Kostas Bekris

Shiyang Lu, Haonan Chang, Eric Pu Jing et al.

This work presents OVIR-3D, a straightforward yet effective method for open-vocabulary 3D object instance retrieval without using any 3D data for training. Given a language query, the proposed method is able to return a ranked set of 3D object instance segments based on the feature similarity of the instance and the text query. This is achieved by a multi-view fusion of text-aligned 2D region proposals into 3D space, where the 2D region proposal network could leverage 2D datasets, which are more accessible and typically larger than 3D datasets. The proposed fusion process is efficient as it can be performed in real-time for most indoor 3D scenes and does not require additional training in 3D space. Experiments on public datasets and a real robot show the effectiveness of the method and its potential for applications in robot navigation and manipulation.

Haonan Chang, Kowndinya Boyalakuntla, Shiyang Lu et al.

We present an Open-Vocabulary 3D Scene Graph (OVSG), a formal framework for grounding a variety of entities, such as object instances, agents, and regions, with free-form text-based queries. Unlike conventional semantic-based object localization approaches, our system facilitates context-aware entity localization, allowing for queries such as ``pick up a cup on a kitchen table" or ``navigate to a sofa on which someone is sitting". In contrast to existing research on 3D scene graphs, OVSG supports free-form text input and open-vocabulary querying. Through a series of comparative experiments using the ScanNet dataset and a self-collected dataset, we demonstrate that our proposed approach significantly surpasses the performance of previous semantic-based localization techniques. Moreover, we highlight the practical application of OVSG in real-world robot navigation and manipulation experiments.

Rui Wang, Kai Gao, Jingjin Yu et al.

Object rearrangement is important for many applications but remains challenging, especially in confined spaces, such as shelves, where objects cannot be accessed from above and they block reachability to each other. Such constraints require many motion planning and collision checking calls, which are computationally expensive. In addition, the arrangement space grows exponentially with the number of objects. To address these issues, this work introduces a lazy evaluation framework with a local monotone solver and a global planner. Monotone instances are those that can be solved by moving each object at most once. A key insight is that reachability constraints at the grasps for objects' starts and goals can quickly reveal dependencies between objects without having to execute expensive motion planning queries. Given that, the local solver builds lazily a search tree that respects these reachability constraints without verifying that the arm paths are collision free. It only collision checks when a promising solution is found. If a monotone solution is not found, the non-monotone planner loads the lazy search tree and explores ways to move objects to intermediate locations from where monotone solutions to the goal can be found. Results show that the proposed framework can solve difficult instances in confined spaces with up to 16 objects, which state-of-the-art methods fail to solve. It also solves problems faster than alternatives, when the alternatives find a solution. It also achieves high-quality solutions, i.e., only 1.8 additional actions on average are needed for non-monotone instances.

Junchi Liang, Bowen Wen, Kostas Bekris et al.

This work aims to learn how to perform complex robot manipulation tasks that are composed of several, consecutively executed low-level sub-tasks, given as input a few visual demonstrations of the tasks performed by a person. The sub-tasks consist of moving the robot's end-effector until it reaches a sub-goal region in the task space, performing an action, and triggering the next sub-task when a pre-condition is met. Most prior work in this domain has been concerned with learning only low-level tasks, such as hitting a ball or reaching an object and grasping it. This paper describes a new neural network-based framework for learning simultaneously low-level policies as well as high-level policies, such as deciding which object to pick next or where to place it relative to other objects in the scene. A key feature of the proposed approach is that the policies are learned directly from raw videos of task demonstrations, without any manual annotation or post-processing of the data. Empirical results on object manipulation tasks with a robotic arm show that the proposed network can efficiently learn from real visual demonstrations to perform the tasks, and outperforms popular imitation learning algorithms.

Kristin J. Dana, Clinton Andrews, Kostas Bekris et al.

Emerging applications of robotics, and concerns about their impact, require the research community to put human-centric objectives front-and-center. To meet this challenge, we advocate an interdisciplinary approach, socially cognizant robotics, which synthesizes technical and social science methods. We argue that this approach follows from the need to empower stakeholder participation (from synchronous human feedback to asynchronous societal assessment) in shaping AI-driven robot behavior at all levels, and leads to a range of novel research perspectives and problems both for improving robots' interactions with individuals and impacts on society. Drawing on these arguments, we develop best practices for socially cognizant robot design that balance traditional technology-based metrics (e.g. efficiency, precision and accuracy) with critically important, albeit challenging to measure, human and society-based metrics.

Shuai Li, Azarakhsh Keipour, Kevin Jamieson et al.

Automating warehouse operations can reduce logistics overhead costs, ultimately driving down the final price for consumers, increasing the speed of delivery, and enhancing the resiliency to market fluctuations. This extended abstract showcases a large-scale package manipulation from unstructured piles in Amazon Robotics' Robot Induction (Robin) fleet, which is used for picking and singulating up to 6 million packages per day and so far has manipulated over 2 billion packages. It describes the various heuristic methods developed over time and their successor, which utilizes a pick success predictor trained on real production data. To the best of the authors' knowledge, this work is the first large-scale deployment of learned pick quality estimation methods in a real production system.

Aggelos Psiris, Vasileios Argyriou, Evangelos K. Markakis et al.

Over the recent years, the field of robotics has been undergoing a transformative paradigm shift from fixed, single-task, domain-specific solutions towards adaptive, multi-function, general-purpose agents, capable of operating in complex, open-world, and dynamic environments. This tremendous advancement is primarily driven by the emergence of Foundation Models (FMs), i.e., large-scale neural-network architectures trained on massive, heterogeneous datasets that provide unprecedented capabilities in multi-modal understanding and reasoning, long-horizon planning, and cross-embodiment generalization. In this context, the current study provides a holistic, systematic, and in-depth review of the research landscape of FMs in robotics. In particular, the evolution of the field is initially delineated through five distinct research phases, spanning from the early incorporation of Natural Language Processing (NLP) and Computer Vision (CV) models to the current frontier of multi-sensory generalization and real-world deployment. Subsequently, a highly-granular taxonomic investigation of the literature is performed, examining the following key aspects: a) the employed FM types, including LLMs, VFMs, VLMs, and VLAs, b) the underlying neural-network architectures, c) the adopted learning paradigms, d) the different learning stages of knowledge incorporation, e) the major robotic tasks, and f) the main real-world application domains. For each aspect, comparative analysis and critical insights are provided. Moreover, a report on the publicly available datasets used for model training and evaluation across the considered robotic tasks is included. Furthermore, a hierarchical discussion on the current open challenges and promising future research directions in the field is incorporated.

Shiyang Lu, Yunfu Deng, Abdeslam Boularias et al.

This work proposes a self-supervised learning system for segmenting rigid objects in RGB images. The proposed pipeline is trained on unlabeled RGB-D videos of static objects, which can be captured with a camera carried by a mobile robot. A key feature of the self-supervised training process is a graph-matching algorithm that operates on the over-segmentation output of the point cloud that is reconstructed from each video. The graph matching, along with point cloud registration, is able to find reoccurring object patterns across videos and combine them into 3D object pseudo labels, even under occlusions or different viewing angles. Projected 2D object masks from 3D pseudo labels are used to train a pixel-wise feature extractor through contrastive learning. During online inference, a clustering method uses the learned features to cluster foreground pixels into object segments. Experiments highlight the method's effectiveness on both real and synthetic video datasets, which include cluttered scenes of tabletop objects. The proposed method outperforms existing unsupervised methods for object segmentation by a large margin.

Shiyang Lu, William R. Johnson, Kun Wang et al.

Tensegrity robots, which are composed of compressive elements (rods) and flexible tensile elements (e.g., cables), have a variety of advantages, including flexibility, low weight, and resistance to mechanical impact. Nevertheless, the hybrid soft-rigid nature of these robots also complicates the ability to localize and track their state. This work aims to address what has been recognized as a grand challenge in this domain, i.e., the state estimation of tensegrity robots through a markerless, vision-based method, as well as novel, onboard sensors that can measure the length of the robot's cables. In particular, an iterative optimization process is proposed to track the 6-DoF pose of each rigid element of a tensegrity robot from an RGB-D video as well as endcap distance measurements from the cable sensors. To ensure that the pose estimates of rigid elements are physically feasible, i.e., they are not resulting in collisions between rods or with the environment, physical constraints are introduced during the optimization. Real-world experiments are performed with a 3-bar tensegrity robot, which performs locomotion gaits. Given ground truth data from a motion capture system, the proposed method achieves less than 1~cm translation error and 3 degrees rotation error, which significantly outperforms alternatives. At the same time, the approach can provide accurate pose estimation throughout the robot's motion, while motion capture often fails due to occlusions.

Bowen Wen, Kostas Bekris

Tracking the 6D pose of objects in video sequences is important for robot manipulation. Most prior efforts, however, often assume that the target object's CAD model, at least at a category-level, is available for offline training or during online template matching. This work proposes BundleTrack, a general framework for 6D pose tracking of novel objects, which does not depend upon 3D models, either at the instance or category-level. It leverages the complementary attributes of recent advances in deep learning for segmentation and robust feature extraction, as well as memory-augmented pose graph optimization for spatiotemporal consistency. This enables long-term, low-drift tracking under various challenging scenarios, including significant occlusions and object motions. Comprehensive experiments given two public benchmarks demonstrate that the proposed approach significantly outperforms state-of-art, category-level 6D tracking or dynamic SLAM methods. When compared against state-of-art methods that rely on an object instance CAD model, comparable performance is achieved, despite the proposed method's reduced information requirements. An efficient implementation in CUDA provides a real-time performance of 10Hz for the entire framework. Code is available at: https://github.com/wenbowen123/BundleTrack

Bowen Wen, Chaitanya Mitash, Kostas Bekris

Tracking the 6D pose of objects in video sequences is important for robot manipulation. This work presents se(3)-TrackNet, a data-driven optimization approach for long term, 6D pose tracking. It aims to identify the optimal relative pose given the current RGB-D observation and a synthetic image conditioned on the previous best estimate and the object's model. The key contribution in this context is a novel neural network architecture, which appropriately disentangles the feature encoding to help reduce domain shift, and an effective 3D orientation representation via Lie Algebra. Consequently, even when the network is trained solely with synthetic data can work effectively over real images. Comprehensive experiments over multiple benchmarks show se(3)-TrackNet achieves consistently robust estimates and outperforms alternatives, even though they have been trained with real images. The approach runs in real time at 90.9Hz. Code, data and supplementary video for this project are available at https://github.com/wenbowen123/iros20-6d-pose-tracking

Edgar Granados, Patrick Meng, Charles Tang et al.

Tensegrity robots offer compliance and adaptability, but their nonlinear, and underconstrained dynamics make state estimation challenging. Reliable continuous-time estimation of all rigid links is crucial for closed-loop control, system identification, and machine learning; however, conventional methods often fall short. This paper proposes a two-stage approach for robust state or trajectory estimation (i.e., filtering or smoothing) of a cable-driven tensegrity robot. For online state estimation, this work introduces a factor-graph-based method, which fuses measurements from an RGB-D camera with on-board cable length sensors. To the best of the authors' knowledge, this is the first application of factor graphs in this domain. Factor graphs are a natural choice, as they exploit the robot's structural properties and provide effective sensor fusion solutions capable of handling nonlinearities in practice. Both the Mahalanobis distance-based clustering algorithm, used to handle noise, and the Chebyshev polynomial method, used to estimate the most probable velocities and intermediate states, are shown to perform well on simulated and real-world data, compared to an ICP-based algorithm. Results show that the approach provides high fidelity, continuous-time state and trajectory estimates for complex tensegrity robot motions.

Shuai Li, Azarakhsh Keipour, Kevin Jamieson et al.

Automating warehouse operations can reduce logistics overhead costs, ultimately driving down the final price for consumers, increasing the speed of delivery, and enhancing the resiliency to workforce fluctuations. The past few years have seen increased interest in automating such repeated tasks but mostly in controlled settings. Tasks such as picking objects from unstructured, cluttered piles have only recently become robust enough for large-scale deployment with minimal human intervention. This paper demonstrates a large-scale package manipulation from unstructured piles in Amazon Robotics' Robot Induction (Robin) fleet, which utilizes a pick success predictor trained on real production data. Specifically, the system was trained on over 394K picks. It is used for singulating up to 5 million packages per day and has manipulated over 200 million packages during this paper's evaluation period. The developed learned pick quality measure ranks various pick alternatives in real-time and prioritizes the most promising ones for execution. The pick success predictor aims to estimate from prior experience the success probability of a desired pick by the deployed industrial robotic arms in cluttered scenes containing deformable and rigid objects with partially known properties. It is a shallow machine learning model, which allows us to evaluate which features are most important for the prediction. An online pick ranker leverages the learned success predictor to prioritize the most promising picks for the robotic arm, which are then assessed for collision avoidance. This learned ranking process is demonstrated to overcome the limitations and outperform the performance of manually engineered and heuristic alternatives. To the best of the authors' knowledge, this paper presents the first large-scale deployment of learned pick quality estimation methods in a real production system.

Kun Wang, William R. Johnson, Shiyang Lu et al.

Tensegrity robots, composed of rigid rods and flexible cables, exhibit high strength-to-weight ratios and significant deformations, which enable them to navigate unstructured terrains and survive harsh impacts. They are hard to control, however, due to high dimensionality, complex dynamics, and a coupled architecture. Physics-based simulation is a promising avenue for developing locomotion policies that can be transferred to real robots. Nevertheless, modeling tensegrity robots is a complex task due to a substantial sim2real gap. To address this issue, this paper describes a Real2Sim2Real (R2S2R) strategy for tensegrity robots. This strategy is based on a differentiable physics engine that can be trained given limited data from a real robot. These data include offline measurements of physical properties, such as mass and geometry for various robot components, and the observation of a trajectory using a random control policy. With the data from the real robot, the engine can be iteratively refined and used to discover locomotion policies that are directly transferable to the real robot. Beyond the R2S2R pipeline, key contributions of this work include computing non-zero gradients at contact points, a loss function for matching tensegrity locomotion gaits, and a trajectory segmentation technique that avoids conflicts in gradient evaluation during training. Multiple iterations of the R2S2R process are demonstrated and evaluated on a real 3-bar tensegrity robot.

Kun Wang, Mridul Aanjaneya, Kostas Bekris

Tensegrity robots, composed of rigid rods and flexible cables, are difficult to accurately model and control given the presence of complex dynamics and high number of DoFs. Differentiable physics engines have been recently proposed as a data-driven approach for model identification of such complex robotic systems. These engines are often executed at a high-frequency to achieve accurate simulation. Ground truth trajectories for training differentiable engines, however, are not typically available at such high frequencies due to limitations of real-world sensors. The present work focuses on this frequency mismatch, which impacts the modeling accuracy. We proposed a recurrent structure for a differentiable physics engine of tensegrity robots, which can be trained effectively even with low-frequency trajectories. To train this new recurrent engine in a robust way, this work introduces relative to prior work: (i) a new implicit integration scheme, (ii) a progressive training pipeline, and (iii) a differentiable collision checker. A model of NASA's icosahedron SUPERballBot on MuJoCo is used as the ground truth system to collect training data. Simulated experiments show that once the recurrent differentiable engine has been trained given the low-frequency trajectories from MuJoCo, it is able to match the behavior of MuJoCo's system. The criterion for success is whether a locomotion strategy learned using the differentiable engine can be transferred back to the ground-truth system and result in a similar motion. Notably, the amount of ground truth data needed to train the differentiable engine, such that the policy is transferable to the ground truth system, is 1% of the data needed to train the policy directly on the ground-truth system.

Bowen Wen, Wenzhao Lian, Kostas Bekris et al.

Promising results have been achieved recently in category-level manipulation that generalizes across object instances. Nevertheless, it often requires expensive real-world data collection and manual specification of semantic keypoints for each object category and task. Additionally, coarse keypoint predictions and ignoring intermediate action sequences hinder adoption in complex manipulation tasks beyond pick-and-place. This work proposes a novel, category-level manipulation framework that leverages an object-centric, category-level representation and model-free 6 DoF motion tracking. The canonical object representation is learned solely in simulation and then used to parse a category-level, task trajectory from a single demonstration video. The demonstration is reprojected to a target trajectory tailored to a novel object via the canonical representation. During execution, the manipulation horizon is decomposed into longrange, collision-free motion and last-inch manipulation. For the latter part, a category-level behavior cloning (CatBC) method leverages motion tracking to perform closed-loop control. CatBC follows the target trajectory, projected from the demonstration and anchored to a dynamically selected category-level coordinate frame. The frame is automatically selected along the manipulation horizon by a local attention mechanism. This framework allows to teach different manipulation strategies by solely providing a single demonstration, without complicated manual programming. Extensive experiments demonstrate its efficacy in a range of challenging industrial tasks in highprecision assembly, which involve learning complex, long-horizon policies. The process exhibits robustness against uncertainty due to dynamics as well as generalization across object instances and scene configurations. The supplementary video is available at https://www.youtube.com/watch?v=WAr8ZY3mYyw

Andrew S. Morgan, Kaiyu Hang, Bowen Wen et al.

Constraining contacts to remain fixed on an object during manipulation limits the potential workspace size, as motion is subject to the hand's kinematic topology. Finger gaiting is one way to alleviate such restraints. It allows contacts to be freely broken and remade so as to operate on different manipulation manifolds. This capability, however, has traditionally been difficult or impossible to practically realize. A finger gaiting system must simultaneously plan for and control forces on the object while maintaining stability during contact switching. This work alleviates the traditional requirement by taking advantage of system compliance, allowing the hand to more easily switch contacts while maintaining a stable grasp. Our method achieves complete SO(3) finger gaiting control of grasped objects against gravity by developing a manipulation planner that operates via orthogonal safe modes of a compliant, underactuated hand absent of tactile sensors or joint encoders. During manipulation, a low-latency 6D pose object tracker provides feedback via vision, allowing the planner to update its plan online so as to adaptively recover from trajectory deviations. The efficacy of this method is showcased by manipulating both convex and non-convex objects on a real robot. Its robustness is evaluated via perturbation rejection and long trajectory goals. To the best of the authors' knowledge, this is the first work that has autonomously achieved full SO(3) control of objects within-hand via finger gaiting and without a support surface, elucidating a valuable step towards realizing true robot in-hand manipulation capabilities.

Shiyang Lu, Rui Wang, Yinglong Miao et al.

This work proposes a robotic pipeline for picking and constrained placement of objects without geometric shape priors. Compared to recent efforts developed for similar tasks, where every object was assumed to be novel, the proposed system recognizes previously manipulated objects and performs online model reconstruction and reuse. Over a lifelong manipulation process, the system keeps learning features of objects it has interacted with and updates their reconstructed models. Whenever an instance of a previously manipulated object reappears, the system aims to first recognize it and then register its previously reconstructed model given the current observation. This step greatly reduces object shape uncertainty allowing the system to even reason for parts of objects, which are currently not observable. This also results in better manipulation efficiency as it reduces the need for active perception of the target object during manipulation. To get a reusable reconstructed model, the proposed pipeline adopts: i) TSDF for object representation, and ii) a variant of the standard particle filter algorithm for pose estimation and tracking of the partial object model. Furthermore, an effective way to construct and maintain a dataset of manipulated objects is presented. A sequence of real-world manipulation experiments is performed. They show how future manipulation tasks become more effective and efficient by reusing reconstructed models of previously manipulated objects, which were generated during their prior manipulation, instead of treating objects as novel every time.

Bowen Wen, Wenzhao Lian, Kostas Bekris et al.

Task-relevant grasping is critical for industrial assembly, where downstream manipulation tasks constrain the set of valid grasps. Learning how to perform this task, however, is challenging, since task-relevant grasp labels are hard to define and annotate. There is also yet no consensus on proper representations for modeling or off-the-shelf tools for performing task-relevant grasps. This work proposes a framework to learn task-relevant grasping for industrial objects without the need of time-consuming real-world data collection or manual annotation. To achieve this, the entire framework is trained solely in simulation, including supervised training with synthetic label generation and self-supervised, hand-object interaction. In the context of this framework, this paper proposes a novel, object-centric canonical representation at the category level, which allows establishing dense correspondence across object instances and transferring task-relevant grasps to novel instances. Extensive experiments on task-relevant grasping of densely-cluttered industrial objects are conducted in both simulation and real-world setups, demonstrating the effectiveness of the proposed framework. Code and data are available at https://sites.google.com/view/catgrasp.

Andrew S. Morgan, Bowen Wen, Junchi Liang et al.

Highly constrained manipulation tasks continue to be challenging for autonomous robots as they require high levels of precision, typically less than 1mm, which is often incompatible with what can be achieved by traditional perception systems. This paper demonstrates that the combination of state-of-the-art object tracking with passively adaptive mechanical hardware can be leveraged to complete precision manipulation tasks with tight, industrially-relevant tolerances (0.25mm). The proposed control method closes the loop through vision by tracking the relative 6D pose of objects in the relevant workspace. It adjusts the control reference of both the compliant manipulator and the hand to complete object insertion tasks via within-hand manipulation. Contrary to previous efforts for insertion, our method does not require expensive force sensors, precision manipulators, or time-consuming, online learning, which is data hungry. Instead, this effort leverages mechanical compliance and utilizes an object agnostic manipulation model of the hand learned offline, off-the-shelf motion planning, and an RGBD-based object tracker trained solely with synthetic data. These features allow the proposed system to easily generalize and transfer to new tasks and environments. This paper describes in detail the system components and showcases its efficacy with extensive experiments involving tight tolerance peg-in-hole insertion tasks of various geometries as well as open-world constrained placement tasks.

Sebastian Höfer, Kostas Bekris, Ankur Handa et al.

This report presents the debates, posters, and discussions of the Sim2Real workshop held in conjunction with the 2020 edition of the "Robotics: Science and System" conference. Twelve leaders of the field took competing debate positions on the definition, viability, and importance of transferring skills from simulation to the real world in the context of robotics problems. The debaters also joined a large panel discussion, answering audience questions and outlining the future of Sim2Real in robotics. Furthermore, we invited extended abstracts to this workshop which are summarized in this report. Based on the workshop, this report concludes with directions for practitioners exploiting this technology and for researchers further exploring open problems in this area.

Kun Wang, Mridul Aanjaneya, Kostas Bekris

Learning policies in simulation is promising for reducing human effort when training robot controllers. This is especially true for soft robots that are more adaptive and safe but also more difficult to accurately model and control. The sim2real gap is the main barrier to successfully transfer policies from simulation to a real robot. System identification can be applied to reduce this gap but traditional identification methods require a lot of manual tuning. Data-driven alternatives can tune dynamical models directly from data but are often data hungry, which also incorporates human effort in collecting data. This work proposes a data-driven, end-to-end differentiable simulator focused on the exciting but challenging domain of tensegrity robots. To the best of the authors' knowledge, this is the first differentiable physics engine for tensegrity robots that supports cable, contact, and actuation modeling. The aim is to develop a reasonably simplified, data-driven simulation, which can learn approximate dynamics with limited ground truth data. The dynamics must be accurate enough to generate policies that can be transferred back to the ground-truth system. As a first step in this direction, the current work demonstrates sim2sim transfer, where the unknown physical model of MuJoCo acts as a ground truth system. Two different tensegrity robots are used for evaluation and learning of locomotion policies, a 6-bar and a 3-bar tensegrity. The results indicate that only 0.25\% of ground truth data are needed to train a policy that works on the ground truth system when the differentiable engine is used for training against training the policy directly on the ground truth system.

Kun Wang, Mridul Aanjaneya, Kostas Bekris

We propose a novel differentiable physics engine for system identification of complex spring-rod assemblies. Unlike black-box data-driven methods for learning the evolution of a dynamical system \emph{and} its parameters, we modularize the design of our engine using a discrete form of the governing equations of motion, similar to a traditional physics engine. We further reduce the dimension from 3D to 1D for each module, which allows efficient learning of system parameters using linear regression. The regression parameters correspond to physical quantities, such as spring stiffness or the mass of the rod, making the pipeline explainable. The approach significantly reduces the amount of training data required, and also avoids iterative identification of data sampling and model training. We compare the performance of the proposed engine with previous solutions, and demonstrate its efficacy on tensegrity systems, such as NASA's icosahedron.

Chaitanya Mitash, Rahul Shome, Bowen Wen et al.

Recent progress in robotic manipulation has dealt with the case of previously unknown objects in the context of relatively simple tasks, such as bin-picking. Existing methods for more constrained problems, however, such as deliberate placement in a tight region, depend more critically on shape information to achieve safe execution. This work deals with pick-and-constrained placement of objects without access to geometric models. The objective is to pick an object and place it safely inside a desired goal region without any collisions, while minimizing the time and the sensing operations required to complete the task. An algorithmic framework is proposed for this purpose, which performs manipulation planning simultaneously over a conservative and an optimistic estimate of the object's volume. The conservative estimate ensures that the manipulation is safe while the optimistic estimate guides the sensor-based manipulation process when no solution can be found for the conservative estimate. To maintain these estimates and dynamically update them during manipulation, objects are represented by a simple volumetric representation, which stores sets of occupied and unseen voxels. The effectiveness of the proposed approach is demonstrated by developing a robotic system that picks a previously unseen object from a table-top and places it in a constrained space. The system comprises of a dual-arm manipulator with heterogeneous end-effectors and leverages hand-offs as a re-grasping strategy. Real-world experiments show that straightforward pick-sense-and-place alternatives frequently fail to solve pick-and-constrained placement problems. The proposed pipeline, however, achieves more than 95% success rate and faster execution times as evaluated over multiple physical experiments.

Kun Wang, Mridul Aanjaneya, Kostas Bekris

We propose a novel differentiable physics engine for system identification of complex spring-rod assemblies. Unlike black-box data-driven methods for learning the evolution of a dynamical system and its parameters, we modularize the design of our engine using a discrete form of the governing equations of motion, similar to a traditional physics engine. We further reduce the dimension from 3D to 1D for each module, which allows efficient learning of system parameters using linear regression. As a side benefit, the regression parameters correspond to physical quantities, such as spring stiffness or the mass of the rod, making the pipeline explainable. The approach significantly reduces the amount of training data required, and also avoids iterative identification of data sampling and model training. We compare the performance of the proposed engine with previous solutions, and demonstrate its efficacy on tensegrity systems, such as NASA's icosahedron.

Malihe Alikhani, Baber Khalid, Rahul Shome et al.

Collaborative robotics requires effective communication between a robot and a human partner. This work proposes a set of interpretive principles for how a robotic arm can use pointing actions to communicate task information to people by extending existing models from the related literature. These principles are evaluated through studies where English-speaking human subjects view animations of simulated robots instructing pick-and-place tasks. The evaluation distinguishes two classes of pointing actions that arise in pick-and-place tasks: referential pointing (identifying objects) and locating pointing (identifying locations). The study indicates that human subjects show greater flexibility in interpreting the intent of referential pointing compared to locating pointing, which needs to be more deliberate. The results also demonstrate the effects of variation in the environment and task context on the interpretation of pointing. Our corpus, experiments and design principles advance models of context, common sense reasoning and communication in embodied communication.

Chaitanya Mitash, Bowen Wen, Kostas Bekris et al.

This paper introduces key machine learning operations that allow the realization of robust, joint 6D pose estimation of multiple instances of objects either densely packed or in unstructured piles from RGB-D data. The first objective is to learn semantic and instance-boundary detectors without manual labeling. An adversarial training framework in conjunction with physics-based simulation is used to achieve detectors that behave similarly in synthetic and real data. Given the stochastic output of such detectors, candidates for object poses are sampled. The second objective is to automatically learn a single score for each pose candidate that represents its quality in terms of explaining the entire scene via a gradient boosted tree. The proposed method uses features derived from surface and boundary alignment between the observed scene and the object model placed at hypothesized poses. Scene-level, multi-instance pose estimation is then achieved by an integer linear programming process that selects hypotheses that maximize the sum of the learned individual scores, while respecting constraints, such as avoiding collisions. To evaluate this method, a dataset of densely packed objects with challenging setups for state-of-the-art approaches is collected. Experiments on this dataset and a public one show that the method significantly outperforms alternatives in terms of 6D pose accuracy while trained only with synthetic datasets.

Rahul Shome, Kiril Solovey, Jingjin Yu et al.

Rearranging objects on a planar surface arises in a variety of robotic applications, such as product packaging. Using two arms can improve efficiency but introduces new computational challenges. This paper studies the structure of dual-arm rearrangement for synchronous, monotone tabletop setups and develops an optimal mixed integer model. It then describes an efficient and scalable algorithm, which first minimizes the cost of object transfers and then of moves between objects. This is motivated by the fact that, asymptotically, object transfers dominate the cost of solutions. Moreover, a lazy strategy minimizes the number of motion planning calls and results in significant speedups. Theoretical arguments support the benefits of using two arms and indicate that synchronous execution, in which the two arms perform together either transfers or moves, introduces only a small overhead. Experiments support these points and show that the scalable method can quickly compute solutions close to the optimal for the considered setup.

Tomas Hodan, Rigas Kouskouridas, Tae-Kyun Kim et al.

This document summarizes the 4th International Workshop on Recovering 6D Object Pose which was organized in conjunction with ECCV 2018 in Munich. The workshop featured four invited talks, oral and poster presentations of accepted workshop papers, and an introduction of the BOP benchmark for 6D object pose estimation. The workshop was attended by 100+ people working on relevant topics in both academia and industry who shared up-to-date advances and discussed open problems.

Chaitanya Mitash, Abdeslam Boularias, Kostas Bekris

This paper focuses on vision-based pose estimation for multiple rigid objects placed in clutter, especially in cases involving occlusions and objects resting on each other. Progress has been achieved recently in object recognition given advancements in deep learning. Nevertheless, such tools typically require a large amount of training data and significant manual effort to label objects. This limits their applicability in robotics, where solutions must scale to a large number of objects and variety of conditions. Moreover, the combinatorial nature of the scenes that could arise from the placement of multiple objects is hard to capture in the training dataset. Thus, the learned models might not produce the desired level of precision required for tasks, such as robotic manipulation. This work proposes an autonomous process for pose estimation that spans from data generation to scene-level reasoning and self-learning. In particular, the proposed framework first generates a labeled dataset for training a Convolutional Neural Network (CNN) for object detection in clutter. These detections are used to guide a scene-level optimization process, which considers the interactions between the different objects present in the clutter to output pose estimates of high precision. Furthermore, confident estimates are used to label online real images from multiple views and re-train the process in a self-learning pipeline. Experimental results indicate that this process is quickly able to identify in cluttered scenes physically-consistent object poses that are more precise than the ones found by reasoning over individual instances of objects. Furthermore, the quality of pose estimates increases over time given the self-learning process.

Andrew Kimmel, Rahul Shome, Zakary Littlefield et al.

Many methods have been developed for planning the motion of robotic arms for picking and placing, ranging from local optimization to global search techniques, which are effective for sparsely placed objects. Dense clutter, however, still adversely affects the success rate, computation times, and quality of solutions in many real-world setups. The current work integrates tools from existing methodologies and proposes a framework that achieves high success ratio in clutter with anytime performance. The idea is to first explore the lower dimensional end effector's task space efficiently by ignoring the arm, and build a discrete approximation of a navigation function, which guides the end effector towards the set of available grasps or object placements. This is performed online, without prior knowledge of the scene. Then, an informed sampling-based planner for the entire arm uses Jacobian-based steering to reach promising end effector poses given the task space guidance. While informed, the method is also comprehensive and allows the exploration of alternative paths over time if the task space guidance does not lead to a solution. This paper evaluates the proposed method against alternatives in picking or placing tasks among varying amounts of clutter for a variety of robotic manipulators with different end-effectors. The results suggest that the method reliably provides higher quality solution paths quicker, with a higher success rate relative to alternatives.

Chaitanya Mitash, Abdeslam Boularias, Kostas Bekris

Object pose estimation is frequently achieved by first segmenting an RGB image and then, given depth data, registering the corresponding point cloud segment against the object's 3D model. Despite the progress due to CNNs, semantic segmentation output can be noisy, especially when the CNN is only trained on synthetic data. This causes registration methods to fail in estimating a good object pose. This work proposes a novel stochastic optimization process that treats the segmentation output of CNNs as a confidence probability. The algorithm, called Stochastic Congruent Sets (StoCS), samples pointsets on the point cloud according to the soft segmentation distribution and so as to agree with the object's known geometry. The pointsets are then matched to congruent sets on the 3D object model to generate pose estimates. StoCS is shown to be robust on an APC dataset, despite the fact the CNN is trained only on synthetic data. In the YCB dataset, StoCS outperforms a recent network for 6D pose estimation and alternative pointset matching techniques.

Athanasios Krontiris, Rahul Shome, Andrew Dobson et al.

This work proposes a method for effectively computing manipulation paths to rearrange similar objects in a cluttered space. The solution can be used to place similar products in a factory floor in a desirable arrangement or for retrieving a particular object from a shelf blocked by similarly sized objects. These are challenging problems as they involve combinatorially large, continuous configuration spaces due to the presence of multiple moving bodies and kinematically complex manipulators. This work leverages ideas from algorithmic theory, multi-robot motion planning and manipulation planning to propose appropriate graphical representations for this challenge. These representations allow to quickly reason whether manipulation paths allow the transition between entire sets of objects arrangements without having to explicitly enumerate the path for each pair of arrangements. The proposed method also allows to take advantage of precomputation given a manipulation roadmap for transferring a single object in the same cluttered space. The resulting approach is evaluated in simulation for a realistic model of a Baxter robot and executed in open-loop on the real system, showing that the approach solves complex instances and is promising in terms of scalability and success ratio.