IROS 2016 论文列表

In a robot mounted camera arrangement, hand-eye calibration estimates the rigid relationship between the robot and camera coordinate frames. Most hand-eye calibration techniques use a calibration object to estimate the relative transformation of the camera in several views of the calibration object and link these to the forward kinematics of the robot to compute the hand-eye transformation. Such approaches achieve good accuracy for general use but for applications such as robotic assisted minimally invasive surgery, acquiring a calibration sequence multiple times during a procedure is not practical. In this paper, we present a new approach to tackle the problem by using the robotic surgical instruments as the calibration object with well known geometry from CAD models used for manufacturing. Our approach removes the requirement of a custom sterile calibration object to be used in the operating room and it simplifies the process of acquiring calibration data when the laparoscope is constrained to move around a remote centre of motion. This is the first demonstration of the feasibility to perform hand-eye calibration using components of the robotic system itself and we show promising validation results on synthetic data as well as data acquired with the da Vinci Research Kit.

We present HPP, a software designed for complex classes of motion planning problems, such as navigation among movable objects, manipulation, contact-rich multiped locomotion, or elastic rods in cluttered environments. HPP is an open-source answer to the lack of a standard framework for these important issues for robotics and graphics communities.

We present an algorithmic framework for the early classification of human intentions, and use it to accurately predict future human motions when planning the path of a robot in an environment that is shared with humans. During an off-line learning phase, a classifier that can recognize when a human intends to interact with the robot is trained. At runtime, this trained classifier allows us to recognize humans who intend to interact with, or obstruct, the robot in some way. We validate our approach using both recorded and simulated data in an environment in which some humans intentionally obstruct the robot. Our classifier identifies these potential blockers, thus allowing the robot to safely and efficiently navigate the environment by minimizing the chances of being blocked.

Exoskeletons offer an intuitive method for actuating multiple DOF of the body; this makes them attractive for applications where generation and coupling of artificial forces to the limbs is needed. Force feedback hand exoskeletons have been continuously considered for whole hand haptic interaction in virtual reality simulators, in teleoperation setups and for rehabilitation. In hand exoskeletons finger tracking, actuation and transmission systems must be embedded in confined spaces, matching at the same time the profound dexterity of the hand transparently and without causing a burden. Most of the design approaches for such systems have remained largely experimental due to hardware limitations, impacting heavily on important functional and ergonomic factors. This paper presents the design of a novel 3-digit hand exoskeleton, which addresses the issues of finger tracking and force feedback. It proposes a new approach for the application of the feedback force with a single attachment at the fingertip through a 6DoF kinematic chain. This kinematic linkage allows for unconstrained reach of the fingers within their full workspace and facilitates a sensor system for high resolution 6DOF tracking of the fingertips. At the same time the highly under-actuated mechanism permits application of a bidirectional feedback force at the fingertips. The hand exoskeleton fits an large range of hand sizes and requires no mechanical alignment between the linkage and the fingers, whatsoever. Preliminary results show the efficacy of this system as a tracking and force feedback device for the hand.

Micro Aerial Vehicles (MAVs) are limited in their operation outdoors near obstacles by their ability to withstand wind gusts. Currently widespread position control methods such as Proportional Integral Derivative control do not perform well under the influence of gusts. Incremental Nonlinear Dynamic Inversion (INDI) is a sensor-based control technique that can control nonlinear systems subject to disturbances. This method was developed for the attitude control of MAVs, but in this paper we generalize this method to the outer loop control of MAVs under gust loads. Significant improvements over a traditional Proportional Integral Derivative (PID) controller are demonstrated in an experiment where the drone flies in and out of a fan's wake. The control method does not rely on frequent position updates, so it is ready to be applied outside with standard GPS modules.

This paper designs a guide rail for a wall-climbing robot based on passive suction cups. The designed guide rail can guarantee stable climbing of a wall-climbing robot. Firstly, to design the parameters of the guide rail, properties of the utilized passive suction cup are experimentally studied. Then a guide rail is designed to distribute the forces of the attached suction cups. It guarantees that the front attached suction cup of the robot obtains a large enough reaction force from the attached surface. Experimental results show that the robot prototype with the designed guide rail is able to stably climb a surface without falling off, since the front suction cup of the robot prototype can be completely attached to the climbing surface.

This paper is devoted to the collective control and group navigation based on hydroacoustic communication intended for marine autonomous robotic complex (MARC) including autonomous underwater vehicle (AUV) and autonomous surface vehicle (ASV). The structures of control and navigation systems are considered and some results of the systems operation in the real marine environment are supplemented. The contribution of this paper is in review of current hydroacoustic (HA) modems for marine robots, implementation details of MARC structure and HA communication and some field test results.

In robotics, robotic arms should be smaller, less expensive, and able to demonstrate a better performance, in order to be utilized in various fields. However, expensive actuators and speed reducers are required to generate sufficient force at the end effector after overcoming the gravitational torque of robotic arm. Robotic arms become compact and cost effective if the gravitational torque can be passively compensated. Especially in the field of robotic laparoscopic surgery, the static equilibrium state of the robot arm in the entire workspace is essential to achieve both safety and ease of use. In this study, we propose a novel passive gravity compensation mechanism based on springs and wires. This mechanism utilizes a scotch yoke mechanism to compensate the gravitational torque changed by two rotational degrees of freedom (DOF). Furthermore, we applied to a surgical robotic arm which was constructed using a roll-pitch joint. According to the experiments, it was proved that the new mechanism effectively compensate the gravitational torque. As a result, compact and cost-effective robotic arms can be developed satisfying safety and ease of use.

This paper presents an algorithm to compute precision grasps for bulky objects using two anthropomorphic hands. We use objects modeled as point clouds obtained from a sensor camera or from a CAD model. We process the point clouds dividing them into two set of slices, one for each hand, where we look for sets of triplets of points. Each triplet must accomplish some physical conditions based on the structure of the hands. Then, triplets of points from each set of slices are evaluated to find a combination that satisfies the force closure condition (FC). Once one valid couple of triplets have been found the inverse kinematics of the system is computed in order to know if the points are kinematically reachable by the hands, if so, motion planning and a collision check are performed to asses if the grasp configuration of the system is suitable. The paper includes application examples of the proposed approach.

Underactuated and synergy-driven hands are gaining attention in the grasping community mainly due to their simple kinematics, intrinsic compliance and versatility for grasping objects even in non structured scenarios. The evaluation of the grasping capabilities of such hands is a challenging task. This paper revisits some traditional quality measures developed for multi-fingered, fully actuated hands, and applies them to the case of underactuated hands. The extension of quality metrics for synergy-driven hands for the case of underactuated grasping is also presented. The performance of both types of measures is evaluated with simulated examples, concluding with a comparative discussion of their main features.

Grasping systems that build upon meticulously planned hand postures rely on precise knowledge of object geometry, mass and frictional properties — assumptions which are often violated in practice. In this work, we propose an alternative solution to the problem of grasp acquisition in simple autonomous pick and place scenarios, by utilizing the concept of grasp envelopes: sets of constraints on gripper postures. We propose a fast method for extracting grasp envelopes for objects that fit within a known shape category, placed in an unknown environment. Our approach is based on grasp envelope primitives, which encode knowledge of human grasping strategies. We use environment models, reconstructed from noisy sensor observations, to refine the grasp envelope primitives and extract bounded envelopes of collision-free gripper postures. Also, we evaluate the envelope extraction procedure both in a stand alone fashion, as well as an integrated component of an autonomous picking system.

Multi-robot cooperative navigation in real-world environments is essential in many applications, including surveillance and search-and-rescue missions. State-of-the-art methods for cooperative navigation are often tested in ideal laboratory conditions and not ready to be deployed in real-world environments, which are often cluttered with static and dynamic obstacles. In this work, we explore a graph-based framework to achieve control of real robot formations moving in a world cluttered with a variety of obstacles by introducing a new distributed algorithm for reconfiguring the formation shape. We systematically validate the reconfiguration algorithm using three real robots in scenarios of increasing complexity.

Modern SLAM systems with a depth sensor are able to reliably reconstruct dense 3D geometric maps of indoor scenes. Representing these maps in terms of meaningful entities is a step towards building semantic maps for autonomous robots. One approach is to segment the 3D maps into semantic objects using Conditional Random Fields (CRF), which requires large 3D ground truth datasets to train the classification model. Additionally, the CRF inference is often computationally expensive. In this paper, we present an unsupervised geometric-based approach for the segmentation of 3D point clouds into objects and meaningful scene structures. We approximate an input point cloud by an adjacency graph over surface patches, whose edges are then classified as being either on or off. We devise an effective classifier which utilises both global planar surfaces and local surface convexities for edge classification. More importantly, we propose a novel global plane extraction algorithm for robustly discovering the underlying planes in the scene. Our algorithm is able to enforce the extracted planes to be mutually orthogonal or parallel which conforms usually with human-made indoor environments. We reconstruct 654 3D indoor scenes from NYUv2 sequences to validate the efficiency and effectiveness of our segmentation method.

Automated guided vehicles (AGVs) are useful for a variety of transportation tasks. They usually detect obstacles on the path they are following using 2D laser scanners. If an AGV should be deployed in a shared space with people, 3D information has to be considered as well to detect unforeseen obstacles. These can be small objects on the floor or overhanging parts of larger objects, which cannot be seen by the standard 2D safety scanners. We propose a generic object detection pipeline using 3D time-of-flight cameras, that can be used in real-time on AGVs of low height and demonstrate its robustness to different measurement artefacts.

This paper proposes a novel method for generating an underactuated bipedal gait that completes in one step. First, we introduce an underactuated biped robot model that has a circular torso as a reaction wheel. Second, we consider an input-output linearization to formulate an output-following control law for achieving collisionless limit cycle walking. We then mathematically analyze the stability of the zero dynamics and investigate the fundamental gait properties through numerical simulations. Furthermore, we discuss an extension to generation of a few-steps walking motion aimed at safe and quick passage from a safety island to the next.

In this paper, we show the generalization of an inverse dynamic model for KUKA LWR IV+ under load mass variations. We use a modular approach based on regression in the model space. First, inverse dynamic models for the known masses are learned using a recently proposed approach called Independent Joint Learning (IJL). In IJL the torque errors due to unmodeled dynamics of the real robot are estimated using only joint-local information. Second, a mapping from load mass to model parameters of torque error model is learned in order to generalize the inverse dynamics to new load masses. The modular approach improves the accuracy of an existing KUKA LWR IV+ inverse dynamic model. The results are compared with a single step IJL approach. The results show the excellent generalization for new load masses using regression in the model space.

In this paper we present a new estimation algorithm that allows for the combination of information from any number of process and measurement models. This adds more flexibility to the design of the estimator and in our case avoids the need for state augmentation. We achieve this by adapting the maximum likelihood formulation of the Kalman Filter, and thereby represent all measurement models as residuals. Posing the problem in this form allows for the straightforward integration of any number of (nonlinear) constraints between two subsequent states. To solve the optimization we present a closed form recursive set of equations that directly marginalizes out information that is not required, this leads to an efficient and generic implementation. The new algorithm is applied to parameter estimation on MAVs which have two dynamic models, the MAV dynamic model and the IMU-driven model. We show the benefits and limitations of the new filtering approach on a simplified simulation example and on a real MAV system.

In this paper, we propose Gaussian random paths by defining a probability distribution over continuous paths interpolating a finite set of anchoring points using Gaussian process regression. By utilizing the generative property of Gaussian random paths, a Gaussian random path planner is developed to safely steer a robot to a goal position. The Gaussian random path planner can be used in a number of applications, including local path planning for a mobile robot and trajectory optimization for whole body motion planning. We have conducted an extensive set of simulations and experiments, showing that the proposed planner outperforms look-ahead planners which use a pre-defined subset of egocentric trajectories in terms of collision rates and trajectory lengths. Furthermore, we apply the proposed method to existing trajectory optimization methods as an initialization step and demonstrate that it can help produce more cost-efficient trajectories.

Dynamic Movement Primitives (DMPs) represent stable goal-directed or periodic movements, which are learned from observations or demonstrations. They rely on proper function approximators, which are sufficiently flexible to represent arbitrary movements but also ensure goal convergence in point-to-point motions. This work shows that Gaussian Processes (GPs) are suitable as a regressor for learning movements with DMPs ensuring stability. In addition, GPs provide a measure for the uncertainty about the current movement, which we exploit by proposing a new cooperation scheme for DMPs: For better reproduction of demonstrations, we follow the intuition, that individuals with more knowledge lead towards the goal, while others follow and focus on cooperation. Along with simulation results, we validate the presented methods in a robotic cooperative object manipulation task.