Wolfram Burgard

Oier Mees, Lukas Hermann, Wolfram Burgard

A long-standing goal in robotics is to build robots that can perform a wide range of daily tasks from perceptions obtained with their onboard sensors and specified only via natural language. While recently substantial advances have been achieved in language-driven robotics by leveraging end-to-end learning from pixels, there is no clear and well-understood process for making various design choices due to the underlying variation in setups. In this paper, we conduct an extensive study of the most critical challenges in learning language conditioned policies from offline free-form imitation datasets. We further identify architectural and algorithmic techniques that improve performance, such as a hierarchical decomposition of the robot control learning, a multimodal transformer encoder, discrete latent plans and a self-supervised contrastive loss that aligns video and language representations. By combining the results of our investigation with our improved model components, we are able to present a novel approach that significantly outperforms the state of the art on the challenging language conditioned long-horizon robot manipulation CALVIN benchmark. We have open-sourced our implementation to facilitate future research in learning to perform many complex manipulation skills in a row specified with natural language. Codebase and trained models available at http://hulc.cs.uni-freiburg.de

Kshitij Sirohi, Sajad Marvi, Daniel Büscher et al.

Reliable scene understanding is indispensable for modern autonomous systems. Current learning-based methods typically try to maximize their performance based on segmentation metrics that only consider the quality of the segmentation. However, for the safe operation of a system in the real world it is crucial to consider the uncertainty in the prediction as well. In this work, we introduce the novel task of uncertainty-aware panoptic segmentation, which aims to predict per-pixel semantic and instance segmentations, together with per-pixel uncertainty estimates. We define two novel metrics to facilitate its quantitative analysis, the uncertainty-aware Panoptic Quality (uPQ) and the panoptic Expected Calibration Error (pECE). We further propose the novel top-down Evidential Panoptic Segmentation Network (EvPSNet) to solve this task. Our architecture employs a simple yet effective panoptic fusion module that leverages the predicted uncertainties. Furthermore, we provide several strong baselines combining state-of-the-art panoptic segmentation networks with sampling-free uncertainty estimation techniques. Extensive evaluations show that our EvPSNet achieves the new state-of-the-art for the standard Panoptic Quality (PQ), as well as for our uncertainty-aware panoptic metrics. We make the code available at: \url{https://github.com/kshitij3112/EvPSNet}

Guangming Wang, Yu Zheng, Yuxuan Wu et al.

Robot localization using a built map is essential for a variety of tasks including accurate navigation and mobile manipulation. A popular approach to robot localization is based on image-to-point cloud registration, which combines illumination-invariant LiDAR-based mapping with economical image-based localization. However, the recent works for image-to-point cloud registration either divide the registration into separate modules or project the point cloud to the depth image to register the RGB and depth images. In this paper, we present I2PNet, a novel end-to-end 2D-3D registration network, which directly registers the raw 3D point cloud with the 2D RGB image using differential modules with a united target. The 2D-3D cost volume module for differential 2D-3D association is proposed to bridge feature extraction and pose regression. The soft point-to-pixel correspondence is implicitly constructed on the intrinsic-independent normalized plane in the 2D-3D cost volume module. Moreover, we introduce an outlier mask prediction module to filter the outliers in the 2D-3D association before pose regression. Furthermore, we propose the coarse-to-fine 2D-3D registration architecture to increase localization accuracy. Extensive localization experiments are conducted on the KITTI, nuScenes, M2DGR, Argoverse, Waymo, and Lyft5 datasets. The results demonstrate that I2PNet outperforms the state-of-the-art by a large margin and has a higher efficiency than the previous works. Moreover, we extend the application of I2PNet to the camera-LiDAR online calibration and demonstrate that I2PNet outperforms recent approaches on the online calibration task. Source codes are released at https://github.com/IRMVLab/I2PNet.

Kshitij Sirohi, Sajad Marvi, Daniel Büscher et al.

Modern autonomous systems often rely on LiDAR scanners, in particular for autonomous driving scenarios. In this context, reliable scene understanding is indispensable. Current learning-based methods typically try to achieve maximum performance for this task, while neglecting a proper estimation of the associated uncertainties. In this work, we introduce a novel approach for solving the task of uncertainty-aware panoptic segmentation using LiDAR point clouds. Our proposed EvLPSNet network is the first to solve this task efficiently in a sampling-free manner. It aims to predict per-point semantic and instance segmentations, together with per-point uncertainty estimates. Moreover, it incorporates methods for improving the performance by employing the predicted uncertainties. We provide several strong baselines combining state-of-the-art panoptic segmentation networks with sampling-free uncertainty estimation techniques. Extensive evaluations show that we achieve the best performance on uncertainty-aware panoptic segmentation quality and calibration compared to these baselines. We make our code available at: https://github.com/kshitij3112/EvLPSNet

Chenguang Huang, Oier Mees, Andy Zeng et al.

Grounding language to the visual observations of a navigating agent can be performed using off-the-shelf visual-language models pretrained on Internet-scale data (e.g., image captions). While this is useful for matching images to natural language descriptions of object goals, it remains disjoint from the process of mapping the environment, so that it lacks the spatial precision of classic geometric maps. To address this problem, we propose VLMaps, a spatial map representation that directly fuses pretrained visual-language features with a 3D reconstruction of the physical world. VLMaps can be autonomously built from video feed on robots using standard exploration approaches and enables natural language indexing of the map without additional labeled data. Specifically, when combined with large language models (LLMs), VLMaps can be used to (i) translate natural language commands into a sequence of open-vocabulary navigation goals (which, beyond prior work, can be spatial by construction, e.g., "in between the sofa and TV" or "three meters to the right of the chair") directly localized in the map, and (ii) can be shared among multiple robots with different embodiments to generate new obstacle maps on-the-fly (by using a list of obstacle categories). Extensive experiments carried out in simulated and real world environments show that VLMaps enable navigation according to more complex language instructions than existing methods. Videos are available at https://vlmaps.github.io.

Oier Mees, Jessica Borja-Diaz, Wolfram Burgard

Recent works have shown that Large Language Models (LLMs) can be applied to ground natural language to a wide variety of robot skills. However, in practice, learning multi-task, language-conditioned robotic skills typically requires large-scale data collection and frequent human intervention to reset the environment or help correcting the current policies. In this work, we propose a novel approach to efficiently learn general-purpose language-conditioned robot skills from unstructured, offline and reset-free data in the real world by exploiting a self-supervised visuo-lingual affordance model, which requires annotating as little as 1% of the total data with language. We evaluate our method in extensive experiments both in simulated and real-world robotic tasks, achieving state-of-the-art performance on the challenging CALVIN benchmark and learning over 25 distinct visuomotor manipulation tasks with a single policy in the real world. We find that when paired with LLMs to break down abstract natural language instructions into subgoals via few-shot prompting, our method is capable of completing long-horizon, multi-tier tasks in the real world, while requiring an order of magnitude less data than previous approaches. Code and videos are available at http://hulc2.cs.uni-freiburg.de

Jannik Zürn, Ingmar Posner, Wolfram Burgard · oxford

Lane graph estimation is a long-standing problem in the context of autonomous driving. Previous works aimed at solving this problem by relying on large-scale, hand-annotated lane graphs, introducing a data bottleneck for training models to solve this task. To overcome this limitation, we propose to use the motion patterns of traffic participants as lane graph annotations. In our AutoGraph approach, we employ a pre-trained object tracker to collect the tracklets of traffic participants such as vehicles and trucks. Based on the location of these tracklets, we predict the successor lane graph from an initial position using overhead RGB images only, not requiring any human supervision. In a subsequent stage, we show how the individual successor predictions can be aggregated into a consistent lane graph. We demonstrate the efficacy of our approach on the UrbanLaneGraph dataset and perform extensive quantitative and qualitative evaluations, indicating that AutoGraph is on par with models trained on hand-annotated graph data. Model and dataset will be made available at redacted-for-review.

Erick Rosete-Beas, Oier Mees, Gabriel Kalweit et al.

Everyday tasks of long-horizon and comprising a sequence of multiple implicit subtasks still impose a major challenge in offline robot control. While a number of prior methods aimed to address this setting with variants of imitation and offline reinforcement learning, the learned behavior is typically narrow and often struggles to reach configurable long-horizon goals. As both paradigms have complementary strengths and weaknesses, we propose a novel hierarchical approach that combines the strengths of both methods to learn task-agnostic long-horizon policies from high-dimensional camera observations. Concretely, we combine a low-level policy that learns latent skills via imitation learning and a high-level policy learned from offline reinforcement learning for skill-chaining the latent behavior priors. Experiments in various simulated and real robot control tasks show that our formulation enables producing previously unseen combinations of skills to reach temporally extended goals by "stitching" together latent skills through goal chaining with an order-of-magnitude improvement in performance upon state-of-the-art baselines. We even learn one multi-task visuomotor policy for 25 distinct manipulation tasks in the real world which outperforms both imitation learning and offline reinforcement learning techniques.

Jannik Zürn, Sebastian Weber, Wolfram Burgard · oxford

Robustly classifying ground infrastructure such as roads and street crossings is an essential task for mobile robots operating alongside pedestrians. While many semantic segmentation datasets are available for autonomous vehicles, models trained on such datasets exhibit a large domain gap when deployed on robots operating in pedestrian spaces. Manually annotating images recorded from pedestrian viewpoints is both expensive and time-consuming. To overcome this challenge, we propose TrackletMapper, a framework for annotating ground surface types such as sidewalks, roads, and street crossings from object tracklets without requiring human-annotated data. To this end, we project the robot ego-trajectory and the paths of other traffic participants into the ego-view camera images, creating sparse semantic annotations for multiple types of ground surfaces from which a ground segmentation model can be trained. We further show that the model can be self-distilled for additional performance benefits by aggregating a ground surface map and projecting it into the camera images, creating a denser set of training annotations compared to the sparse tracklet annotations. We qualitatively and quantitatively attest our findings on a novel large-scale dataset for mobile robots operating in pedestrian areas. Code and dataset will be made available at http://trackletmapper.cs.uni-freiburg.de.

Chenguang Huang, Oier Mees, Andy Zeng et al.

While interacting in the world is a multi-sensory experience, many robots continue to predominantly rely on visual perception to map and navigate in their environments. In this work, we propose Audio-Visual-Language Maps (AVLMaps), a unified 3D spatial map representation for storing cross-modal information from audio, visual, and language cues. AVLMaps integrate the open-vocabulary capabilities of multimodal foundation models pre-trained on Internet-scale data by fusing their features into a centralized 3D voxel grid. In the context of navigation, we show that AVLMaps enable robot systems to index goals in the map based on multimodal queries, e.g., textual descriptions, images, or audio snippets of landmarks. In particular, the addition of audio information enables robots to more reliably disambiguate goal locations. Extensive experiments in simulation show that AVLMaps enable zero-shot multimodal goal navigation from multimodal prompts and provide 50% better recall in ambiguous scenarios. These capabilities extend to mobile robots in the real world - navigating to landmarks referring to visual, audio, and spatial concepts. Videos and code are available at: https://avlmaps.github.io.

Jessica Borja-Diaz, Oier Mees, Gabriel Kalweit et al.

Robots operating in human-centered environments should have the ability to understand how objects function: what can be done with each object, where this interaction may occur, and how the object is used to achieve a goal. To this end, we propose a novel approach that extracts a self-supervised visual affordance model from human teleoperated play data and leverages it to enable efficient policy learning and motion planning. We combine model-based planning with model-free deep reinforcement learning (RL) to learn policies that favor the same object regions favored by people, while requiring minimal robot interactions with the environment. We evaluate our algorithm, Visual Affordance-guided Policy Optimization (VAPO), with both diverse simulation manipulation tasks and real world robot tidy-up experiments to demonstrate the effectiveness of our affordance-guided policies. We find that our policies train 4x faster than the baselines and generalize better to novel objects because our visual affordance model can anticipate their affordance regions.

Nikhil Gosala, Kürsat Petek, Paulo L. J. Drews-Jr et al.

Bird's-Eye-View (BEV) semantic maps have become an essential component of automated driving pipelines due to the rich representation they provide for decision-making tasks. However, existing approaches for generating these maps still follow a fully supervised training paradigm and hence rely on large amounts of annotated BEV data. In this work, we address this limitation by proposing the first self-supervised approach for generating a BEV semantic map using a single monocular image from the frontal view (FV). During training, we overcome the need for BEV ground truth annotations by leveraging the more easily available FV semantic annotations of video sequences. Thus, we propose the SkyEye architecture that learns based on two modes of self-supervision, namely, implicit supervision and explicit supervision. Implicit supervision trains the model by enforcing spatial consistency of the scene over time based on FV semantic sequences, while explicit supervision exploits BEV pseudolabels generated from FV semantic annotations and self-supervised depth estimates. Extensive evaluations on the KITTI-360 dataset demonstrate that our self-supervised approach performs on par with the state-of-the-art fully supervised methods and achieves competitive results using only 1% of direct supervision in the BEV compared to fully supervised approaches. Finally, we publicly release both our code and the BEV datasets generated from the KITTI-360 and Waymo datasets.

Martin Büchner, Jannik Zürn, Ion-George Todoran et al.

Lane graph estimation is an essential and highly challenging task in automated driving and HD map learning. Existing methods using either onboard or aerial imagery struggle with complex lane topologies, out-of-distribution scenarios, or significant occlusions in the image space. Moreover, merging overlapping lane graphs to obtain consistent large-scale graphs remains difficult. To overcome these challenges, we propose a novel bottom-up approach to lane graph estimation from aerial imagery that aggregates multiple overlapping graphs into a single consistent graph. Due to its modular design, our method allows us to address two complementary tasks: predicting ego-respective successor lane graphs from arbitrary vehicle positions using a graph neural network and aggregating these predictions into a consistent global lane graph. Extensive experiments on a large-scale lane graph dataset demonstrate that our approach yields highly accurate lane graphs, even in regions with severe occlusions. The presented approach to graph aggregation proves to eliminate inconsistent predictions while increasing the overall graph quality. We make our large-scale urban lane graph dataset and code publicly available at http://urbanlanegraph.cs.uni-freiburg.de.

Niclas Vödisch, Kürsat Petek, Wolfram Burgard et al.

Operating a robot in the open world requires a high level of robustness with respect to previously unseen environments. Optimally, the robot is able to adapt by itself to new conditions without human supervision, e.g., automatically adjusting its perception system to changing lighting conditions. In this work, we address the task of continual learning for deep learning-based monocular depth estimation and panoptic segmentation in new environments in an online manner. We introduce CoDEPS to perform continual learning involving multiple real-world domains while mitigating catastrophic forgetting by leveraging experience replay. In particular, we propose a novel domain-mixing strategy to generate pseudo-labels to adapt panoptic segmentation. Furthermore, we explicitly address the limited storage capacity of robotic systems by leveraging sampling strategies for constructing a fixed-size replay buffer based on rare semantic class sampling and image diversity. We perform extensive evaluations of CoDEPS on various real-world datasets demonstrating that it successfully adapts to unseen environments without sacrificing performance on previous domains while achieving state-of-the-art results. The code of our work is publicly available at http://codeps.cs.uni-freiburg.de.

Markus Käppeler, Kürsat Petek, Niclas Vödisch et al.

Current state-of-the-art methods for panoptic segmentation require an immense amount of annotated training data that is both arduous and expensive to obtain posing a significant challenge for their widespread adoption. Concurrently, recent breakthroughs in visual representation learning have sparked a paradigm shift leading to the advent of large foundation models that can be trained with completely unlabeled images. In this work, we propose to leverage such task-agnostic image features to enable few-shot panoptic segmentation by presenting Segmenting Panoptic Information with Nearly 0 labels (SPINO). In detail, our method combines a DINOv2 backbone with lightweight network heads for semantic segmentation and boundary estimation. We show that our approach, albeit being trained with only ten annotated images, predicts high-quality pseudo-labels that can be used with any existing panoptic segmentation method. Notably, we demonstrate that SPINO achieves competitive results compared to fully supervised baselines while using less than 0.3% of the ground truth labels, paving the way for learning complex visual recognition tasks leveraging foundation models. To illustrate its general applicability, we further deploy SPINO on real-world robotic vision systems for both outdoor and indoor environments. To foster future research, we make the code and trained models publicly available at http://spino.cs.uni-freiburg.de.

Monish R. Nallapareddy, Kshitij Sirohi, Paulo L. J. Drews-Jr et al.

Uncertainty estimation is crucial in safety-critical settings such as automated driving as it provides valuable information for several downstream tasks including high-level decision making and path planning. In this work, we propose EvCenterNet, a novel uncertainty-aware 2D object detection framework using evidential learning to directly estimate both classification and regression uncertainties. To employ evidential learning for object detection, we devise a combination of evidential and focal loss functions for the sparse heatmap inputs. We introduce class-balanced weighting for regression and heatmap prediction to tackle the class imbalance encountered by evidential learning. Moreover, we propose a learning scheme to actively utilize the predicted heatmap uncertainties to improve the detection performance by focusing on the most uncertain points. We train our model on the KITTI dataset and evaluate it on challenging out-of-distribution datasets including BDD100K and nuImages. Our experiments demonstrate that our approach improves the precision and minimizes the execution time loss in relation to the base model.

D. Adriana Gómez-Rosal, Max Bergau, Georg K. J. Fischer et al.

In today's chemical plants, human field operators perform frequent integrity checks to guarantee high safety standards, and thus are possibly the first to encounter dangerous operating conditions. To alleviate their task, we present a system consisting of an autonomously navigating robot integrated with various sensors and intelligent data processing. It is able to detect methane leaks and estimate its flow rate, detect more general gas anomalies, recognize oil films, localize sound sources and detect failure cases, map the environment in 3D, and navigate autonomously, employing recognition and avoidance of dynamic obstacles. We evaluate our system at a wastewater facility in full working conditions. Our results demonstrate that the system is able to robustly navigate the plant and provide useful information about critical operating conditions.

Johan Vertens, Nicolai Dorka, Tim Welschehold et al.

The safe deployment of autonomous vehicles relies on their ability to effectively react to environmental changes. This can require maneuvering on varying surfaces which is still a difficult problem, especially for slippery terrains. To address this issue we propose a new approach that learns a surface-aware dynamics model by conditioning it on a latent variable vector storing surface information about the current location. A latent mapper is trained to update these latent variables during inference from multiple modalities on every traversal of the corresponding locations and stores them in a map. By training everything end-to-end with the loss of the dynamics model, we enforce the latent mapper to learn an update rule for the latent map that is useful for the subsequent dynamics model. We implement and evaluate our approach on a real miniature electric car. The results show that the latent map is updated to allow more accurate predictions of the dynamics model compared to a model without this information. We further show that by using this model, the driving performance can be improved on varying and challenging surfaces.

Nicolai Dorka, Tim Welschehold, Wolfram Burgard

Early stopping based on the validation set performance is a popular approach to find the right balance between under- and overfitting in the context of supervised learning. However, in reinforcement learning, even for supervised sub-problems such as world model learning, early stopping is not applicable as the dataset is continually evolving. As a solution, we propose a new general method that dynamically adjusts the update to data (UTD) ratio during training based on under- and overfitting detection on a small subset of the continuously collected experience not used for training. We apply our method to DreamerV2, a state-of-the-art model-based reinforcement learning algorithm, and evaluate it on the DeepMind Control Suite and the Atari $100$k benchmark. The results demonstrate that one can better balance under- and overestimation by adjusting the UTD ratio with our approach compared to the default setting in DreamerV2 and that it is competitive with an extensive hyperparameter search which is not feasible for many applications. Our method eliminates the need to set the UTD hyperparameter by hand and even leads to a higher robustness with regard to other learning-related hyperparameters further reducing the amount of necessary tuning.

Johan Vertens, Wolfram Burgard

In this paper we propose USegScene, a framework for semantically guided unsupervised learning of depth, optical flow and ego-motion estimation for stereo camera images using convolutional neural networks. Our framework leverages semantic information for improved regularization of depth and optical flow maps, multimodal fusion and occlusion filling considering dynamic rigid object motions as independent SE(3) transformations. Furthermore, complementary to pure photo-metric matching, we propose matching of semantic features, pixel-wise classes and object instance borders between the consecutive images. In contrast to previous methods, we propose a network architecture that jointly predicts all outputs using shared encoders and allows passing information across the task-domains, e.g., the prediction of optical flow can benefit from the prediction of the depth. Furthermore, we explicitly learn the depth and optical flow occlusion maps inside the network, which are leveraged in order to improve the predictions in therespective regions. We present results on the popular KITTI dataset and show that our approach outperforms other methods by a large margin.

Michael G. Adam, Sebastian Eger, Martin Piccolrovazzi et al.

As labor shortage increases in the health sector, the demand for assistive robotics grows. However, the needed test data to develop those robots is scarce, especially for the application of active 3D object detection, where no real data exists at all. This short paper counters this by introducing such an annotated dataset of real environments. The captured environments represent areas which are already in use in the field of robotic health care research. We further provide ground truth data within one room, for assessing SLAM algorithms running directly on a health care robot.

Iman Nematollahi, Erick Rosete-Beas, Seyed Mahdi B. Azad et al.

For autonomous skill acquisition, robots have to learn about the physical rules governing the 3D world dynamics from their own past experience to predict and reason about plausible future outcomes. To this end, we propose a transformation-based 3D video prediction (T3VIP) approach that explicitly models the 3D motion by decomposing a scene into its object parts and predicting their corresponding rigid transformations. Our model is fully unsupervised, captures the stochastic nature of the real world, and the observational cues in image and point cloud domains constitute its learning signals. To fully leverage all the 2D and 3D observational signals, we equip our model with automatic hyperparameter optimization (HPO) to interpret the best way of learning from them. To the best of our knowledge, our model is the first generative model that provides an RGB-D video prediction of the future for a static camera. Our extensive evaluation with simulated and real-world datasets demonstrates that our formulation leads to interpretable 3D models that predict future depth videos while achieving on-par performance with 2D models on RGB video prediction. Moreover, we demonstrate that our model outperforms 2D baselines on visuomotor control. Videos, code, dataset, and pre-trained models are available at http://t3vip.cs.uni-freiburg.de.

Iman Nematollahi, Kirill Yankov, Wolfram Burgard et al.

A long-standing challenge for a robotic manipulation system operating in real-world scenarios is adapting and generalizing its acquired motor skills to unseen environments. We tackle this challenge employing hybrid skill models that integrate imitation and reinforcement paradigms, to explore how the learning and adaptation of a skill, along with its core grounding in the scene through a learned keypoint, can facilitate such generalization. To that end, we develop Keypoint Integrated Soft Actor-Critic Gaussian Mixture Models (KIS-GMM) approach that learns to predict the reference of a dynamical system within the scene as a 3D keypoint, leveraging visual observations obtained by the robot's physical interactions during skill learning. Through conducting comprehensive evaluations in both simulated and real-world environments, we show that our method enables a robot to gain a significant zero-shot generalization to novel environments and to refine skills in the target environments faster than learning from scratch. Importantly, this is achieved without the need for new ground truth data. Moreover, our method effectively copes with scene displacements.

Jonas Schramm, Niclas Vödisch, Kürsat Petek et al.

Semantic scene segmentation from a bird's-eye-view (BEV) perspective plays a crucial role in facilitating planning and decision-making for mobile robots. Although recent vision-only methods have demonstrated notable advancements in performance, they often struggle under adverse illumination conditions such as rain or nighttime. While active sensors offer a solution to this challenge, the prohibitively high cost of LiDARs remains a limiting factor. Fusing camera data with automotive radars poses a more inexpensive alternative but has received less attention in prior research. In this work, we aim to advance this promising avenue by introducing BEVCar, a novel approach for joint BEV object and map segmentation. The core novelty of our approach lies in first learning a point-based encoding of raw radar data, which is then leveraged to efficiently initialize the lifting of image features into the BEV space. We perform extensive experiments on the nuScenes dataset and demonstrate that BEVCar outperforms the current state of the art. Moreover, we show that incorporating radar information significantly enhances robustness in challenging environmental conditions and improves segmentation performance for distant objects. To foster future research, we provide the weather split of the nuScenes dataset used in our experiments, along with our code and trained models at http://bevcar.cs.uni-freiburg.de.

Siting Zhu, Ziyun Lu, Guangming Wang et al.

Open-vocabulary scene understanding is crucial for robotic applications, enabling robots to comprehend complex 3D environmental contexts and supporting various downstream tasks such as navigation and manipulation. However, existing methods require pre-built complete 3D semantic maps to construct scene graphs for scene understanding, which limits their applicability in robotic scenarios where environments are explored incrementally. To address this challenge, we propose OGScene3D, an open-vocabulary scene understanding system that achieves accurate 3D semantic mapping and scene graph construction incrementally. Our system employs a confidence-based Gaussian semantic representation that jointly models semantic predictions and their reliability, enabling robust scene modeling. Building on this representation, we introduce a hierarchical 3D semantic optimization strategy that achieves semantic consistency through local correspondence establishment and global refinement, thereby constructing globally consistent semantic maps. Moreover, we design a long-term global optimization method that leverages temporal memory of historical observations to enhance semantic predictions. By integrating 2D-3D semantic consistency with Gaussian rendering contribution, this method continuously refines the semantic understanding of the entire scene. Furthermore, we develop a progressive graph construction approach that dynamically creates and updates both nodes and semantic relationships, allowing continuous updating of the 3D scene graphs. Extensive experiments on widely used datasets and real-world scenes demonstrate the effectiveness of our OGScene3D on open-vocabulary scene understanding.

Yue Yao, Shengchao Yan, Daniel Goehring et al.

Robustness against Out-of-Distribution (OoD) samples is a key performance indicator of a trajectory prediction model. However, the development and ranking of state-of-the-art (SotA) models are driven by their In-Distribution (ID) performance on individual competition datasets. We present an OoD testing protocol that homogenizes datasets and prediction tasks across two large-scale motion datasets. We introduce a novel prediction algorithm based on polynomial representations for agent trajectory and road geometry on both the input and output sides of the model. With a much smaller model size, training effort, and inference time, we reach near SotA performance for ID testing and significantly improve robustness in OoD testing. Within our OoD testing protocol, we further study two augmentation strategies of SotA models and their effects on model generalization. Highlighting the contrast between ID and OoD performance, we suggest adding OoD testing to the evaluation criteria of trajectory prediction models.

Sai Prasanna, Daniel Honerkamp, Kshitij Sirohi et al.

Embodied AI has made significant progress acting in unexplored environments. However, tasks such as object search have largely focused on efficient policy learning. In this work, we identify several gaps in current search methods: They largely focus on dated perception models, neglect temporal aggregation, and transfer from ground truth directly to noisy perception at test time, without accounting for the resulting overconfidence in the perceived state. We address the identified problems through calibrated perception probabilities and uncertainty across aggregation and found decisions, thereby adapting the models for sequential tasks. The resulting methods can be directly integrated with pretrained models across a wide family of existing search approaches at no additional training cost. We perform extensive evaluations of aggregation methods across both different semantic perception models and policies, confirming the importance of calibrated uncertainties in both the aggregation and found decisions. We make the code and trained models available at https://semantic-search.cs.uni-freiburg.de.

Tobias Jülg, Wolfram Burgard, Florian Walter

Vision-Language-Action Models (VLAs) have demonstrated remarkable generalization capabilities in real-world experiments. However, their success rates are often not on par with expert policies, and they require fine-tuning when the setup changes. In this work, we introduce Refined Policy Distillation (RPD), a novel Reinforcement Learning (RL)-based policy refinement method that bridges this performance gap through a combination of on-policy RL with behavioral cloning. The core idea of RPD is to distill and refine VLAs into compact, high-performing expert policies by guiding the student policy during RL exploration using the actions of a teacher VLA, resulting in increased sample efficiency and faster convergence. We complement our method by fine-tuned versions of Octo and OpenVLA for ManiSkill3 to evaluate RPD in simulation. While this is a key requirement for applying RL, it also yields new insights beyond existing studies on VLA performance in real-world settings. Our experimental results across various manipulation tasks show that RPD enables the RL student to learn expert policies that outperform the VLA teacher in both dense and sparse reward settings, while also achieving faster convergence than the RL baseline. Our approach is even robust to changes in camera perspective and can generalize to task variations that the underlying VLA cannot solve. Our code, dataset, VLA checkpoints, and videos are available at https://refined-policy-distillation.github.io

Yannik Blei, Michael Krawez, Nisarga Nilavadi et al.

Nowadays, unmanned aerial vehicles (UAVs) are commonly used in search and rescue scenarios to gather information in the search area. The automatic identification of the person searched for in aerial footage could increase the autonomy of such systems, reduce the search time, and thus increase the missed person's chances of survival. In this paper, we present a novel approach to perform semantically conditioned open vocabulary object tracking that is specifically designed to cope with the limitations of UAV hardware. Our approach has several advantages. It can run with verbal descriptions of the missing person, e.g., the color of the shirt, it does not require dedicated training to execute the mission and can efficiently track a potentially moving person. Our experimental results demonstrate the versatility and efficacy of our approach.

Pierre Krack, Tobias Jülg, Wolfram Burgard et al.

Well-designed dense reward functions in robot manipulation not only indicate whether a task is completed but also encode progress along the way. Generally, designing dense rewards is challenging and usually requires access to privileged state information available only in simulation, not in real-world experiments. This makes reward prediction models that infer task state information from camera images attractive. A common approach is to predict rewards from expert demonstrations based on visual similarity or sequential frame ordering. However, this biases the resulting reward function towards a specific solution and leaves it undefined in states not covered by the demonstrations. In this work, we introduce Rewarding DINO, a method for language-conditioned reward modeling that learns actual reward functions rather than specific trajectories. The model's compact size allows it to serve as a direct replacement for analytical reward functions with comparatively low computational overhead. We train our model on data sampled from 24 Meta-World+ tasks using a rank-based loss and evaluate pairwise accuracy, rank correlation, and calibration. Rewarding DINO achieves competitive performance in tasks from the training set and generalizes to new settings in simulation and the real world, indicating that it learns task semantics. We also test the model with off-the-shelf reinforcement learning algorithms to solve tasks from our Meta-World+ training set.

Johannes Hechtl, Philipp Schmitt, Georg von Wichert et al.

While vision-language-action (VLA) models have shown great promise for robot manipulation, their deployment on rigid industrial robots remains challenging due to the inherent trade-off between compliance and responsiveness. Standard Behavior Cloning (BC) approaches predict discrete poses at low frequencies, omitting the velocity and acceleration feedforward terms typically used by low-level compliant controllers. This requires to rely on high stiffness for accurate tracking, thereby sacrificing safe contact dynamics. In this paper, we demonstrate the importance of integrating velocity feedforward terms into VLA policies to resolve this trade-off. We propose two methods for extracting velocity targets from VLAs: a time-discrete finite-difference approximation that serves as a highly effective bridge for existing models, and a continuous Cubic B-Spline action space that natively yields $C^2$ continuous trajectories for high-frequency control. Crucially, both approaches are strictly model-agnostic and compatible with any standard action-chunking architecture, requiring modifications only to teleoperation, data processing, and the low-level controller. We fine-tune the $π_{0.5}$ model and evaluate both of our approaches on a demanding, contact-rich cube-in-hole task. Our results indicate that incorporating the velocity feedforward term via finite differences significantly improves task execution speed, while the continuous B-Spline approach maintains high overall success rates and provides a foundation for smoother higher-order derivatives without compromising compliance.

Abdelrhman Werby, Chenguang Huang, Martin Büchner et al.

Recent open-vocabulary robot mapping methods enrich dense geometric maps with pre-trained visual-language features. While these maps allow for the prediction of point-wise saliency maps when queried for a certain language concept, large-scale environments and abstract queries beyond the object level still pose a considerable hurdle, ultimately limiting language-grounded robotic navigation. In this work, we present HOV-SG, a hierarchical open-vocabulary 3D scene graph mapping approach for language-grounded robot navigation. Leveraging open-vocabulary vision foundation models, we first obtain state-of-the-art open-vocabulary segment-level maps in 3D and subsequently construct a 3D scene graph hierarchy consisting of floor, room, and object concepts, each enriched with open-vocabulary features. Our approach is able to represent multi-story buildings and allows robotic traversal of those using a cross-floor Voronoi graph. HOV-SG is evaluated on three distinct datasets and surpasses previous baselines in open-vocabulary semantic accuracy on the object, room, and floor level while producing a 75% reduction in representation size compared to dense open-vocabulary maps. In order to prove the efficacy and generalization capabilities of HOV-SG, we showcase successful long-horizon language-conditioned robot navigation within real-world multi-storage environments. We provide code and trial video data at http://hovsg.github.io/.

Tobias Jülg, Pierre Krack, Seongjin Bien et al.

Vision-Language-Action models (VLAs) mark a major shift in robot learning. They replace specialized architectures and task-tailored components of expert policies with large-scale data collection and setup-specific fine-tuning. In this machine learning-focused workflow that is centered around models and scalable training, traditional robotics software frameworks become a bottleneck, while robot simulations offer only limited support for transitioning from and to real-world experiments. In this work, we close this gap by introducing Robot Control Stack (RCS), a lean ecosystem designed from the ground up to support research in robot learning with large-scale generalist policies. At its core, RCS features a modular and easily extensible layered architecture with a unified interface for simulated and physical robots, facilitating sim-to-real transfer. Despite its minimal footprint and dependencies, it offers a complete feature set, enabling both real-world experiments and large-scale training in simulation. Our contribution is twofold: First, we introduce the architecture of RCS and explain its design principles. Second, we evaluate its usability and performance along the development cycle of VLA and RL policies. Our experiments also provide an extensive evaluation of Octo, OpenVLA, and Pi Zero on multiple robots and shed light on how simulation data can improve real-world policy performance. Our code, datasets, weights, and videos are available at: https://robotcontrolstack.github.io/

Oier Mees, Lukas Hermann, Erick Rosete-Beas et al.

General-purpose robots coexisting with humans in their environment must learn to relate human language to their perceptions and actions to be useful in a range of daily tasks. Moreover, they need to acquire a diverse repertoire of general-purpose skills that allow composing long-horizon tasks by following unconstrained language instructions. In this paper, we present CALVIN (Composing Actions from Language and Vision), an open-source simulated benchmark to learn long-horizon language-conditioned tasks. Our aim is to make it possible to develop agents that can solve many robotic manipulation tasks over a long horizon, from onboard sensors, and specified only via human language. CALVIN tasks are more complex in terms of sequence length, action space, and language than existing vision-and-language task datasets and supports flexible specification of sensor suites. We evaluate the agents in zero-shot to novel language instructions and to novel environments and objects. We show that a baseline model based on multi-context imitation learning performs poorly on CALVIN, suggesting that there is significant room for developing innovative agents that learn to relate human language to their world models with this benchmark.

Artemij Amiranashvili, Nicolai Dorka, Wolfram Burgard et al.

Imitation learning is a powerful family of techniques for learning sensorimotor coordination in immersive environments. We apply imitation learning to attain state-of-the-art performance on hard exploration problems in the Minecraft environment. We report experiments that highlight the influence of network architecture, loss function, and data augmentation. An early version of our approach reached second place in the MineRL competition at NeurIPS 2019. Here we report stronger results that can be used as a starting point for future competition entries and related research. Our code is available at https://github.com/amiranas/minerl_imitation_learning.

Alexander Schaefer, Daniel Büscher, Johan Vertens et al.

Due to their ubiquity and long-term stability, pole-like objects are well suited to serve as landmarks for vehicle localization in urban environments. In this work, we present a complete mapping and long-term localization system based on pole landmarks extracted from 3-D lidar data. Our approach features a novel pole detector, a mapping module, and an online localization module, each of which are described in detail, and for which we provide an open-source implementation at www.github.com/acschaefer/polex. In extensive experiments, we demonstrate that our method improves on the state of the art with respect to long-term reliability and accuracy: First, we prove reliability by tasking the system with localizing a mobile robot over the course of 15~months in an urban area based on an initial map, confronting it with constantly varying routes, differing weather conditions, seasonal changes, and construction sites. Second, we show that the proposed approach clearly outperforms a recently published method in terms of accuracy.

Alexander Schaefer, Johan Vertens, Daniel Büscher et al.

Whether it is object detection, model reconstruction, laser odometry, or point cloud registration: Plane extraction is a vital component of many robotic systems. In this paper, we propose a strictly probabilistic method to detect finite planes in organized 3-D laser range scans. An agglomerative hierarchical clustering technique, our algorithm builds planes from bottom up, always extending a plane by the point that decreases the measurement likelihood of the scan the least. In contrast to most related methods, which rely on heuristics like orthogonal point-to-plane distance, we leverage the ray path information to compute the measurement likelihood. We evaluate our approach not only on the popular SegComp benchmark, but also provide a challenging synthetic dataset that overcomes SegComp's deficiencies. Both our implementation and the suggested dataset are available at www.github.com/acschaefer/ppe.

Alexander Schaefer, Daniel Büscher, Lukas Luft et al.

Man-made environments such as households, offices, or factory floors are typically composed of linear structures. Accordingly, polylines are a natural way to accurately represent their geometry. In this paper, we propose a novel probabilistic method to extract polylines from raw 2-D laser range scans. The key idea of our approach is to determine a set of polylines that maximizes the likelihood of a given scan. In extensive experiments carried out on publicly available real-world datasets and on simulated laser scans, we demonstrate that our method substantially outperforms existing state-of-the-art approaches in terms of accuracy, while showing comparable computational requirements. Our implementation is available under https://github.com/acschaefer/ple.

Robin Tibor Schirrmeister, Jost Tobias Springenberg, Lukas Dominique Josef Fiederer et al.

PLEASE READ AND CITE THE REVISED VERSION at Human Brain Mapping: http://onlinelibrary.wiley.com/doi/10.1002/hbm.23730/full Code available here: https://github.com/robintibor/braindecode

Eugenio Chisari, Nick Heppert, Tim Welschehold et al.

Reliable object grasping is a crucial capability for autonomous robots. However, many existing grasping approaches focus on general clutter removal without explicitly modeling objects and thus only relying on the visible local geometry. We introduce CenterGrasp, a novel framework that combines object awareness and holistic grasping. CenterGrasp learns a general object prior by encoding shapes and valid grasps in a continuous latent space. It consists of an RGB-D image encoder that leverages recent advances to detect objects and infer their pose and latent code, and a decoder to predict shape and grasps for each object in the scene. We perform extensive experiments on simulated as well as real-world cluttered scenes and demonstrate strong scene reconstruction and 6-DoF grasp-pose estimation performance. Compared to the state of the art, CenterGrasp achieves an improvement of 38.5 mm in shape reconstruction and 33 percentage points on average in grasp success. We make the code and trained models publicly available at http://centergrasp.cs.uni-freiburg.de.

Iman Nematollahi, Branton DeMoss, Akshay L Chandra et al.

We introduce LUMOS, a language-conditioned multi-task imitation learning framework for robotics. LUMOS learns skills by practicing them over many long-horizon rollouts in the latent space of a learned world model and transfers these skills zero-shot to a real robot. By learning on-policy in the latent space of the learned world model, our algorithm mitigates policy-induced distribution shift which most offline imitation learning methods suffer from. LUMOS learns from unstructured play data with fewer than 1% hindsight language annotations but is steerable with language commands at test time. We achieve this coherent long-horizon performance by combining latent planning with both image- and language-based hindsight goal relabeling during training, and by optimizing an intrinsic reward defined in the latent space of the world model over multiple time steps, effectively reducing covariate shift. In experiments on the difficult long-horizon CALVIN benchmark, LUMOS outperforms prior learning-based methods with comparable approaches on chained multi-task evaluations. To the best of our knowledge, we are the first to learn a language-conditioned continuous visuomotor control for a real-world robot within an offline world model. Videos, dataset and code are available at http://lumos.cs.uni-freiburg.de.

Abdelrhman Werby, Martin Büchner, Adrian Röfer et al.

Understanding the 3D motion of articulated objects is essential in robotic scene understanding, mobile manipulation, and motion planning. Prior methods for articulation estimation have primarily focused on controlled settings, assuming either fixed camera viewpoints or direct observations of various object states, which tend to fail in more realistic unconstrained environments. In contrast, humans effortlessly infer articulation by watching others manipulate objects. Inspired by this, we introduce ArtiPoint, a novel estimation framework that can infer articulated object models under dynamic camera motion and partial observability. By combining deep point tracking with a factor graph optimization framework, ArtiPoint robustly estimates articulated part trajectories and articulation axes directly from raw RGB-D videos. To foster future research in this domain, we introduce Arti4D, the first ego-centric in-the-wild dataset that captures articulated object interactions at a scene level, accompanied by articulation labels and ground-truth camera poses. We benchmark ArtiPoint against a range of classical and learning-based baselines, demonstrating its superior performance on Arti4D. We make code and Arti4D publicly available at https://artipoint.cs.uni-freiburg.de.

Akshay L Chandra, Iman Nematollahi, Chenguang Huang et al.

Fine-tuning diffusion policies with reinforcement learning (RL) presents significant challenges. The long denoising sequence for each action prediction impedes effective reward propagation. Moreover, standard RL methods require millions of real-world interactions, posing a major bottleneck for practical fine-tuning. Although prior work frames the denoising process in diffusion policies as a Markov Decision Process to enable RL-based updates, its strong dependence on environment interaction remains highly inefficient. To bridge this gap, we introduce DiWA, a novel framework that leverages a world model for fine-tuning diffusion-based robotic skills entirely offline with reinforcement learning. Unlike model-free approaches that require millions of environment interactions to fine-tune a repertoire of robot skills, DiWA achieves effective adaptation using a world model trained once on a few hundred thousand offline play interactions. This results in dramatically improved sample efficiency, making the approach significantly more practical and safer for real-world robot learning. On the challenging CALVIN benchmark, DiWA improves performance across eight tasks using only offline adaptation, while requiring orders of magnitude fewer physical interactions than model-free baselines. To our knowledge, this is the first demonstration of fine-tuning diffusion policies for real-world robotic skills using an offline world model. We make the code publicly available at https://diwa.cs.uni-freiburg.de.

Aude Billard, Alin Albu-Schaeffer, Michael Beetz et al.

AI technologies, including deep learning, large-language models have gone from one breakthrough to the other. As a result, we are witnessing growing excitement in robotics at the prospect of leveraging the potential of AI to tackle some of the outstanding barriers to the full deployment of robots in our daily lives. However, action and sensing in the physical world pose greater and different challenges than analysing data in isolation. As the development and application of AI in robotic products advances, it is important to reflect on which technologies, among the vast array of network architectures and learning models now available in the AI field, are most likely to be successfully applied to robots; how they can be adapted to specific robot designs, tasks, environments; which challenges must be overcome. This article offers an assessment of what AI for robotics has achieved since the 1990s and proposes a short- and medium-term research roadmap listing challenges and promises. These range from keeping up-to-date large datasets, representatives of a diversity of tasks robots may have to perform, and of environments they may encounter, to designing AI algorithms tailored specifically to robotics problems but generic enough to apply to a wide range of applications and transfer easily to a variety of robotic platforms. For robots to collaborate effectively with humans, they must predict human behavior without relying on bias-based profiling. Explainability and transparency in AI-driven robot control are not optional but essential for building trust, preventing misuse, and attributing responsibility in accidents. We close on what we view as the primary long-term challenges, that is, to design robots capable of lifelong learning, while guaranteeing safe deployment and usage, and sustainable computational costs.

Shengchao Yan, Lukas König, Wolfram Burgard

Active traffic management with autonomous vehicles offers the potential for reduced congestion and improved traffic flow. However, developing effective algorithms for real-world scenarios requires overcoming challenges related to infinite-horizon traffic flow and partial observability. To address these issues and further decentralize traffic management, we propose an asymmetric actor-critic model that learns decentralized cooperative driving policies for autonomous vehicles using single-agent reinforcement learning. By employing attention neural networks with masking, our approach efficiently manages real-world traffic dynamics and partial observability, eliminating the need for predefined agents or agent-specific experience buffers in multi-agent reinforcement learning. Extensive evaluations across various traffic scenarios demonstrate our method's significant potential in improving traffic flow at critical bottleneck points. Moreover, we address the challenges posed by conservative autonomous vehicle driving behaviors that adhere strictly to traffic rules, showing that our cooperative policy effectively alleviates potential slowdowns without compromising safety.

Débora N. P. Oliveira, Joshua Knights, Sebastián Barbas Laina et al.

Loop closures are essential for correcting odometry drift and creating consistent maps, especially in the context of large-scale navigation. Current methods using dense point clouds for accurate place recognition do not scale well due to computationally expensive scan-to-scan comparisons. Alternative object-centric approaches are more efficient but often struggle with sensitivity to viewpoint variation. In this work, we introduce REGRACE, a novel approach that addresses these challenges of scalability and perspective difference in re-localization by using LiDAR-based submaps. We introduce rotation-invariant features for each labeled object and enhance them with neighborhood context through a graph neural network. To identify potential revisits, we employ a scalable bag-of-words approach, pooling one learned global feature per submap. Additionally, we define a revisit with geometrical consistency cues rather than embedding distance, allowing us to recognize far-away loop closures. Our evaluations demonstrate that REGRACE achieves similar results compared to state-of-the-art place recognition and registration baselines while being twice as fast. Code and models are publicly available.

Ahmet Selim Çanakçı, Niclas Vödisch, Kürsat Petek et al.

A main bottleneck of learning-based robotic scene understanding methods is the heavy reliance on extensive annotated training data, which often limits their generalization ability. In LiDAR panoptic segmentation, this challenge becomes even more pronounced due to the need to simultaneously address both semantic and instance segmentation from complex, high-dimensional point cloud data. In this work, we address the challenge of LiDAR panoptic segmentation with very few labeled samples by leveraging recent advances in label-efficient vision panoptic segmentation. To this end, we propose a novel method, Limited-Label LiDAR Panoptic Segmentation (L3PS), which requires only a minimal amount of labeled data. Our approach first utilizes a label-efficient 2D network to generate panoptic pseudo-labels from a small set of annotated images, which are subsequently projected onto point clouds. We then introduce a novel 3D refinement module that capitalizes on the geometric properties of point clouds. By incorporating clustering techniques, sequential scan accumulation, and ground point separation, this module significantly enhances the accuracy of the pseudo-labels, improving segmentation quality by up to +10.6 PQ and +7.9 mIoU. We demonstrate that these refined pseudo-labels can be used to effectively train off-the-shelf LiDAR segmentation networks. Through extensive experiments, we show that L3PS not only outperforms existing methods but also substantially reduces the annotation burden. We release the code of our work at https://l3ps.cs.uni-freiburg.de.

Chenguang Huang, Shengchao Yan, Wolfram Burgard

Dynamic scene understanding remains a persistent challenge in robotic applications. Early dynamic mapping methods focused on mitigating the negative influence of short-term dynamic objects on camera motion estimation by masking or tracking specific categories, which often fall short in adapting to long-term scene changes. Recent efforts address object association in long-term dynamic environments using neural networks trained on synthetic datasets, but they still rely on predefined object shapes and categories. Other methods incorporate visual, geometric, or semantic heuristics for the association but often lack robustness. In this work, we introduce BYE, a class-agnostic, per-scene point cloud encoder that removes the need for predefined categories, shape priors, or extensive association datasets. Trained on only a single sequence of exploration data, BYE can efficiently perform object association in dynamically changing scenes. We further propose an ensembling scheme combining the semantic strengths of Vision Language Models (VLMs) with the scene-specific expertise of BYE, achieving a 7% improvement and a 95% success rate in object association tasks. Code and dataset are available at https://byencoder.github.io.

Johannes Meyer, Jasper Hoffmann, Felix Schulz et al.

Semantic segmentation of 3D point cloud data often comes with high annotation costs. Active learning automates the process of selecting which data to annotate, reducing the total amount of annotation needed to achieve satisfactory performance. Recent approaches to active learning for 3D point clouds are often based on sophisticated heuristics for both, splitting point clouds into annotatable regions and selecting the most beneficial for further neural network training. In this work, we propose a novel and easy-to-implement strategy to separate the point cloud into annotatable regions. In our approach, we utilize a 2D grid to subdivide the point cloud into columns. To identify the next data to be annotated, we employ a network ensemble to estimate the uncertainty in the network output. We evaluate our method on the S3DIS dataset, the Toronto-3D dataset, and a large-scale urban 3D point cloud of the city of Freiburg, which we labeled in parts manually. The extensive evaluation shows that our method yields performance on par with, or even better than, complex state-of-the-art methods on all datasets. Furthermore, we provide results suggesting that in the context of point clouds the annotated area can be a more meaningful measure for active learning algorithms than the number of annotated points.

Noah Steinkrüger, Nisarga Nilavadi, Wolfram Burgard et al.

Object transportation in cluttered environments is a fundamental task in various domains, including domestic service and warehouse logistics. In cooperative object transport, multiple robots must coordinate to move objects that are too large for a single robot. One transport strategy is pushing, which only requires simple robots. However, careful selection of robot-object contact points is necessary to push the object along a preplanned path. Although this selection can be solved analytically, the solution space grows combinatorially with the number of robots and object size, limiting scalability. Inspired by how humans rely on common-sense reasoning for cooperative transport, we propose combining the reasoning capabilities of Large Language Models with local search to select suitable contact points. Our LLM-guided local search method for contact point selection, ConPoSe, successfully selects contact points for a variety of shapes, including cuboids, cylinders, and T-shapes. We demonstrate that ConPoSe scales better with the number of robots and object size than the analytical approach, and also outperforms pure LLM-based selection.