically tune reward functions. This study offers important insights into the driving style optimization of general-purpose planners with maximum entropy inverse reinforcement learning. We evaluate our approach based on the expected value difference between learned and demonstrated policies. Furthermore, we compare the similarity of human driven trajectories with optimal policies of our planner under learned and expert-tuned reward functions. Our experiments show that we are able to learn reward functions exceeding the level of manual expert tuning without prior domain knowledge.

The use of Floating Car Data (FCD) as a particular case of Probe Vehicle Data (PVD) has been the object of extensive research for estimating traffic conditions, travel times and Origin to Destination trip matrices. It is based on data collected from a GPS-equipped vehicle fleet or available cell phones. Cooperative Cars with vehicle-to-vehicle (V2V) and vehicle-to-infrastructure (V2I) communication capabilities represent a step forward, as they also allow tracking vehicles surrounding the equipped car. This paper presents the results of a limited experiment with a small fleet of cooperative cars in Barcelona's Central Business District (CBD) known as L'Eixample. Data collected from the experiment were used to build and calibrate the emulation of cooperative functions in a microscopic simulation model that captured the behavior of vehicle sensors in Barcelona's CBD. Such a calibrated model allows emulating fleet data on a large scale that goes far beyond what a small fleet of cooperative vehicles could capture. To determine the traffic state, several approaches are developed for estimating traffic variables based on extensions of Edie's definition of the fundamental traffic variables with the emulated data, whose accuracy depends on the penetration level of the technology.

No abstract