ICRA 2015 论文列表

We introduce a novel RGB-D descriptor called local ordinal intensity and normal descriptor (LOIND) with the integration of texture information in RGB image and geometric information in depth image. We implement the descriptor with a 3-D histogram supported by orders of intensities and angles between normal vectors, in addition with the spatial sub-divisions. The former ordering information which is invariant under the transformation of illumination, scale and rotation provides the robustness of our descriptor, while the latter spatial distribution provides higher information capacity so that the discriminative performance is promoted. Comparable experiments with the state-of-art descriptors, e.g. SIFT, SURF, CSHOT and BRAND, show the effectiveness of our LOIND to the complex illumination changes and scale transformation. We also provide a new method to estimate the dominant orientation with only the geometric information, which can ensure the rotation invariance under extremely poor illumination.

This paper is about building maps which not only contain the traditional information useful for localising — such as point features — but also embeds a spatial model of expected localiser performance. This often overlooked second-order information provides vital context when it comes to map use and planning. Our motivation here is to improve the performance of the popular Teach and Repeat paradigm [1] which has been shown to enable truly large-scale field operation. When using the taught route for localisation, it is often assumed the robot is following exactly, or is sufficiently close to, the original path, enabling successful localisation. However, what happens if it is not possible, or not desirable to exactly follow the mapped path? How far off the beaten track can the robot travel before it gets lost? We present an approach for assessing this localisation area around a taught route, which we refer to as the localisation envelope. Using a combination of physical sampling and a Gaussian Process model, we are able to accurately predict the localisation performance at unseen points.

In this work we present a fast kinodynamic RRT-planner that uses dynamic nonprehensile actions to rearrange cluttered environments. In contrast to many previous works, the presented planner is not restricted to quasi-static interactions and monotonicity. Instead the results of dynamic robot actions are predicted using a black box physics model. Given a general set of primitive actions and a physics model, the planner randomly explores the configuration space of the environment to find a sequence of actions that transform the environment into some goal configuration.

Due to excellent dynamic properties and high precision, parallel manipulators are particularly suited for high-speed and high-accuracy object handling. For this reason, the design process of parallel manipulators is usually supposed to optimize the manipulators' dexterity or structural stiffness. Accordingly, this research intends to analyze the effects of changing actuator and frame-configurations regarding the handling performance of a class of translational 3-RPC-manipulators. Since the RPC-structure can be arranged with translational and/or rotational actuators, several different drive configurations are possible. Additionally, four different frame configurations can be identified, expanding the number of possible manipulators. In order to find an optimized actuator and frame configuration for a given task, this research approach systematically reduces the number of possible configurations by means of validity detection and kinetostatic performance analysis. Hence, it is shown that good performance can be achieved for all frame configurations by selecting a suitable actuator combination.

This paper considers how a multi-limbed robot can carry out manipulation tasks involving simultaneous and compatible end-effector velocity and force goals, while also maintaining quasi-static stance stability. The formulation marries a local optimization process with an assumption of a compliant model of the environment. For purposes of illustration, we first develop the formulation for a single fixed based manipulator arm. Some of the basic kinematic variables we previously introduced for multi-limbed robot mechanism analysis in [1] are extended to accomodate this new formulation. Using these extensions, we provide a novel definition for static equilibrium of multi-limbed robot with actuator limits, and provide general conditions that guarantee the ability to apply arbitrary end-effector forces. Using these extended definitions, we present the local optimization problem and its solution for combined manipulation and stance. We also develop, using the theory of strong alternatives, a new definition and a computable test for quasi-static stance feasibility in the presence of manipulation forces. Simulations illustrate the concepts and method.

It is well known that kinematics can significantly affect the manipulation capabilities of robotic arms, traditionally illustrated by performance indices such as workspace volume, kinematic and force manipulability, and isotropy within the arm workspace. In the case of dual-arm systems and bimanual manipulation tasks, the kinematics effects to the above indices becomes even more apparent. However, in spite of the large number of dual-arm systems developed in the past, there is a little literature on the kinematic design analysis for the development of such systems. Particularly, the effects of configuration/ orientation of the shoulders' placement with respect to the torso structure have not sufficiently studied or considered, while many dual-arm systems with upward and/or forward tilt angle in shoulder base frame have been introduced. This paper addresses this problem and quantifies the effect of shoulders base frame orientation in a dual-arm manipulation system by looking at its effect on several important manipulation indices, such as the overall and common workspace, redundancy, global isotropy, dual-arm manipulability, and inertia ellipsoid index within the common workspace of the two arms. Consequently, a range of upward and forward tilt angles for the shoulder frames is identified for the design of a dual-arm torso system to render the most desired manipulation performance.

A simple and reliable keypoint matching method is proposed in this paper. Our research is motivated by the desire to improve the performance of multi-view geometry (MVG) based verification in visual loop closure detection under significant illumination change, where traditional methods may fail due to their inability to either find a sufficient number of correctly matched keypoints or identify correct underlying camera motion to verify the matches. Our method is inspired by research on the spatial statistics of optical flow. By observing that the displacement of matching keypoints between a pair of images is equivalent to the optical flow under the assumption of small camera motion (which is true in applications such as loop closure detection), we exploit the fact that the displacement of correctly matched keypoints between two images must follow a well-defined distribution. This paves the way to a keypoint matching method that uses this distribution to screen or prune potential matching keypoints, so as to remove the incorrect matches (outliers) and retain the true matches (inliers) without being overly and solely dependent on keypoint descriptors. The proposed method is validated on the outdoor image sequences and shows superior performance to the standard keypoint matching method based on distance ratio test.

This paper presents a probabilistically rigorous method for jointly estimating the shape and kinematic states of arbitrary extended objects. A non-parametric shape model is defined as a set of points sampled from the object surface, and the joint probability density function over the surface samples is estimated recursively over time from lidar data. The presented work is demonstrated for a single maneuvering, non-convex object, highlighting key advantages over existing methods and motivating further development.

Stereo Visual Teach & Repeat (VT&R) is a system for long-range, autonomous route following in unstructured 3D environments. As this system relies on a passive sensor to localize, it is highly susceptible to changes in lighting conditions. Recent work in the optics community has provided a method to transform images collected from a three-channel passive sensor into color-constant images that are resistant to changes in outdoor lighting conditions. This paper presents a lighting-resistant VT&R system that uses experimentally trained color-constant images to autonomously navigate difficult outdoor terrain despite changes in lighting. We show through an extensive field trial that our algorithm is capable of autonomously following a 1km outdoor route spanning sandy/rocky terrain, grassland, and wooded areas. Using a single visual map created at midday, the route was autonomously repeated 26 times over a period of four days, from sunrise to sunset with an autonomy rate (by distance) of over 99.9%. These experiments show that a simple image transformation can extend the operation of VT&R from a few hours to multiple days.

A soft wearable robot, which is an emerging type of wearable robot, can take advantage of tendon-driven mechanisms with a Bowden cable. These tendon-driven mechanisms benefits soft wearable robots because the actuator can be remotely placed and the transmission is very compact. However, it is difficult to compensate the friction along the Bowden cable which makes it hard to control. This study proposes the use of a position-based impedance controller, which is robust to the nonlinear dynamics of the system and provides compliant interaction between robot, human, and environment. Additionally, to eliminate disturbances from unexpected tension of the antagonistic wire arising from friction, this study proposes a new type of slack enabling tendon actuator. It can eliminate friction force along the antagonistic wire by actively pushing the wire while preventing derailment of the wire from the spool.

Last 15 years, a wide range of self-healing (SH) materials has been developed and recently these materials are increasingly used in applications in multiple fields, like the automotive industry and aerospace. However, so far this material technology is not yet explored in robotics. The introduction of these materials in robotics will potentially reduce the over-dimensioning of current robotic systems, leading to lighter systems and eventually to more efficient designs. Compliant elements used in next generation soft robots, can be constructed from available SH-materials, making them able to autonomously heal cuts and perforations caused by sharp objects in unstructured environments. In addition, the use of SH-materials will have a beneficial impact on the life span of robotic components, reducing the required maintenance drastically. This paper presents the innovative concept of implementing a SH-mechanism in compliant actuators, using dynamic covalent polymer network systems based on the reversible Diels-Alder (DA) reaction. For two entirely different compliant actuators, a series elastic actuator (SEA) and a soft pneumatic actuator (SPA), an analysis is presented on the integration of the DA-polymers in the actuator designs. For both actuator types, a prototype was designed, developed and validated.

A new methodology that employs the rate-dependent Prandtl-Ishlinskii model (RDPI) as a model and a compensator is suggested in this study for modeling and compensation of rate-dependent hysteresis nonlinearities of a piezoelectric actuator. The technique employs a restructuration of the model that ignores the need to derive an inverse model, which avoids the additional calculations required to formulate a compensator. The simulation results are presented to demonstrate the effectiveness of the strategy on modeling and compensation of hysteresis nonlinearities at various frequencies. The simulation results were followed by experimental study on a piezoelectric actuator that exhibits rate-dependent hysteresis nonlinearities. The results demonstrate that the proposed methodology can be employed effectively for compensation of rate-dependent hysteresis nonlinearities without developing an inverse model.

To increase the acceptance of autonomous systems in populated environments, it is indispensable to teach them social behavior. We would expect a social robot, which plans its motions among humans, to consider both the social acceptability of its behavior as well as task constraints, such as time limits. These requirements are often contradictory and therefore resulting in a trade-off. For example, a robot has to decide whether it is more important to quickly achieve its goal or to comply with social conventions, such as the proximity to humans, i.e., the robot has to react adaptively to task-specific priorities. In this paper, we present a method for priority-adaptive navigation of mobile autonomous systems, which optimizes the social acceptability of the behavior while meeting task constraints. We learn acceptability-dependent behavioral models from human demonstrations by using maximum entropy (MaxEnt) inverse reinforcement learning (IRL). These models are generative and describe the learned stochastic behavior. We choose the optimum behavioral model by maximizing the social acceptability under constraints on expected time-limits and reliabilities. This approach is evaluated in the context of driving behaviors based on the highway scenario of Levine et al. [1].

In this paper we present a dense visual odometry system for RGB-D cameras performing both photometric and geometric error minimisation to estimate the camera motion between frames. Contrary to most works in the literature, we parametrise the geometric error by the inverse depth instead of the depth, which translates into a better fit of the distribution of the geometric error to the used robust cost functions. We also provide a unified evaluation under the same framework of different estimators and ways of computing the scale of the residuals which can be found spread along the related literature. For the comparison of our approach with state-of-the-art approaches we use the popular dataset from the TUM for RGB-D benchmarking. Our approach shows to be competitive with state-of-the-art methods in terms of drift in meters per second, even compared to methods performing loop closure too. When comparing to approaches performing pure odometry like ours, our method outperforms them in the majority of the tested datasets. Additionally we show that our approach is able to work in real time and we provide a qualitative evaluation on our own sequences showing a low drift in the 3D reconstructions.

For numerous real-world applications teleoperated unmanned guided vehicles (UGVs) can quite successfully assist a human in fulfilling her mission objectives. It is important for the specialists to get an overview of the site quickly and with as intuitive means as possible. Our approach to an intuitive human-machine interface for visually teleoperating UGVs makes use of a spherical camera in combination with the Virtual Reality Head-Mounted Display (HMD) Oculus Rift. The Oculus Rift is equipped with an inertial measurement unit to track the teleoperator's head orientation. With this orientation information, the current field of view is synthesized from the panoramic image coming from the spherical camera allowing a free look for the operator. In this paper, we present the hardware setup and the software system of our free-look HMD approach on our UGV. Our approach allows for multi-view, i.e, several operators can collaborate on a given task. What is more, we augment the spherical image with heading and orientation information of the robot as well as of other viewers. Our preliminary evaluation with a number of untrained user suggests that our free-look HMD offers an intuitive multi-view human-machine interface for teleoperating UGVs.

Real-time dense computer vision and SLAM offer great potential for a new level of scene modelling, tracking and real environmental interaction for many types of robot, but their high computational requirements mean that use on mass market embedded platforms is challenging. Meanwhile, trends in low-cost, low-power processing are towards massive parallelism and heterogeneity, making it difficult for robotics and vision researchers to implement their algorithms in a performance-portable way. In this paper we introduce SLAMBench, a publicly-available software framework which represents a starting point for quantitative, comparable and validatable experimental research to investigate trade-offs in performance, accuracy and energy consumption of a dense RGB-D SLAM system. SLAMBench provides a KinectFusion implementation in C++, OpenMP, OpenCL and CUDA, and harnesses the ICL-NUIM dataset of synthetic RGB-D sequences with trajectory and scene ground truth for reliable accuracy comparison of different implementation and algorithms. We present an analysis and breakdown of the constituent algorithmic elements of KinectFusion, and experimentally investigate their execution time on a variety of multicore and GPU-accelerated platforms. For a popular embedded platform, we also present an analysis of energy efficiency for different configuration alternatives.

Tooth extraction is a typical process in clinical dental operations. Interactive haptic simulation of tooth extraction may provide a useful tool for dental students to learn the correct force pattern and tool posture to accomplish a safe tooth extraction. In this paper, we extended our previous configuration-based optimization approach to simulate the six Degree-of-Freedom (DoF) haptic interaction process of tooth extraction. A multi-phase model was proposed to simulate progressive changes of the connection strength between the target tooth and its surrounding gingiva. An energy accumulation model was proposed to compute the small-scale rotation and translation of the target tooth under active forces from a dental forceps. The proposed approach could support training of coordinated force and motion control skill required for tooth extraction. Experimental results validated the stability and efficiency of the proposed approach to simulate various force-displacement profiles for extracting diversified target teeth.

Identifying the relevant functional degrees of freedom is a key prerequisite for the proper handling of everyday objects. Recognizing and exploiting these degrees of freedom in the context of non-rigid objects poses challenges that are significantly different from the rigid case. As a major generic subtask, we consider the identification and exploitation of boundary components during clothes manipulation, combining RGBD vision with uni- and bi-manual handling through a robot. Specifically, we present a novel graph-based approach to detecting boundary components by extracting closed contours from depth images. Based on that, we suggest a planner minimizing a heuristic energy function for an optimal grasp pose of a robot hand around the boundary of a garment. We demonstrate the effectiveness of the approach in interactive perception and regrasping experiments with a dual arm and two attached anthropomorphic hands. Furthermore, we show how to make use of these capabilities to implement a basic skill for a coat-check robot: hanging up a knit cap on a hat-stand.

This paper presents an intention-aware online planning approach for autonomous driving amid many pedestrians. To drive near pedestrians safely, efficiently, and smoothly, autonomous vehicles must estimate unknown pedestrian intentions and hedge against the uncertainty in intention estimates in order to choose actions that are effective and robust. A key feature of our approach is to use the partially observable Markov decision process (POMDP) for systematic, robust decision making under uncertainty. Although there are concerns about the potentially high computational complexity of POMDP planning, experiments show that our POMDP-based planner runs in near real time, at 3 Hz, on a robot golf cart in a complex, dynamic environment. This indicates that POMDP planning is improving fast in computational efficiency and becoming increasingly practical as a tool for robot planning under uncertainty.