The precise localization of vehicles is an important requirement for autonomous driving or advanced driver assistance systems. Using common GNSS the ego position can be measured but not with the reliability and precision necessary. An alternative approach to achieve precise localization is the usage of visual landmarks observed by a camera mounted in the vehicle. However, this raises the necessity of reliable visual landmarks that are easily recognizable and persistent. We propose a novel SLAM algorithm that focuses on learning and mapping such visual long-term landmarks (LLamas). The algorithm therefore processes stereo image streams from several recording sessions in the same spatial area. The key part within LLama-SLAM is the assessment of the landmarks with quality values that are inferred as viewpoint dependent probabilities from observation statistics. By adding solely landmarks of high quality to the final LLama Map, it can be kept compact while still allowing reliable localization. Due to the long-term evaluation of the GNSS measurement during the sessions, the landmarks can be positioned precisely in a global referenced coordinate system. For a first assessment of the algorithm's capabilities, we present some experimental results from the mapping process combining three sessions recorded over two months on the same route.

Accurate localization is a vital prerequisite for future assistance or autonomous driving functions in intelligent vehicles. To achieve the required localization accuracy and availability, long-term visual SLAM algorithms like LLama-SLAM are a promising option. In such algorithms visual feature tracks, i. e. landmark observations over several consecutive image frames, have to be matched to feature tracks recorded days, weeks or months earlier. This leads to a more challenging matching problem than in short-term visual localization and known descriptor matching methods cannot be applied directly. In this paper, we devise several approaches to compare and match feature tracks and evaluate their performance on a long-term data set. With the proposed descriptor combination and masking ("CoMa") method the best track matching performance is achieved with minor computational cost. This method creates a single combined descriptor for each feature track and furthermore increases the robustness by capturing the appearance variations of this track in a descriptor mask.

This paper proposes an algorithm for camera based roadway mapping in urban areas. With a convolutional neural network the roadway is detected in images taken by a camera mounted in the vehicle. The detected roadway masks from all images of one driving session are combined according to their corresponding GPS position to create a probabilistic grid map of the roadway. Finally, maps from several driving sessions are merged by a feature matching algorithm to compensate for errors in the roadway detection and localization inaccuracies. Hence, this approach utilizes solely low-cost sensors common in usual production vehicles and can generate highly detailed roadway maps from crowdsourced data.

A necessary prerequisite for future driver assistance systems as well as automated driving is a suitable and accurate representation of the environment around the vehicle with a sufficient range. To extend the range of the environment representation, sharing the environment detections of multiple vehicles via vehicle-to-vehicle (V2V) communication is a promising approach. In this paper, we present a method to fuse shared free space detections from multiple vehicles. The detections are represented as Parametric Free Space (PFS) maps, which are especially suitable for real-time radio V2Vtransmission due to their compactness. A graph-based algorithm to fuse PFS maps is proposed that solves possible contradictions between the maps and incorporates the maps' uncertainty attributes. By solely operating on the contour, the fusion can be carried out by a simple path search in a fusion graph that is constructed from the maps. This results in an efficient method that finds the fusion result within few iterations. To account for possible errors in the relative poses between the PFS maps, we furthermore present an adapted Iterative Closest Point (ICP) matching to align the maps before the fusion. Therein we employ a modified soft-assign scheme for robust outlier rejection, and incorporate the PFS maps' boundary orientation to improve the matching process. We show the capabilities of our method by presenting results on real test drive data.

In this paper, we present a method to generate compact geometric track-maps for train-borne localization applications. Therefore, we first give a brief overview on the purpose of track maps in train-positioning applications. It becomes apparent that there are hardly any adequate methods to generate suitable geometric track-maps. This is why we present a novel map generation procedure. It uses an optimization formulation to find the continuous sequence of track geometries that fits the available measurement data best. The optimization is initialized with the results from a localization filter [1] developed in our previous work. The localization filter also provides the required information for shape identification and measurement association. The presented approach will be evaluated on simulated data as well as on real measurements.

The use of machine learning in driver assistance systems allows to significantly enhance their functionalities. In particular, it allows to personalize systems by evaluating the driver’s past behavior. Such personalization is especially relevant for recommendations in maneuvers where the specific maneuver embodiment strongly depends on the driver’s momentary driving style and attention. Led by this idea, PRORETA 4 developed a prototypical City Assistant System, which gives the driver a personalized recommendation in urban scenarios. To adapt the recommendations and warnings appropriately, the system incorporates the learned momentary driving style and the driver’s gaze behavior. In this work, we describe the main functional blocks of the system, present our solutions to major implementation challenges and also discuss the safety of the used learning algorithm.

No abstract