Although accurate and comprehensive maps are indispensable for highly automated driving, especially in complex urban scenarios, there are hardly any publications in which requirements for these maps are discussed. In our opinion, such maps must meet high demands in terms of accuracy, completeness, verifiability and extensibility, so that the resulting complexity can only be handled by an enclosing, carefully designed software framework. In this paper we therefore introduce the open-source map framework Lanelet2 implemented in C++ and explain the underlying concept. The goal of Lanelet2 is not only to be usable for typical, isolated applications such as localization or motion planning, but for various potential applications of maps for highly automated driving. On the basis of both abstract and real examples we show the concrete structure of Lanelet2 maps and its use for automated driving.

Lanelet2: A high-definition map framework for the future of automated driving

This paper reviews the state of the art and future trends of indoor Positioning, Localization, and Navigation (PLAN). It covers the requirements, the main players, sensors, and techniques for indoor PLAN. Other than the navigation sensors such as Inertial Navigation System (INS) and Global Navigation Satellite System (GNSS), the environmental-perception sensors such as High-Definition map (HD map), Light Detection and Ranging (LiDAR), camera, the fifth generation of mobile network communication technology (5G), and Internet-of-Things (IoT) signals are becoming important aiding sensors for PLAN. The PLAN systems are expected to be more intelligent and robust under the emergence of more advanced sensors, multi-platform/multi-device/multi-sensor information fusion, self-learning systems, and the integration with artificial intelligence, 5G, IoT, and edge/fog computing.

/pdf/indoor-navigation-state-of-the-art-and-future-trends-2ozl1wnwm5.pdf

Indoor navigation: state of the art and future trends

Reliable and accurate 3D object detection is a necessity for safe autonomous driving. Although LiDAR sensors can provide accurate 3D point cloud estimates of the environment, they are also prohibitively expensive for many settings. Recently, the introduction of pseudo-LiDAR (PL) has led to a drastic reduction in the accuracy gap between methods based on LiDAR sensors and those based on cheap stereo cameras. PL combines state-of-the-art deep neural networks for 3D depth estimation with those for 3D object detection by converting 2D depth map outputs to 3D point cloud inputs. However, so far these two networks have to be trained separately. In this paper, we introduce a new framework based on differentiable Change of Representation (CoR) modules that allow the entire PL pipeline to be trained end-to-end. The resulting framework is compatible with most state-of-the-art networks for both tasks and in combination with PointRCNN improves over PL consistently across all benchmarks --- yielding the highest entry on the KITTI image-based 3D object detection leaderboard at the time of submission. Our code will be made available at https://github.com/mileyan/pseudo-LiDAR_e2e.

/pdf/end-to-end-pseudo-lidar-for-image-based-3d-object-detection-5lmhbt7d5u.pdf

End-to-End Pseudo-LiDAR for Image-Based 3D Object Detection

Reliable and accurate 3D object detection is a necessity for safe autonomous driving. Although LiDAR sensors can provide accurate 3D point cloud estimates of the environment, they are also prohibitively expensive for many settings. Recently, the introduction of pseudo-LiDAR (PL) has led to a drastic reduction in the accuracy gap between methods based on LiDAR sensors and those based on cheap stereo cameras. PL combines state-of-the-art deep neural networks for 3D depth estimation with those for 3D object detection by converting 2D depth map outputs to 3D point cloud inputs. However, so far these two networks have to be trained separately. In this paper, we introduce a new framework based on differentiable Change of Representation (CoR) modules that allow the entire PL pipeline to be trained end-to-end. The resulting framework is compatible with most state-of-the-art networks for both tasks and in combination with PointRCNN improves over PL consistently across all benchmarks -- yielding the highest entry on the KITTI image-based 3D object detection leaderboard at the time of submission. Our code will be made available at this https URL.

In the domain of autonomous driving, deep learning has substantially improved the 3D object detection accuracy for LiDAR and stereo camera data alike. While deep networks are great at generalization, they are also notorious to over-fit to all kinds of spurious artifacts, such as brightness, car sizes and models, that may appear consistently throughout the data. In fact, most datasets for autonomous driving are collected within a narrow subset of cities within one country, typically under similar weather conditions. In this paper we consider the task of adapting 3D object detectors from one dataset to another. We observe that naively, this appears to be a very challenging task, resulting in drastic drops in accuracy levels. We provide extensive experiments to investigate the true adaptation challenges and arrive at a surprising conclusion: the primary adaptation hurdle to overcome are differences in car sizes across geographic areas. A simple correction based on the average car size yields a strong correction of the adaptation gap. Our proposed method is simple and easily incorporated into most 3D object detection frameworks. It provides a first baseline for 3D object detection adaptation across countries, and gives hope that the underlying problem may be more within grasp than one may have hoped to believe. Our code is available at this https URL.

Train in Germany, Test in The USA: Making 3D Object Detectors Generalize

We propose a 3D object detection and pose estimation method for automated driving using stereo images. In contrast to existing stereo-based approaches, we focus not only on cars, but on all types of road users and can ensure real-time capability through GPU implementation of the entire processing chain. These are essential conditions to exploit an algorithm for highly automated driving. Semantic information is provided by a deep convolutional neural network and used together with disparity and geometric constraints to recover accurate 3D bounding boxes. Experiments on the challenging KITTI 3D object detection benchmark show results that are within the range of the best image-based algorithms, while the runtime is only about a fifth. This makes our algorithm the first real-time image-based approach on KITTI.

Realtime 3D Object Detection for Automated Driving Using Stereo Vision and Semantic Information

This paper presents the concepts and methods utilized by Team AnnieWAY for the 2016 Grand Cooperative Driving Challenge. The paper introduces the automated vehicle BerthaOne . The vehicle, even though being based on the Bertha platform, distinguishes itself from its siblings by its software modules and algorithms. We, therefore, describe its system architecture and algorithms for perception, cooperation and motion planning. In Particular, we present a motion planner that plans different maneuvers flexibly by augmenting the cost function with situation specific cost terms. We subsequently describe the requirements of the 2016 GCDC and evaluate our performance during the competition.

Making Bertha Cooperate–Team AnnieWAY’s Entry to the 2016 Grand Cooperative Driving Challenge

The future of automated driving in urban areas will most probably rely on highly accurate road maps. However, the necessary precision of a localization in such maps has so far only been reached using extra, sensor specific feature layers for localization. In this paper we want to show that it is possible to achieve sufficient accuracy without a separate localization layer. Instead, elements are used that are already contained in high-resolution road maps, such as markings and road borders. For this, we introduce a modular approach in which detections from different detection algorithms are associated with elements in the map and then fused to an absolute pose using an Unscented Kalman Filter. We evaluate our approach using a sensor setup that employs a stereo camera, vehicle odometry and a low-cost GNSS module on a 5km test route covering both narrow urban roads and multi-lane main roads under varying weather conditions. The results show that this approach is capable to be used for highly automated driving, showing an accuracy of 0.08m in typical road scenarios and a is available 98% of the time.

Precise Localization in High-Definition Road Maps for Urban Regions

Automated vehicles are complex systems with a high degree of interdependencies between its components. This complexity sets increasing demands for the underlying software framework. This paper firstly analyzes the requirements for software frameworks. Afterwards an overview on existing software frameworks, that have been used for automated driving projects, is provided with an in-depth introduction into an emerging open-source software framework, the Robot Operating System (ROS). After discussing the main features, advantages and disadvantages of ROS, the communication overhead of ROS is analyzed quantitatively in various configurations showing its applicability for systems with a high data load.

Robot operating system: A modular software framework for automated driving

Most state of the art object detectors output multiple detections per object. The duplicates are removed in a post-processing step called Non-Maximum Suppression. Classical Non-Maximum Suppression has shortcomings in scenes that contain objects with high overlap: This heuristic assumes that a high overlap between two bounding boxes corresponds to a high probability of one being a duplicate. We propose FeatureNMS to solve this problem. FeatureNMS recognizes duplicates not only based on the intersection over union between the bounding boxes, but also based on the difference of feature vectors. These feature vectors can encode more information like visual appearance. Our approach outperforms classical NMS and derived approaches and achieves state of the art performance.

Niels Ole Salscheider

Papers

Realtime 3D Object Detection for Automated Driving Using Stereo Vision and Semantic Information

Making Bertha Cooperate–Team AnnieWAY’s Entry to the 2016 Grand Cooperative Driving Challenge

Precise Localization in High-Definition Road Maps for Urban Regions

Robot operating system: A modular software framework for automated driving

FeatureNMS: Non-Maximum Suppression by Learning Feature Embeddings