Abstract — This paper combines grasp analysis and manipulation planning techniques to perform fast grasp planning in complex scenes. In much previous work on grasping, the object being grasped is assumed to be the only object in the environment. Hence the grasp quality metrics and grasping strategies developed do not perform well when the object is close to obstacles and many good grasps are infeasible. We introduce a framework for finding valid grasps in cluttered environments that combines a grasp quality metric for the object with information about the local environment around the object and information about the robot’s kinematics. We encode these factors in a grasp-scoring function which we use to rank a precomputed set of grasps in terms of their appropriateness for a given scene. We show that this ranking is essential for efficient grasp selection and present experiments in simulation and on the HRP2 robot. I.

Abstract — We present an algorithm for efficiently generating collision-free force-closure grasps for dexterous hands in cluttered environments. Computing a grasp is complicated by the high dimensionality of the hand configuration space, and the high cost of validating a candidate grasp by collision-checking and testing for force-closure. When an object is placed in a new scene, we use a novel cost function to focus our search to good regions of hand pose space for a given preshape. The proposed cost function is fast to compute and encapsulates aspects of the object, the scene, and the force-closure of the ensuing grasp. The low cost candidate grasps produced by the search are then validated. We demonstrate the generality of our approach by testing on the 3-fingered 4DOF Barrett hand and the anthropomorphic 22DOF Shadow hand. Our results show that the candidate grasps generated by our algorithm consistently have high probability of being valid for various hands, objects and scenes. Finally, we describe an implementation on a WAM arm with a Barrett Hand. I.

Abstract — To grasp a novel object, we can index it into a database of known 3D models and use precomputed grasp data for those models to suggest a new grasp. We refer to this idea as data-driven grasping, and we have previously introduced the Columbia Grasp Database for this purpose. In this paper we demonstrate a data-driven grasp planner that requires only partial 3D data of an object in order to grasp it. To achieve this, we introduce a new shape descriptor for partial 3D range data, along with an alignment method that can rigidly register partial 3D models to models that are globally similar but not identical. Our method uses SIFT features of depth images, and encapsulates “nearby ” views of an object in a compact shape descriptor. I.

In this paper, we propose a prior for hand pose estimation that integrates the direct relation between a manipulating hand and a 3d object. This is of particular interest for a variety of applications since many tasks performed by humans require hand-object interaction. Inspired by the ability of humans to learn the handling of an object from a single example, our focus lies on very sparse training data. We express estimated hand poses in local object coordinates and extract for each individual hand segment, the relative position and orientation as well as contact points on the object. The prior is then modeled as a spatial distribution conditioned to the object. Given a new object of the same object class and new hand dimensions, we can transfer the prior by a procedure involving a geometric warp. In our experiments, we demonstrate that the prior may be used to improve the robustness of a 3d hand tracker and to synthesize a new hand grasping a new object. For this, we integrate the prior into a unified belief propagation framework for tracking and synthesis. 1.

Every planning problem in robotics involves constraints. Whether the robot must avoid collision or joint limits, there are always states that are not permissible. Some constraints are straightforward to satisfy while others can be so stringent that feasible states are very difficult to find. What makes planning with constraints challenging is that, for many constraints, it is impossible or impractical to provide the planning algorithm with the allowed states explicitly; it must discover these states as it plans. The goal of this thesis is to develop a framework for representing and exploring feasible states in the context of manipulation planning. Planning for manipulation gives rise to a rich variety of tasks that include constraints on collisionavoidance, torque, balance, closed-chain kinematics, and end-effector pose. While many researchers have developed representations and strategies to plan with a specific constraint, the goal of this thesis is to develop a broad representation of constraints on a robot’s configuration and identify general strategies to manage these constraints during the planning process. Some of the most important constraints in manipulation planning are functions of the pose of the manipulator’s end-effector, so we devote a large part of this thesis to end-effector placement for grasping and transport tasks. We present an efficient approach to generating paths that uses Task Space Regions (TSRs) to specify manipulation

In this paper we focus on the concept of low-dimensional posture subspaces for artificial hands. We begin by discussing the applicability of a hand configuration subspace to the problem of automated grasp synthesis � our results show that low-dimensional optimization can be instrumental in deriving effective pre-grasp shapes for a number of complex robotic hands. We then show that the computational advantages of using a reduced dimensionality framework enable it to serve as an interface between the human and automated components of an interactive grasping system. We present an on-line grasp planner that allows a human operator to perform dexterous grasping tasks using an artificial hand. In order to achieve the computational rates required for effective user interaction, grasp planning is performed in a hand posture subspace of highly reduced dimensionality. The system also uses real-time input provided by the operator, further simplifying

Animating hand-object interactions is a frequent task in applications such as the production of 3d movies. Unfortunately this task is difficult due to the hand’s many degrees of freedom and the constraints on the hand motion imposed by the geometry of the object. However, the causality between the object state and the hand’s pose can be exploited in order to simplify the animation process. In this paper, we present a method that takes an animation of an object as input and automatically generates the corresponding hand motion. This approach is based on the simple observation that objects are easier to animate than hands, since they usually have fewer degrees of freedom. The method is data-driven; sequences of hands manipulating an object are captured semiautomatically with a structured-light setup. The training data is then combined with a new animation of the object in order to generate a plausible animation featuring the hand-object interaction.

Articulated hand-tracking systems have been widely used in virtual reality but are rarely deployed in consumer applications due to their price and complexity. In this paper, we propose an easy-to-use and inexpensive system that facilitates 3-D articulated user-input using the hands. Our approach uses a single camera to track a hand wearing an ordinary cloth glove that is imprinted with a custom pattern. The pattern is designed to simplify the pose estimation problem, allowing us to employ a nearest-neighbor approach to track hands at interactive rates. We describe several proof-of-concept applications enabled by our system that we hope will provide a foundation for new interactions in modeling, animation control and augmented reality.

Figure 1: We describe a system that can reconstruct the pose of the hand from a single image of the hand wearing a multi-colored glove. We demonstrate our system as a user-input device for desktop virtual reality applications. Articulated hand-tracking systems have been widely used in virtual reality but are rarely deployed in consumer applications due to their price and complexity. In this paper, we propose an easy-to-use and inexpensive system that facilitates 3-D articulated user-input using the hands. Our approach uses a single camera to track a hand wearing an ordinary cloth glove that is imprinted with a custom pattern. The pattern is designed to simplify the pose estimation problem, allowing us to employ a nearest-neighbor approach to track hands at interactive rates. We describe several proof-of-concept applications enabled by our system that we hope will provide a foundation for new interactions in modeling, animation control and augmented reality.

This thesis introduces new methods for enabling the effective use of highly dexterous robotic hands, interfacing with the upcoming generation of neurally controlled hand prostheses, and designing a new class of simple yet effective grasping devices based on underactuation and mechanical adaptation. These methods share a common goal: reducing the complexity that has traditionally been associated, at both computational and mechanical levels, with robotic grasping in unstructured environments. A key prerequisite for robot operation in human settings is versatility, which, in terms of autonomous grasping, translates into the ability to reliably acquire and interact with a wide range of objects. In an attempt to match the abilities of the most versatile end-effector known, the human hand, many anthropomorphic robotic models have been proposed, with the number of degrees of freedom starting to approach that of their human counterpart. However, these models have proven difficult to use in practice, as the high dimensionality of the posture space means that finding adequate grasps for a target object is often an intractable problem. In this thesis, we propose using low-dimensional posture subspaces for dexterous or