Claudius Gläser

Felicia Ruppel, Florian Faion, Claudius Gläser et al.

We present TransMOT, a novel transformer-based end-to-end trainable online tracker and detector for point cloud data. The model utilizes a cross- and a self-attention mechanism and is applicable to lidar data in an automotive context, as well as other data types, such as radar. Both track management and the detection of new tracks are performed by the same transformer decoder module and the tracker state is encoded in feature space. With this approach, we make use of the rich latent space of the detector for tracking rather than relying on low-dimensional bounding boxes. Still, we are able to retain some of the desirable properties of traditional Kalman-filter based approaches, such as an ability to handle sensor input at arbitrary timesteps or to compensate frame skips. This is possible due to a novel module that transforms the track information from one frame to the next on feature-level and thereby fulfills a similar task as the prediction step of a Kalman filter. Results are presented on the challenging real-world dataset nuScenes, where the proposed model outperforms its Kalman filter-based tracking baseline.

Michael Ulrich, Sascha Braun, Daniel Köhler et al.

This paper presents novel hybrid architectures that combine grid- and point-based processing to improve the detection performance and orientation estimation of radar-based object detection networks. Purely grid-based detection models operate on a bird's-eye-view (BEV) projection of the input point cloud. These approaches suffer from a loss of detailed information through the discrete grid resolution. This applies in particular to radar object detection, where relatively coarse grid resolutions are commonly used to account for the sparsity of radar point clouds. In contrast, point-based models are not affected by this problem as they process point clouds without discretization. However, they generally exhibit worse detection performances than grid-based methods. We show that a point-based model can extract neighborhood features, leveraging the exact relative positions of points, before grid rendering. This has significant benefits for a subsequent grid-based convolutional detection backbone. In experiments on the public nuScenes dataset our hybrid architecture achieves improvements in terms of detection performance (19.7% higher mAP for car class than next-best radar-only submission) and orientation estimates (11.5% relative orientation improvement) over networks from previous literature.

Felicia Ruppel, Florian Faion, Claudius Gläser et al.

Group regression is commonly used in 3D object detection to predict box parameters of similar classes in a joint head, aiming to benefit from similarities while separating highly dissimilar classes. For query-based perception methods, this has, so far, not been feasible. We close this gap and present a method to incorporate multi-class group regression, especially designed for the 3D domain in the context of autonomous driving, into existing attention and query-based perception approaches. We enhance a transformer based joint object detection and tracking model with this approach, and thoroughly evaluate its behavior and performance. For group regression, the classes of the nuScenes dataset are divided into six groups of similar shape and prevalence, each being regressed by a dedicated head. We show that the proposed method is applicable to many existing transformer based perception approaches and can bring potential benefits. The behavior of query group regression is thoroughly analyzed in comparison to a unified regression head, e.g. in terms of class-switching behavior and distribution of the output parameters. The proposed method offers many possibilities for further research, such as in the direction of deep multi-hypotheses tracking.

Florian Drews, Di Feng, Florian Faion et al.

We propose DeepFusion, a modular multi-modal architecture to fuse lidars, cameras and radars in different combinations for 3D object detection. Specialized feature extractors take advantage of each modality and can be exchanged easily, making the approach simple and flexible. Extracted features are transformed into bird's-eye-view as a common representation for fusion. Spatial and semantic alignment is performed prior to fusing modalities in the feature space. Finally, a detection head exploits rich multi-modal features for improved 3D detection performance. Experimental results for lidar-camera, lidar-camera-radar and camera-radar fusion show the flexibility and effectiveness of our fusion approach. In the process, we study the largely unexplored task of faraway car detection up to 225 meters, showing the benefits of our lidar-camera fusion. Furthermore, we investigate the required density of lidar points for 3D object detection and illustrate implications at the example of robustness against adverse weather conditions. Moreover, ablation studies on our camera-radar fusion highlight the importance of accurate depth estimation.

Daniel Niederlöhner, Michael Ulrich, Sascha Braun et al.

This paper presents a method to learn the Cartesian velocity of objects using an object detection network on automotive radar data. The proposed method is self-supervised in terms of generating its own training signal for the velocities. Labels are only required for single-frame, oriented bounding boxes (OBBs). Labels for the Cartesian velocities or contiguous sequences, which are expensive to obtain, are not required. The general idea is to pre-train an object detection network without velocities using single-frame OBB labels, and then exploit the network's OBB predictions on unlabelled data for velocity training. In detail, the network's OBB predictions of the unlabelled frames are updated to the timestamp of a labelled frame using the predicted velocities and the distances between the updated OBBs of the unlabelled frame and the OBB predictions of the labelled frame are used to generate a self-supervised training signal for the velocities. The detection network architecture is extended by a module to account for the temporal relation of multiple scans and a module to represent the radars' radial velocity measurements explicitly. A two-step approach of first training only OBB detection, followed by training OBB detection and velocities is used. Further, a pre-training with pseudo-labels generated from radar radial velocity measurements bootstraps the self-supervised method of this paper. Experiments on the publicly available nuScenes dataset show that the proposed method almost reaches the velocity estimation performance of a fully supervised training, but does not require expensive velocity labels. Furthermore, we outperform a baseline method which uses only radial velocity measurements as labels.

Felicia Ruppel, Florian Faion, Claudius Gläser et al.

Transformers have recently been utilized to perform object detection and tracking in the context of autonomous driving. One unique characteristic of these models is that attention weights are computed in each forward pass, giving insights into the model's interior, in particular, which part of the input data it deemed interesting for the given task. Such an attention matrix with the input grid is available for each detected (or tracked) object in every transformer decoder layer. In this work, we investigate the distribution of these attention weights: How do they change through the decoder layers and through the lifetime of a track? Can they be used to infer additional information about an object, such as a detection uncertainty? Especially in unstructured environments, or environments that were not common during training, a reliable measure of detection uncertainty is crucial to decide whether the system can still be trusted or not.

Felicia Ruppel, Florian Faion, Claudius Gläser et al.

We present TransLPC, a novel detection model for large point clouds that is based on a transformer architecture. While object detection with transformers has been an active field of research, it has proved difficult to apply such models to point clouds that span a large area, e.g. those that are common in autonomous driving, with lidar or radar data. TransLPC is able to remedy these issues: The structure of the transformer model is modified to allow for larger input sequence lengths, which are sufficient for large point clouds. Besides this, we propose a novel query refinement technique to improve detection accuracy, while retaining a memory-friendly number of transformer decoder queries. The queries are repositioned between layers, moving them closer to the bounding box they are estimating, in an efficient manner. This simple technique has a significant effect on detection accuracy, which is evaluated on the challenging nuScenes dataset on real-world lidar data. Besides this, the proposed method is compatible with existing transformer-based solutions that require object detection, e.g. for joint multi-object tracking and detection, and enables them to be used in conjunction with large point clouds.

Joachim Börger, Marc Patrick Zapf, Marat Kopytjuk et al.

Sensor visibility is crucial for safety-critical applications in automotive, robotics, smart infrastructure and others: In addition to object detection and occupancy mapping, visibility describes where a sensor can potentially measure or is blind. This knowledge can enhance functional safety and perception algorithms or optimize sensor topologies. Despite its significance, to the best of our knowledge, neither a common definition of visibility nor performance metrics exist yet. We close this gap and provide a definition of visibility, derived from a use case review. We introduce metrics and a framework to assess the performance of visibility estimators. Our metrics are verified with labeled real-world and simulation data from infrastructure radars and cameras: The framework easily identifies false visible or false invisible estimations which are safety-critical. Applying our metrics, we enhance the radar and camera visibility estimators by modeling the 3D elevation of sensor and objects. This refinement outperforms the conventional planar 2D approach in trustfulness and thus safety.

Michael Kösel, Marcel Schreiber, Michael Ulrich et al.

LiDAR-based 3D object detection has become an essential part of automated driving due to its ability to localize and classify objects precisely in 3D. However, object detectors face a critical challenge when dealing with unknown foreground objects, particularly those that were not present in their original training data. These out-of-distribution (OOD) objects can lead to misclassifications, posing a significant risk to the safety and reliability of automated vehicles. Currently, LiDAR-based OOD object detection has not been well studied. We address this problem by generating synthetic training data for OOD objects by perturbing known object categories. Our idea is that these synthetic OOD objects produce different responses in the feature map of an object detector compared to in-distribution (ID) objects. We then extract features using a pre-trained and fixed object detector and train a simple multilayer perceptron (MLP) to classify each detection as either ID or OOD. In addition, we propose a new evaluation protocol that allows the use of existing datasets without modifying the point cloud, ensuring a more authentic evaluation of real-world scenarios. The effectiveness of our method is validated through experiments on the newly proposed nuScenes OOD benchmark. The source code is available at https://github.com/uulm-mrm/mmood3d.

Michael Kösel, Marcel Schreiber, Michael Ulrich et al.

LiDAR-based 3D object detection plays a critical role for reliable and safe autonomous driving systems. However, existing detectors often produce overly confident predictions for objects not belonging to known categories, posing significant safety risks. This is caused by so-called out-of-distribution (OOD) objects, which were not part of the training data, resulting in incorrect predictions. To address this challenge, we propose ALOOD (Aligned LiDAR representations for Out-Of-Distribution Detection), a novel approach that incorporates language representations from a vision-language model (VLM). By aligning the object features from the object detector to the feature space of the VLM, we can treat the detection of OOD objects as a zero-shot classification task. We demonstrate competitive performance on the nuScenes OOD benchmark, establishing a novel approach to OOD object detection in LiDAR using language representations. The source code is available at https://github.com/uulm-mrm/mmood3d.

Markus Käppeler, Özgün Çiçek, Daniele Cattaneo et al.

Camera-based 3D object detection and tracking are essential for perception in autonomous driving. Current state-of-the-art approaches often rely exclusively on either perspective-view (PV) or bird's-eye-view (BEV) features, limiting their ability to leverage both fine-grained object details and spatially structured scene representations. In this work, we propose DualViewDistill, a hybrid detection and tracking framework that incorporates both PV and BEV camera image features to leverage their complementary strengths. Our approach introduces BEV maps guided by foundation models, leveraging descriptive DINOv2 features that are distilled into BEV representations through a novel distillation process. By integrating PV features with BEV maps enriched with semantic and geometric features from DINOv2, our model leverages this hybrid representation via deformable aggregation to enhance 3D object detection and tracking. Extensive experiments on the nuScenes and Argoverse 2 benchmarks demonstrate that DualViewDistill achieves state-of-the-art performance. The results showcase the potential of foundation model BEV maps to enable more reliable perception for autonomous driving. We make the code and pre-trained models available at https://dualviewdistill.cs.uni-freiburg.de .

Rohit Mohan, Julia Hindel, Florian Drews et al.

Autonomous vehicles that navigate in open-world environments may encounter previously unseen object classes. However, most existing LiDAR panoptic segmentation models rely on closed-set assumptions, failing to detect unknown object instances. In this work, we propose ULOPS, an uncertainty-guided open-set panoptic segmentation framework that leverages Dirichlet-based evidential learning to model predictive uncertainty. Our architecture incorporates separate decoders for semantic segmentation with uncertainty estimation, embedding with prototype association, and instance center prediction. During inference, we leverage uncertainty estimates to identify and segment unknown instances. To strengthen the model's ability to differentiate between known and unknown objects, we introduce three uncertainty-driven loss functions. Uniform Evidence Loss to encourage high uncertainty in unknown regions. Adaptive Uncertainty Separation Loss ensures a consistent difference in uncertainty estimates between known and unknown objects at a global scale. Contrastive Uncertainty Loss refines this separation at the fine-grained level. To evaluate open-set performance, we extend benchmark settings on KITTI-360 and introduce a new open-set evaluation for nuScenes. Extensive experiments demonstrate that ULOPS consistently outperforms existing open-set LiDAR panoptic segmentation methods.

Mohamed Abdelsamad, Michael Ulrich, Claudius Gläser et al.

Masked autoencoders (MAE) have shown tremendous potential for self-supervised learning (SSL) in vision and beyond. However, point clouds from LiDARs used in automated driving are particularly challenging for MAEs since large areas of the 3D volume are empty. Consequently, existing work suffers from leaking occupancy information into the decoder and has significant computational complexity, thereby limiting the SSL pre-training to only 2D bird's eye view encoders in practice. In this work, we propose the novel neighborhood occupancy MAE (NOMAE) that overcomes the aforementioned challenges by employing masked occupancy reconstruction only in the neighborhood of non-masked voxels. We incorporate voxel masking and occupancy reconstruction at multiple scales with our proposed hierarchical mask generation technique to capture features of objects of different sizes in the point cloud. NOMAEs are extremely flexible and can be directly employed for SSL in existing 3D architectures. We perform extensive evaluations on the nuScenes and Waymo Open datasets for the downstream perception tasks of semantic segmentation and 3D object detection, comparing with both discriminative and generative SSL methods. The results demonstrate that NOMAE sets the new state-of-the-art on multiple benchmarks for multiple point cloud perception tasks.

Karim Armanious, Maurice Quach, Michael Ulrich et al.

This paper introduces the Bosch street dataset (BSD), a novel multi-modal large-scale dataset aimed at promoting highly automated driving (HAD) and advanced driver-assistance systems (ADAS) research. Unlike existing datasets, BSD offers a unique integration of high-resolution imaging radar, lidar, and camera sensors, providing unprecedented 360-degree coverage to bridge the current gap in high-resolution radar data availability. Spanning urban, rural, and highway environments, BSD enables detailed exploration into radar-based object detection and sensor fusion techniques. The dataset is aimed at facilitating academic and research collaborations between Bosch and current and future partners. This aims to foster joint efforts in developing cutting-edge HAD and ADAS technologies. The paper describes the dataset's key attributes, including its scalability, radar resolution, and labeling methodology. Key offerings also include initial benchmarks for sensor modalities and a development kit tailored for extensive data analysis and performance evaluation, underscoring our commitment to contributing valuable resources to the HAD and ADAS research community.

Marcel Schreiber, Vasileios Belagiannis, Claudius Gläser et al.

A common approach for modeling the environment of an autonomous vehicle are dynamic occupancy grid maps, in which the surrounding is divided into cells, each containing the occupancy and velocity state of its location. Despite the advantage of modeling arbitrary shaped objects, the used algorithms rely on hand-designed inverse sensor models and semantic information is missing. Therefore, we introduce a multi-task recurrent neural network to predict grid maps providing occupancies, velocity estimates, semantic information and the driveable area. During training, our network architecture, which is a combination of convolutional and recurrent layers, processes sequences of raw lidar data, that is represented as bird's eye view images with several height channels. The multi-task network is trained in an end-to-end fashion to predict occupancy grid maps without the usual preprocessing steps consisting of removing ground points and applying an inverse sensor model. In our evaluations, we show that our learned inverse sensor model is able to overcome some limitations of a geometric inverse sensor model in terms of representing object shapes and modeling freespace. Moreover, we report a better runtime performance and more accurate semantic predictions for our end-to-end approach, compared to our network relying on measurement grid maps as input data.

Michael Herman, Jörg Wagner, Vishnu Prabhakaran et al.

Automated vehicles require a comprehensive understanding of traffic situations to ensure safe and anticipatory driving. In this context, the prediction of pedestrians is particularly challenging as pedestrian behavior can be influenced by multiple factors. In this paper, we thoroughly analyze the requirements on pedestrian behavior prediction for automated driving via a system-level approach. To this end we investigate real-world pedestrian-vehicle interactions with human drivers. Based on human driving behavior we then derive appropriate reaction patterns of an automated vehicle and determine requirements for the prediction of pedestrians. This includes a novel metric tailored to measure prediction performance from a system-level perspective. The proposed metric is evaluated on a large-scale dataset comprising thousands of real-world pedestrian-vehicle interactions. We furthermore conduct an ablation study to evaluate the importance of different contextual cues and compare these results to ones obtained using established performance metrics for pedestrian prediction. Our results highlight the importance of a system-level approach to pedestrian behavior prediction.

Marcel Schreiber, Vasileios Belagiannis, Claudius Gläser et al.

Modeling and understanding the environment is an essential task for autonomous driving. In addition to the detection of objects, in complex traffic scenarios the motion of other road participants is of special interest. Therefore, we propose to use a recurrent neural network to predict a dynamic occupancy grid map, which divides the vehicle surrounding in cells, each containing the occupancy probability and a velocity estimate. During training, our network is fed with sequences of measurement grid maps, which encode the lidar measurements of a single time step. Due to the combination of convolutional and recurrent layers, our approach is capable to use spatial and temporal information for the robust detection of static and dynamic environment. In order to apply our approach with measurements from a moving ego-vehicle, we propose a method for ego-motion compensation that is applicable in neural network architectures with recurrent layers working on different resolutions. In our evaluations, we compare our approach with a state-of-the-art particle-based algorithm on a large publicly available dataset to demonstrate the improved accuracy of velocity estimates and the more robust separation of the environment in static and dynamic area. Additionally, we show that our proposed method for ego-motion compensation leads to comparable results in scenarios with stationary and with moving ego-vehicle.

Michael Ulrich, Claudius Gläser, Fabian Timm

This paper presents an novel object type classification method for automotive applications which uses deep learning with radar reflections. The method provides object class information such as pedestrian, cyclist, car, or non-obstacle. The method is both powerful and efficient, by using a light-weight deep learning approach on reflection level radar data. It fills the gap between low-performant methods of handcrafted features and high-performant methods with convolutional neural networks. The proposed network exploits the specific characteristics of radar reflection data: It handles unordered lists of arbitrary length as input and it combines both extraction of local and global features. In experiments with real data the proposed network outperforms existing methods of handcrafted or learned features. An ablation study analyzes the impact of the proposed global context layer.

Thomas Michalke, Di Feng, Claudius Gläser et al.

Lane detection is an essential part of the perception sub-architecture of any automated driving (AD) or advanced driver assistance system (ADAS). When focusing on low-cost, large scale products for automated driving, model-driven approaches for the detection of lane markings have proven good performance. More recently, data-driven approaches have been proposed that target the drivable area / freespace mainly in inner-city applications. Focus of these approaches is less on lane-based driving due to the fact that the lane concept does not fully apply in unstructured, residential inner-city environments. So-far the concept of drivable area is seldom used for highway and inter-urban applications due to the specific requirements of these scenarios that require clear lane associations of all traffic participants. We believe that lane-based, mapless driving in inter-urban and highway scenarios is still not fully handled with sufficient robustness and availability. Especially for challenging weather situations such as heavy rain, fog, low-standing sun, darkness or reflections in puddles, the mapless detection of lane markings decreases significantly or completely fails. We see potential in applying specifically designed data-driven freespace approaches in more lane-based driving applications for highways and inter-urban use. Therefore, we propose to classify specifically a drivable corridor of the ego lane on pixel level with a deep learning approach. Our approach is kept computationally efficient with only 0.66 million parameters allowing its application in large scale products. Thus, we were able to easily integrate into an online AD system of a test vehicle. We demonstrate the performance of our approach under challenging conditions qualitatively and quantitatively in comparison to a state-of-the-art model-driven approach.