We introduce a novel approach that combines tactile estimation and control for in-hand object manipulation. By integrating measurements from robot kinematics and an image-based tactile sensor, our framework estimates and tracks object pose while simultaneously generating motion plans in a receding horizon fashion to control the pose of a grasped object. This approach consists of a discrete pose estimator that tracks the most likely sequence of object poses in a coarsely discretized grid, and a continuous pose estimator-controller to refine the pose estimate and accurately manipulate the pose of the grasped object. Our method is tested on diverse objects and configurations, achieving desired manipulation objectives and outperforming single-shot methods in estimation accuracy. The proposed approach holds potential for tasks requiring precise manipulation and limited intrinsic in-hand dexterity under visual occlusion, laying the foundation for closed-loop behavior in applications such as regrasping, insertion, and tool use. Please see supplementary multimedia for videos of real-world demonstrations.

We introduce a novel approach that combines tactile estimation and control for in-hand object manipulation. By integrating measurements from robot kinematics and an image-based tactile sensor, our framework estimates and tracks object pose while simultaneously generating motion plans in a receding horizon fashion to control the pose of a grasped object. This approach consists of a discrete pose estimator that tracks the most likely sequence of object poses in a coarsely discretized grid, and a continuous pose estimator-controller to refine the pose estimate and accurately manipulate the pose of the grasped object. Our method is tested on diverse objects and configurations, achieving desired manipulation objectives and outperforming single-shot methods in estimation accuracy. The proposed approach holds potential for tasks requiring precise manipulation and limited intrinsic in-hand dexterity under visual occlusion, laying the foundation for closed-loop behavior in applications such as regrasping, insertion, and tool use. Please see this url for videos of real-world demonstrations.

We propose a method that simultaneously estimates and controls extrinsic contact with tactile feedback. The method enables challenging manipulation tasks that require controlling light forces and accurate motions in contact, such as balancing an unknown object on a thin rod standing upright. A factor graph-based framework fuses a sequence of tactile and kinematic measurements to estimate and control the interaction between gripper-object-environment, including the location and wrench at the extrinsic contact between the grasped object and the environment and the grasp wrench transferred from the gripper to the object. The same framework simultaneously plans the gripper motions that make it possible to estimate the state while satisfying regularizing control objectives to prevent slip, such as minimizing the grasp wrench and minimizing frictional force at the extrinsic contact. We show results with sub-millimeter contact localization error and good slip prevention even on slippery environments, for multiple contact formations (point, line, patch contact) and transitions between them. See supplementary video and results at https://sites.google.com/view/sim-tact.

We propose a framework for optimizing a planar parallel-jaw gripper for use with multiple objects. While optimizing general-purpose grippers and contact locations for grasps are both well studied, co-optimizing grasps and the gripper geometry to execute them receives less attention. As such, our framework synthesizes grippers optimized to stably grasp sets of polygonal objects. Given a fixed number of contacts and their assignments to object faces and gripper jaws, our framework optimizes contact locations along these faces, gripper pose for each grasp, and gripper shape. Our key insights are to pose shape and contact constraints in frames fixed to the gripper jaws, and to leverage the linearity of constraints in our grasp stability and gripper shape models via an augmented Lagrangian formulation. Together, these enable a tractable nonlinear program implementation. We apply our method to several examples. The first illustrative problem shows the discovery of a geometrically simple solution where possible. In another, space is constrained, forcing multiple objects to be contacted by the same features as each other. Finally a toolset-grasping example shows that our framework applies to complex, real-world objects. We provide a physical experiment of the toolset grasps.

In this work, we build on our method for manipulating unknown objects via contact configuration regulation: the estimation and control of the location, geometry, and mode of all contacts between the robot, object, and environment. We further develop our estimator and controller to enable manipulation through more complex contact interactions, including intermittent contact between the robot/object, and multiple contacts between the object/environment. In addition, we support a larger set of contact geometries at each interface. This is accomplished through a factor graph based estimation framework that reasons about the complementary kinematic and wrench constraints of contact to predict the current contact configuration. We are aided by the incorporation of a limited amount of visual feedback; which when combined with the available F/T sensing and robot proprioception, allows us to differentiate contact modes that were previously indistinguishable. We implement this revamped framework on our manipulation platform, and demonstrate that it allows the robot to perform a wider set of manipulation tasks. This includes, using a wall as a support to re-orient an object, or regulating the contact geometry between the object and the ground. Finally, we conduct ablation studies to understand the contributions from visual and tactile feedback in our manipulation framework. Our code can be found at: https://github.com/mcubelab/pbal.

We present an approach to robotic manipulation of unknown objects through regulation of the object's contact configuration: the location, geometry, and mode of all contacts between the object, robot, and environment. A contact configu-ration constrains the forces and motions that can be applied to the object; however, synthesizing these constraints generally requires knowledge of the object's pose and geometry. We develop an object-agnostic approach for estimation and control that circumvents this need. Our framework directly estimates a set of wrench and motion constraints which it uses to regulate the contact configuration. We use this to reactively manipulate unknown planar objects in the gravity plane. A video describing our work can be found on our project page: http://mcube.mit.edu/research/contactConfig.html.

We propose a method that actively estimates contact location between a grasped rigid object and its environment and uses this as input to a peg-in-hole insertion policy. An estimation model and an active tactile feedback controller work collaboratively to estimate the external contacts accurately. The controller helps the estimation model get a better estimate by regulating a consistent contact mode. The better estimation makes it easier for the controller to regulate the contact. We then train an object-agnostic insertion policy that learns to use the series of contact estimates to guide the insertion of an unseen peg into a hole. In contrast with previous works that learn a policy directly from tactile signals, since this policy is in contact configuration space, it can be learned directly in simulation. Lastly, we demonstrate and evaluate the active extrinsic contact line estimation and the trained insertion policy together in a real experiment. We show that the proposed method inserts various-shaped test objects with higher success rates and fewer insertion attempts than previous work with end-to-end approaches. See supplementary video and results at https://sites.google.com/view/active-extrinsic-contact.

Thin, reflective objects such as forks and whisks are common in our daily lives, but they are particularly chal-lenging for robot perception because it is hard to reconstruct them using commodity RGB-D cameras or multi-view stereo techniques. While traditional pipelines struggle with objects like these, Neural Radiance Fields (NeRFs) have recently been shown to be remarkably effective for performing view synthesis on objects with thin structures or reflective materials. In this paper we explore the use of NeRF as a new source of supervision for robust robot vision systems. In particular, we demonstrate that a NeRF representation of a scene can be used to train dense object descriptors. We use an optimized NeRF to extract dense correspondences between multiple views of an object, and then use these correspondences as training data for learning a view-invariant representation of the object. NeRF's usage of a density field allows us to reformulate the correspondence problem with a novel distribution-of-depths formulation, as opposed to the conventional approach of using a depth map. Dense correspondence models supervised with our method significantly outperform off-the-shelf learned descriptors by 106% (PCK@3px metric, more than doubling performance) and outperform our baseline supervised with multi-view stereo by 29%. Furthermore, we demonstrate the learned dense descriptors enable robots to perform accurate 6-degree of freedom (6-DoF) pick and place of thin and reflective objects.

We present Neural Descriptor Fields (NDFs), an object representation that encodes both points and relative poses between an object and a target (such as a robot gripper or a rack used for hanging) via category-level descriptors. We employ this representation for object manipulation, where given a task demonstration, we want to repeat the same task on a new object instance from the same category. We propose to achieve this objective by searching (via optimization) for the pose whose descriptor matches that observed in the demonstration. NDFs are conveniently trained in a self-supervised fashion via a 3D auto-encoding task that does not rely on expert-labeled keypoints. Further, NDFs are SE(3)-equivariant, guaranteeing performance that generalizes across all possible 3D object translations and rotations. We demonstrate learning of manipulation tasks from few (∼5-10) demonstrations both in simulation and on a real robot. Our performance generalizes across both object instances and 6-DoF object poses, and significantly outperforms a recent baseline that relies on 2D descriptors. Project website: https://yilundu.github.io/ndf/

This work presents a new version of tactile-sensing finger, GelSlim 3.0, which integrates the ability to sense high-resolution shape, force, and slip in a more compact form factor than previous implementations, designed for cluttered bin-picking scenarios. The novel design integrates real-time model-based algorithms to measure shape, estimate the 3-D contact force distribution, and detect incipient slip. The constraints imposed by the photometric stereo algorithm used for depth reconstruction and the implementation of a planar sensing surface make the miniaturization of previous designs nontrivial. To achieve a compact integration, we optimize the optical path from illumination source to camera. Using an optical simulation environment, we develop an illumination shaping lens and position the source LEDs and camera. The optimized optical configuration is integrated into a finger design composed of a robust and easily replaceable snap-to-fit fingertip module that facilitates manufacture, assembly, use, and repair. To stimulate future research in tactile-sensing and provide the robotics community access to a reliable and easily reproducible tactile finger with a diversity of sensing modalities, we open-source the design, fabrication methods, and software at https://github.com/mcubelab/gelslim.

Given an object, an environment, and a goal pose, how should a robot make contact to move it? Solving this problem requires reasoning about rigid-body dynamics, object and environment geometries, and hybrid contact mechanics. This paper proposes a hierarchical framework that solves this problem in 2D worlds, with polygonal objects and point fingers. To achieve this, we decouple the problem in three stages: 1) a high-level graph search over regions of free-space, 2) a medium-level randomized motion planner for the object motion, and 3) a low-level contact-trajectory optimization for the robot and environment contacts. In contrast to the state of the art, this approach does not rely on handcrafted primitives and can still be solved efficiently. This algorithm does not require seeding and can be applied to complex object shapes and environments. We validate this framework with extensive simulated experiments showcasing long-horizon and contact-rich interactions. We demonstrate how our algorithm can reliably solve complex planar manipulation problems in the order of seconds.

We propose a framework for co-optimizing the shape and motion of rigid robotic effectors for planar tasks. While planning object and robot-object contact trajectories is extensively studied, designing an effector that can execute the planned trajectories receives less attention. As such, our framework synthesizes an object trajectory and object-effector contact trajectory into an effector trajectory and shape that (a) does not penetrate the object, (b) makes contact with the object as specified, and (c) optimizes a user-specified objective. This simplifies manipulator control by encoding task-specific contact information in the effector's geometry. Our key insight is posing these requirements as constraints in the effector's reference frame, preventing the need for explicit parameterization of the effector shape. This prevents artificial restrictions on the shape design space. Importantly, it also facilitates posing the shape and motion design problem as a tractable nonlinear program. Our method is particularly useful for problems where the shape of the effector surface must be precisely chosen to achieve a task. We apply our method to several such problems, including jar-opening and picking up objects in constrained spaces. We evaluate the performance and computational cost of our method, and provide a physical experiment of a robotic arm picking up a screwdriver from a table with a designed tool.

We present the design, implementation, and evaluation of RF-Grasp, a robotic system that can grasp fully-occluded objects in unknown and unstructured environments. Unlike prior systems that are constrained by the line-of-sight perception of vision and infrared sensors, RF-Grasp employs RF (Radio Frequency) perception to identify and locate target objects through occlusions, and perform efficient exploration and complex manipulation tasks in non-line-of-sight settings.RF-Grasp relies on an eye-in-hand camera and batteryless RFID tags attached to objects of interest. It introduces two main innovations: (1) an RF-visual servoing controller that uses the RFID’s location to selectively explore the environment and plan an efficient trajectory toward an occluded target, and (2) an RF-visual deep reinforcement learning network that can learn and execute efficient, complex policies for decluttering and grasping.We implemented and evaluated an end-to-end physical prototype of RF-Grasp. We demonstrate it improves success rate and efficiency by up to 40-50% over a state-of-the-art baseline. We also demonstrate RF-Grasp in novel tasks such mechanical search of fully-occluded objects behind obstacles, opening up new possibilities for robotic manipulation. Qualitative results (videos) available at rfgrasp.media.mit.edu

Object insertion is a classic contact-rich manipulation task. The task remains challenging, especially when considering general objects of unknown geometry, which significantly limits the ability to understand the contact configuration between the object and the environment. We study the problem of aligning the object and environment with a tactile-based feedback insertion policy. The insertion process is modeled as an episodic policy that iterates between insertion attempts followed by pose corrections. We explore different mechanisms to learn such a policy based on Reinforcement Learning. The key contribution of this paper is to demonstrate that it is possible to learn a tactile insertion policy that generalizes across different object geometries, and an ablation study of the key design choices for the learning agent: 1) the type of learning scheme: supervised vs. reinforcement learning; 2) the type of learning schedule: unguided vs. curriculum learning ; 3) the type of sensing modality: force/torque vs. tactile; and 4) the type of tactile representation: tactile RGB vs. tactile flow. We show that the optimal configuration of the learning agent (RL + curriculum + tactile flow) exposed to 4 training objects yields an closed-loop insertion policy that inserts 4 novel objects with over 85.0% success rate and within 3~4 consecutive attempts. Comparisons between F/T and tactile sensing, shows that while an F/T-based policy learns more efficiently, a tactile-based policy provides better generalization. See supplementary video and results at https://sites.google.com/view/tactileinsertion.

Multi-stage forceful manipulation tasks, such as twisting a nut on a bolt, require reasoning over interlocking constraints over discrete and continuous choices. The robot must choose a sequence of discrete actions, or strategy, such as whether to pick up an object, and the continuous parameters of each of those actions, such as how to grasp that object. In forceful manipulation tasks, the force requirements substantially impact the choices of both strategy and parameters. To enable planning and executing forceful manipulation, we augment an existing task and motion planner with controllers that exert wrenches and constraints that explicitly consider torque and frictional limits. In two domains, opening a childproof bottle and twisting a nut, we demonstrate how the system considers a combinatorial number of strategies and how choosing actions that are robust to parameter variations impacts the choice of strategy. https://mcube.mit.edu/forceful-manipulation/

This paper addresses the localization of contacts of an unknown grasped rigid object with its environment, i.e., extrinsic to the robot. We explore the key role that distributed tactile sensing plays in localizing contacts external to the robot, in contrast to the role that aggregated force/torque measurements traditionally play in localizing contacts on the robot. When in contact with the environment, an object will move in accordance with the kinematic and possibly frictional constraints imposed by that contact. Small motions of the object, which are observable with tactile sensors, indirectly encode those constraints and the geometry that defines them.We formulate the extrinsic contact sensing problem as a constraint-based estimation problem. The estimation is subject to the kinematic constraints imposed by the tactile measurements of object motion, as well as the kinematic (e.g., non-penetration) and possibly frictional (e.g., sticking) constraints imposed by rigid-body mechanics. We validate the approach in simulation and with real experiments on the case studies of fixed point and line contacts.This paper discusses the theoretical basis for the value of distributed tactile sensing in contrast to aggregated force/torque measurements. It also provides an estimation framework for localizing environmental contacts with potential impact in contact-rich manipulation scenarios such as assembling or packing.

Tactile perception is central to robot manipulation in unstructured environments. However, it requires contact, and a mature implementation must infer object models while also accounting for the motion induced by the interaction. In this work, we present a method to estimate both object shape and pose in real-time from a stream of tactile measurements. This is applied towards tactile exploration of an unknown object by planar pushing. We consider this as an online SLAM problem with a nonparametric shape representation. Our formulation of tactile inference alternates between Gaussian process implicit surface regression and pose estimation on a factor graph. Through a combination of local Gaussian processes and fixed-lag smoothing, we infer object shape and pose in real-time. We evaluate our system across different objects in both simulated and real-world planar pushing tasks.

We present iNeRF, a framework that performs mesh-free pose estimation by "inverting" a Neural Radiance Field (NeRF). NeRFs have been shown to be remarkably effective for the task of view synthesis — synthesizing photorealistic novel views of real-world scenes or objects. In this work, we investigate whether we can apply analysis-by-synthesis via NeRF for mesh-free, RGB-only 6DoF pose estimation – given an image, find the translation and rotation of a camera relative to a 3D object or scene. Our method assumes that no object mesh models are available during either training or test time. Starting from an initial pose estimate, we use gradient descent to minimize the residual between pixels rendered from a NeRF and pixels in an observed image. In our experiments, we first study 1) how to sample rays during pose refinement for iNeRF to collect informative gradients and 2) how different batch sizes of rays affect iNeRF on a synthetic dataset. We then show that for complex real-world scenes from the LLFF dataset [21], iNeRF can improve NeRF by estimating the camera poses of novel images and using these images as additional training data for NeRF. Finally, we show iNeRF can perform categorylevel object pose estimation, including object instances not seen during training, with RGB images by inverting a NeRF model inferred from a single view.

We present the design, implementation, and evaluation of RFusion, a robotic system that can search for and retrieve RFID-tagged items in line-of-sight, non-line-of-sight, and fully-occluded settings. RFusion consists of a robotic arm that has a camera and antenna strapped around its gripper. Our design introduces two key innovations: the first is a method that geometrically fuses RF and visual information to reduce uncertainty about the target object's location, even when the item is fully occluded. The second is a novel reinforcement-learning network that uses the fused RF-visual information to efficiently localize, maneuver toward, and grasp target items. We built an end-to-end prototype of RFusion and tested it in challenging real-world environments. Our evaluation demonstrates that RFusion localizes target items with centimeter-scale accuracy and achieves 96% success rate in retrieving fully occluded objects, even if they are under a pile. The system paves the way for novel robotic retrieval tasks in complex environments such as warehouses, manufacturing plants, and smart homes.

We present a hybrid differential dynamic programming (DDP) algorithm for closed-loop execution of manipulation primitives with frictional contact switches. Planning and control of these primitives is challenging as they are hybrid, under-actuated, and stochastic. We address this by developing hybrid DDP both to plan finite horizon trajectories with a few contact switches and to create linear stabilizing controllers. We evaluate the performance and computational cost of our framework in ablations studies for two primitives: planar pushing and planar pivoting. We find that generating pose-to-pose closed-loop trajectories from most configurations requires only a couple (one to two) hybrid switches and can be done in reasonable time (one to five seconds). We further demonstrate that our controller stabilizes these hybrid trajectories on a real pushing system. A video describing our work can be found at https://youtu.be/YGSe4cUfq6Q.

The ability to simulate and predict the outcome of contacts is paramount to the successful execution of many robotic tasks. Simulators are powerful tools for the design of robots and their behaviors, yet the discrepancy between their predictions and observed data limit their usability. In this paper, we propose a self-supervised approach to learning residual models for rigid-body simulators that exploits corrections of contact models to refine predictive performance and propagate uncertainty. We empirically evaluate the framework by predicting the outcomes of planar dice rolls and compare it's performance to state-of-the-art techniques.

This paper develops closed-loop tactile controllers for dexterous robotic manipulation with a dual-palm robotic system. Tactile dexterity is an approach to dexterous manipulation that plans for robot/object interactions that render interpretable tactile information for control. We divide the role of tactile control into two goals: 1) control the contact state between the end-effector and the object (contact/no-contact, stick/slip) by regulating the stability of planned contact configurations and monitoring undesired slip events; and 2) control the object state by tactile-based tracking and iterative replanning of the object and robot trajectories. Key to this formulation is the decomposition of manipulation plans into sequences of manipulation primitives with simple mechanics and efficient planners. We consider the scenario of manipulating an object from an initial pose to a target pose on a flat surface while correcting for external perturbations and uncertainty in the initial pose of the object. We experimentally validate the approach with an ABB YuMi dual-arm robot and demonstrate the ability of the tactile controller to react to external perturbations.

Planning contact interactions is one of the core challenges of many robotic tasks. Optimizing contact locations while taking dynamics into account is computationally costly and, in environments that are only partially observable, executing contact-based tasks often suffers from low accuracy. We present an approach that addresses these two challenges for the problem of vision-based manipulation. First, we propose to disentangle contact from motion optimization. Thereby, we improve planning efficiency by focusing computation on promising contact locations. Second, we use a hybrid approach for perception and state estimation that combines neural networks with a physically meaningful state representation. In simulation and real-world experiments on the task of planar pushing, we show that our method is more efficient and achieves a higher manipulation accuracy than previous vision-based approaches.

We present the design of an active two-phase finger for mechanically mediated dexterous manipulation. The finger enables re-orientation of a grasped object by using a pneumatic braking mechanism to transition between free-rotating and fixed (i.e., braked) phases. Our design allows controlled high-bandwidth (5 Hz) phase transitions independent of the grasping force for manipulation of a variety of objects. Moreover, its thin profile (1 cm) facilitates picking and placing in clutter. Finally, the design features a sensor for measuring fingertip rotation to support feedback control. We experimentally characterize the finger's load handling capacity in the brake phase and rotational resistance in the free phase. We also demonstrate several pick-and-place manipulations common to industrial and laboratory automation settings that are simplified by our design.

We investigate whether a robot arm can learn to pick and throw arbitrary rigid objects into selected boxes quickly and accurately. Throwing has the potential to increase the physical reachability and picking speed of a robot arm. However, precisely throwing arbitrary objects in unstructured settings presents many challenges: from acquiring objects in grasps suitable for reliable throwing, to handling varying object-centric properties (e.g., mass distribution, friction, shape) and complex aerodynamics. In this work, we propose an end-to-end formulation that jointly learns to infer control parameters for grasping and throwing motion primitives from visual observations (RGB-D images of arbitrary objects in a bin) through trial and error. Within this formulation, we investigate the synergies between grasping and throwing (i.e., learning grasps that enable more accurate throws) and between simulation and deep learning (i.e., using deep networks to predict residuals on top of control parameters predicted by a physics simulator). The resulting system, TossingBot, is able to grasp and successfully throw arbitrary objects into boxes located outside its maximum reach range at 500+ mean picks per hour (600+ grasps per hour with 85% throwing accuracy); and generalizes to new objects and target locations.

We explore the use of graph neural networks (GNNs) to model spatial processes in which there is no a priori graphical structure. Similar to finite element analysis, we assign nodes of a GNN to spatial locations and use a computational process defined on the graph to model the relationship between an initial function defined over a space and a resulting function in the same space. We use GNNs as a computational substrate, and show that the locations of the nodes in space as well as their connectivity can be optimized to focus on the most complex parts of the space. Moreover, this representational strategy allows the learned input-output relationship to generalize over the size of the underlying space and run the same model at different levels of precision, trading computation for accuracy. We demonstrate this method on a traditional PDE problem, a physical prediction problem from robotics, and learning to predict scene images from novel viewpoints.

Physics engines play an important role in robot planning and control; however, many real-world control problems involve complex contact dynamics that cannot be characterized analytically. Most physics engines therefore employ approximations that lead to a loss in precision. In this paper, we propose a hybrid dynamics model, simulator-augmented interaction networks (SAIN), combining a physics engine with an object-based neural network for dynamics modeling. Compared with existing models that are purely analytical or purely data-driven, our hybrid model captures the dynamics of interacting objects in a more accurate and data-efficient manner. Experiments both in simulation and on a real robot suggest that it also leads to better performance when used in complex control tasks. Finally, we show that our model generalizes to novel environments with varying object shapes and materials.

This work studies the problem of shape reconstruction and object localization using a vision-based tactile sensor, GelSlim. The main contributions are the recovery of local shapes from contact, an approach to reconstruct the tactile shape of objects from tactile imprints, and an accurate method for object localization of previously reconstructed objects. The algorithms can be applied to a large variety of 3D objects and provide accurate tactile feedback for in-hand manipulation. Results show that by exploiting the dense tactile information we can reconstruct the shape of objects with high accuracy and do on-line object identification and localization, opening the door to reactive manipulation guided by tactile sensing. We provide videos and supplemental information in the project's website web.mit.edu/mcube/research/tactile localization.html.

In this paper, we propose an approach to detect incipient slip, i.e. predict slip, by using a high-resolution vision-based tactile sensor, GelSlim. The sensor dynamically captures the tactile imprints of the grasped object and their changes with a soft gel pad. The method assumes the object is mostly rigid and expects the motion of object's imprint on the sensor surface to be a 2D rigid-body motion. We use the deviation of the true motion field from that of a 2D planar rigid transformation as a measure of slip. The output is a dense slip field which we monitor in real time to detect when small areas of the contact patch start to slip (incipient slip). The method can detect incipient slip in any direction without any prior knowledge of the object at 24 Hz. We test the method on 10 objects for 240 times and achieve 86.25% detection accuracy with the vast majority of failure cases occurring when grasping highly deformable objects. We further show how the slip feedback can be used to adjust the gripping force to avoid slip with a closed-loop bottle-cap screwing and unscrewing experiment. The method can be used to enable many manipulation tasks in both structured and unstructured environments.

In this paper, we present a new version of tactile sensor GelSlim 2.0 with the capability to estimate the contact force distribution in real time. The sensor is vision-based and uses an array of markers to track deformations on a gel pad due to contact. A new hardware design makes the sensor more rugged, parametrically adjusTable AND Improves illumination. leveraging the sensor's increased functionality, we propose to use inverse finite element method (ifem), a numerical method to reconstruct the contact force distribution based on marker displacements. the sensor is able to provide force distribution of contact with high spatial density. experiments and comparison with ground truth show that the reconstructed force distribution is physically reasonable with good accuracy.A sequence of Kendama manipulations with corresponding displacement field (yellow) and force field (red). Video can be found on Youtube: https://youtu.be/hWw9A0ZBZuU.

Caging is a promising tool which allows a robot to manipulate an object without directly reasoning about the contact dynamics involved. Furthermore, caging also provides useful guarantees in terms of robustness to uncertainty, and often serves as a way-point to a grasp. However, caging is traditionally difficult to integrate as part of larger manipulation frameworks, where caging is not the goal but an intermediate condition. In this paper, we develop a convex-combinatorial model to characterize caging from an optimization perspective. More specifically, we derive a set of sufficient constraints to enclose the configuration of the object in a compact-connected component of its free-space. The convex-combinatorial nature of this approach provides guarantees on optimality and convergence, and its optimization nature makes it versatile for further applications on robot manipulation tasks. To the best of our knowledge, this is the first optimization-based approach to formulate the caging condition.

Pushing is a fundamental robotic skill. Existing work has shown how to exploit models of pushing to achieve a variety of tasks, including grasping under uncertainty, in-hand manipulation and clearing clutter. Such models, however, are approximate, which limits their applicability.Learning-based methods can reason directly from raw sensory data with accuracy, and have the potential to generalize to a wider diversity of scenarios. However, developing and testing such methods requires rich-enough datasets. In this paper we introduce Omnipush, a dataset with high variety of planar pushing behavior.In particular, we provide 250 pushes for each of 250 objects, all recorded with RGB-D and a high precision tracking system. The objects are constructed so as to systematically explore key factors that affect pushing-the shape of the object and its mass distribution-which have not been broadly explored in previous datasets, and allow to study generalization in model learning.Omnipush includes a benchmark for meta-learning dynamic models, which requires algorithms that make good predictions and estimate their own uncertainty. We also provide an RGB video prediction benchmark and propose other relevant tasks that can be suited with this dataset. Data and code are available at https://web.mit.edu/mcube/omnipush-dataset/.

We study the problem of using high-resolution tactile sensors to control the insertion of objects in a boxpacking scenario. In this paper, we propose an insertion strategy that leverages tactile sensing to: 1) safely probe the box with the grasped object while monitoring incipient slip to maintain a stable grasp on the object. 2) estimate and correct for residual position uncertainties to insert the object into a designated gap without disturbing the environment.Our proposed methodology is based on two neural networks that estimate the error direction and error magnitude, from a stream of tactile imprints, acquired by two GelSlim fingers, during the insertion process. The system is trained on four objects with basic geometric shapes, which we show generalizes to four other common objects. Based on the estimated positional errors, a heuristic controller iteratively adjusts the position of the object and eventually inserts it successfully without requiring prior knowledge of the geometry of the object. The key insight is that dense tactile feedback contains useful information with respect to the contact interaction between the grasped object and its environment. We achieve high success rate and show that unknown objects can be inserted with an average of 6 attempts of the probe-correct loop. The method's ability to generalize to novel objects makes it a good fit for box packing in warehouse automation.

The use of hand tools presents a challenge for robot manipulation in part because it calls for motions requiring continuous force application over a whole trajectory, usually involving large joint-angle excursions. The feasible application of a tool, such as pulling a nail with a hammer claw, requires careful coordination of the choice of grasp and joint trajectories to ensure kinematic and force limits are not exceeded - in the grasp as well as the robot mechanism. In this paper, we formulate this type of problem as choosing the values of decision variables in the presence of various constraints. We evaluate the impact of the various constraints in some representative instances of tool use. To aid others in further investigating this class of problems, we have released materials such as printable tool models and experimental data. We hope that these can serve as the basis of a benchmark problem for investigating tasks that involve many kinematic, actuation, friction, and environment constraints.

This paper presents an approach to in-hand manipulation planning that exploits the mechanics of alternating sticking contact. Particularly, we consider the problem of manipulating a grasped object using external pushes for which the pusher sticks to the object. Given the physical properties of the object, frictional coefficients at contacts and a desired regrasp on the object, we propose a sampling-based planning framework that builds a pushing strategy concatenating different feasible stable pushes to achieve the desired regrasp. An efficient dynamics formulation allows us to plan in-hand manipulations 100-1000 times faster than our previous work which builds upon a complementarity formulation. Experimental observations for the generated plans show that the object precisely moves in the grasp as expected by the planner.

This paper presents a reactive controller for planar manipulation tasks that leverages machine learning to achieve real-time performance. The approach is based on a Model Predictive Control (MPC) formulation, where the goal is to find an optimal sequence of robot motions to achieve a desired object motion. Due to the multiple contact modes associated with frictional interactions, the resulting optimization program suffers from combinatorial complexity when tasked with determining the optimal sequence of modes. To overcome this difficulty, we formulate the search for the optimal mode sequences offline, separately from the search for optimal control inputs online. Using tools from machine learning, this leads to a convex hybrid MPC program that can be solved in real-time. We validate our algorithm on a planar manipulation experimental setup where results show that the convex hybrid MPC formulation with learned modes achieves good closed-loop performance on a trajectory tracking problem.

Accurate and robust object state estimation enables successful object manipulation. Visual sensing is widely used to estimate object poses. However, in a cluttered scene or in a tight workspace, the robot's end-effector often occludes the object from the visual sensor. The robot then loses visual feedback and must fall back on open-loop execution. In this paper, we integrate both tactile and visual input using a framework for solving the SLAM problem, incremental smoothing and mapping (iSAM), to provide a fast and flexible solution. Visual sensing provides global pose information but is noisy in general, whereas contact sensing is local, but its measurements are more accurate relative to the end-effector. By combining them, we aim to exploit their advantages and overcome their limitations. We explore the technique in the context of a pusher-slider system. We adapt iSAM's measurement cost and motion cost to the pushing scenario, and use an instrumented setup to evaluate the estimation quality with different object shapes, on different surface materials, and under different contact modes.

This paper presents a robotic pick-and-place system that is capable of grasping and recognizing both known and novel objects in cluttered environments. The key new feature of the system is that it handles a wide range of object categories without needing any task-specific training data for novel objects. To achieve this, it first uses a category-agnostic affordance prediction algorithm to select and execute among four different grasping primitive behaviors. It then recognizes picked objects with a cross-domain image classification framework that matches observed images to product images. Since product images are readily available for a wide range of objects (e.g., from the web), the system works out-of-the-box for novel objects without requiring any additional training data. Exhaustive experimental results demonstrate that our multi-affordance grasping achieves high success rates for a wide variety of objects in clutter, and our recognition algorithm achieves high accuracy for both known and novel grasped objects. The approach was part of the MIT-Princeton Team system that took 1st place in the stowing task at the 2017 Amazon Robotics Challenge. All code, datasets, and pre-trained models are available online at http://arc.cs.princeton.edu.

An efficient, generalizable physical simulator with universal uncertainty estimates has wide applications in robot state estimation, planning, and control. In this paper, we build such a simulator for two scenarios, planar pushing and ball bouncing, by augmenting an analytical rigid-body simulator with a neural network that learns to model uncertainty as residuals. Combining symbolic, deterministic simulators with learnable, stochastic neural nets provides us with expressiveness, efficiency, and generalizability simultaneously. Our model outperforms both purely analytical and purely learned simulators consistently on real, standard benchmarks. Compared with methods that model uncertainty using Gaussian processes, our model runs much faster, generalizes better to new object shapes, and is able to characterize the complex distribution of object trajectories.

This work describes the development of a high-resolution tactile-sensing finger for robot grasping. This finger, inspired by previous GelSight sensing techniques (Johnson and Adelson 2009), features an integration that is slimmer, more robust, and with more homogeneous output than previous vision-based tactile sensors. To achieve a compact integration, we redesign the optical path from illumination source to camera by combining light guides and an arrangement of mirror reflections. We parameterize the optical path with geometric design variables and describe the tradeoffs between the finger thickness, camera depth of field, and size of the tactile sensing area. The sensor sustains the wear from continuous use - and abuse - in grasping tasks by combining tougher materials for the compliant gel, a textured fabric skin, a structurally rigid body, and a calibration process that maintains homogeneous illumination and contrast of the tactile images during use. Finally, we evaluate the sensor's durability along four metrics that track the signal quality during more than 3000 grasping experiments.

This paper presents a novel regrasp control policy that makes use of tactile sensing to plan local grasp adjustments. Our approach determines regrasp actions by virtually searching for local transformations of tactile measurements that improve the quality of the grasp. First, we construct a tactile-based grasp quality metric using a deep convolutional neural network trained on over 2800 grasps. The quality of each grasp, a continuous value between 0 and 1, is determined experimentally by measuring its resistance to external perturbations. Second, we simulate the tactile imprints associated with robot motions relative to the initial grasp by performing rigid-body transformations of the given tactile measurements. The newly generated tactile imprints are evaluated with the learned grasp quality network and the regrasp action is chosen to maximize the grasp quality. Results show that the grasp quality network can predict the outcome of grasps with an average accuracy of 85% on known objects and 75% on novel objects. The regrasp control policy improves the success rate of grasp actions by an average relative increase of 70% on a test set of 8 objects. We provide a video summarizing our approach at https://youtu.be/gjn7DmfpwDk.

We develop a real-time state estimation system to recover the pose and contact formation of an object relative to its environment. In this paper, we focus on the application of inserting an object picked by a suction cup into a tight space, a key technology for robotic packaging. We propose a framework that fuses tactile and visual sensing. Visual sensing is versatile and non-intrusive, but suffers from occlusions and limited accuracy, especially for tasks involving contact. Tactile sensing is local, but provides accuracy and robustness to occlusions. The proposed algorithm to fuse them is based on iSAM, an on-line estimation technique, which we use to incorporate kinematic measurements from the robot, contact geometry of the object and the container, and visual tracking. In this paper, we generalize previous results in planar settings [1] to a 3D task with more complex contact interactions. A key challenge is that we do not observe contact locations between the suction-held object and the container directly. We propose a data-driven method to infer the contact formation, which is then used in real-time by the state estimator. We demonstrate and evaluate the algorithm in a setup instrumented to provide groundtruth.

Skilled robotic manipulation benefits from complex synergies between non-prehensile (e.g. pushing) and prehensile (e.g. grasping) actions: pushing can help rearrange cluttered objects to make space for arms and fingers; likewise, grasping can help displace objects to make pushing movements more precise and collision-free. In this work, we demonstrate that it is possible to discover and learn these synergies from scratch through model-free deep reinforcement learning. Our method involves training two fully convolutional networks that map from visual observations to actions: one infers the utility of pushes for a dense pixel-wise sampling of end-effector orientations and locations, while the other does the same for grasping. Both networks are trained jointly in a Q-learning framework and are entirely self-supervised by trial and error, where rewards are provided from successful grasps. In this way, our policy learns pushing motions that enable future grasps, while learning grasps that can leverage past pushes. During picking experiments in both simulation and real-world scenarios, we find that our system quickly learns complex behaviors even amid challenging cases of tightly packed clutter, and achieves better grasping success rates and picking efficiencies than baseline alternatives after a few hours of training. We further demonstrate that our method is capable of generalizing to novel objects. Qualitative results (videos), code, pre-trained models, and simulation environments are available at http://vpg.cs.princeton.edu/

This paper presents an overview of the inaugural Amazon Picking Challenge along with a summary of a survey conducted among the 26 participating teams. The challenge goal was to design an autonomous robot to pick items from a warehouse shelf. This task is currently performed by human workers, and there is hope that robots can someday help increase efficiency and throughput while lowering cost. We report on a 28-question survey posed to the teams to learn about each team's background, mechanism design, perception apparatus, planning, and control approach. We identify trends in this data, correlate it with each team's success in the competition, and discuss observations and lessons learned based on survey results and the authors' personal experiences during the challenge.

Automated grasping has a long history of research that is increasing due to interest from industry. One grand challenge for robotics is Universal Picking: the ability to robustly grasp a broad variety of objects in diverse environments for applications from warehouses to assembly lines to homes. Although many researchers now openly share code and data, it is challenging to compare and/or reproduce experimental results to identify which aspects of which approaches work best due to variations in assumptions and experimental protocols, e.g., sensors, lighting, robot arms, grippers, and objects.

This paper presents a data-driven approach to model planar pushing interaction to predict both the most likely outcome of a push and its expected variability. The learned models rely on a variation of Gaussian processes with input-dependent noise called Variational Heteroscedastic Gaussian processes (VHGP) [1] that capture the mean and variance of a stochastic function. We show that we can learn accurate models that outperform analytical models after less than 100 samples and saturate in performance with less than 1000 samples. We validate the results against a collected dataset of repeated trajectories, and use the learned models to study questions such as the nature of the variability in pushing, and the validity of the quasi-static assumption.

In this paper we evaluate the predictive performance of six commonly used rigid body impact models on real planar impacts captured with a motion tracking system. We propose a metric to evaluate the performance of impact models on a task (based on predicting post impact momentum) and use this metric to tune the six parametric models. We evaluate model performance in predicting impact outcomes against the defined metric and discuss the implications of uncertainty in geometric models and initial conditions. We show that the models can fairly effectively predict the outcomes of single impacts on our chosen task. We motivate further study into consensus and hybrid impact models by showing that a hypothetical hybrid model would significantly outperform the isolated models by providing a post-hoc model that demonstrates an upper bound on the combined predictive power of the models. We use perturbation analysis to compute the predictive range of the models and show that bifurcations can cause the model predictions to cluster into regions of the state space.

Robot warehouse automation has attracted significant interest in recent years, perhaps most visibly in the Amazon Picking Challenge (APC) [1]. A fully autonomous warehouse pick-and-place system requires robust vision that reliably recognizes and locates objects amid cluttered environments, self-occlusions, sensor noise, and a large variety of objects. In this paper we present an approach that leverages multiview RGB-D data and self-supervised, data-driven learning to overcome those difficulties. The approach was part of the MIT-Princeton Team system that took 3rd- and 4th-place in the stowing and picking tasks, respectively at APC 2016. In the proposed approach, we segment and label multiple views of a scene with a fully convolutional neural network, and then fit pre-scanned 3D object models to the resulting segmentation to get the 6D object pose. Training a deep neural network for segmentation typically requires a large amount of training data. We propose a self-supervised method to generate a large labeled dataset without tedious manual segmentation. We demonstrate that our system can reliably estimate the 6D pose of objects under a variety of scenarios. All code, data, and benchmarks are available at http://apc.cs.princeton.edu/

Pushing is a motion primitive useful to handle objects that are too large, too heavy, or too cluttered to be grasped. It is at the core of much of robotic manipulation, in particular when physical interaction is involved. It seems reasonable then to wish for robots to understand how pushed objects move. In reality, however, robots often rely on approximations which yield models that are computable, but also restricted and inaccurate. Just how close are those models? How reasonable are the assumptions they are based on? To help answer these questions, and to get a better experimental understanding of pushing, we present a comprehensive and high-fidelity dataset of planar pushing experiments. The dataset contains time-stamped poses of a circular pusher and a pushed object, as well as forces at the interaction. We vary the push interaction in 6 dimensions: surface material, shape of the pushed object, contact position, pushing direction, pushing speed, and pushing acceleration. An industrial robot automates the data capturing along precisely controlled position-velocity-acceleration trajectories of the pusher, which give dense samples of positions and forces of uniform quality. We finish the paper by characterizing the variability of friction, and evaluating the most common assumptions and simplifications made by models of frictional pushing in robotics.

This paper explores the manipulation of a grasped object by pushing it against its environment. Relying on precise arm motions and detailed models of frictional contact, prehensile pushing enables dexterous manipulation with simple manipulators, such as those currently available in industrial settings, and those likely affordable by service and field robots.

Tactile exploration refers to the use of physical interaction to infer object properties. In this work, we study the feasibility of recovering the shape and pose of a movable object from observing a series of contacts. In particular, we approach the problem of estimating the shape and trajectory of a planar object lying on a frictional surface, and being pushed by a frictional probe. The probe, when in contact with the object, makes observations of the location of contact and the contact normal.

This paper presents a novel nonlinear compliant link. It has two major properties: bi-directionality and stiffening compliance. Bi-directionality means it can be stretched and compressed, and is realized by antagonistic arrangement of an extension spring and a compression spring. Stiffening compliance means it becomes stiffer as it is stretched, and is realized by asymmetric geometry. The links are parts of Simple Hand. Because Simple Hand gives limited space for links, current iteration of links is not obviously nonlinear. However, nonlinearity should be more obvious if links are designed for larger grippers.

“In-hand manipulation” is the ability to reposition an object in the hand, for example when adjusting the grasp of a hammer before hammering a nail. The common approach to in-hand manipulation with robotic hands, known as dexterous manipulation [1], is to hold an object within the fingertips of the hand and wiggle the fingers, or walk them along the object's surface. Dexterous manipulation, however, is just one of the many techniques available to the robot. The robot can also roll the object in the hand by using gravity, or adjust the object's pose by pressing it against a surface, or if fast enough, it can even toss the object in the air and catch it in a different pose. All these techniques have one thing in common: they rely on resources extrinsic to the hand, either gravity, external contacts or dynamic arm motions. We refer to them as “extrinsic dexterity”. In this paper we study extrinsic dexterity in the context of regrasp operations, for example when switching from a power to a precision grasp, and we demonstrate that even simple grippers are capable of ample in-hand manipulation. We develop twelve regrasp actions, all open-loop and hand-scripted, and evaluate their effectiveness with over 1200 trials of regrasps and sequences of regrasps, for three different objects (see video [2]). The long-term goal of this work is to develop a general repertoire of these behaviors, and to understand how such a repertoire might eventually constitute a general-purpose in-hand manipulation capability.

This video presents the application of Extrinsic Dexterity to change the pose of an object in the hand, i.e., to regrasp the object.

Given a desired function for an effector, what is its appropriate shape? This paper formulates mechanical function as a product of both effector's shape and motion, and, assuming a fixed motion model, explores the role of shape in satisfying it. We assume that the desired mechanical function is expressed as a set of constraints on the geometry of contact, and develop the tools for transforming these constraints into an effector shape. A previous paper [1] addressed the special case of revolute or prismatic fingers. This paper develops the more general case, including all smooth 1DOF planar mechanisms. The technique is illustrated with the design of finger shapes to improve the stability of a planar grasp of an object.

Sensing the forces applied to the palm of a robot manipulator requires robust compliant sensors guarded against collision. The force sensing palm for the MLab Simple Hand measures three components of net contact force in a simple robust design using flexure springs and optical position sensing. The hand is calibrated in an automatic process. Measurements are included to compare performance to the theoretical model.

This paper revisits the notion of free configuration space and reviews some of its path-connectivity-related properties. The literature on motion planning reveals at least three different definitions for the free configuration space of a robot in the presence of obstacles. This paper shows that, assuming regularity of both object and obstacles, those three definitions are equivalent. We show that the three definitions regularize the free space and therefore prevent the existence of “thin bits,” or low-dimensional strata. The paper concludes by discussing a series of properties regarding the existence and smoothability of contact-free paths between pairs of configurations.

Iteration is often sufficient for a simple hand to accomplish complex tasks, at the cost of an increase in the expected time to completion. In this paper, we minimize that overhead time by allowing a simple hand to abort early and retry as soon as it realizes that the task is likely to fail. We present two key contributions. First, we learn a probabilistic model of the relationship between the likelihood of success of a grasp and its grasp signature—the trace of the state of the hand along the entire grasp motion. Second, we model the iterative process of early abort and retry as a Markov chain and optimize the expected time to completion of the grasping task by effectively thresholding the likelihood of success. Experiments with our simple hand prototype tasked with grasping and singulating parts from a bin show that early abort and retry significantly increases efficiency.