IROS 2017 论文列表

Climbing robots have many potential applications including maintenance, monitoring, search and rescue, and self-assembly. While numerous climbing designs have been investigated, most are limited to stiff components. Flippy (Fig. 1) is a small, flipping biped robot with a soft, flexible body and on-board power and control. Due to its built-in compliance, flipping gait, and corkscrew gripper, it can autonomously climb up and down surfaces held at any angle relative to gravity and transition from one surface to another, without complex sensing or control. In this paper, we demonstrate the robot's ability to flip consistently over a flat Velcro surface and 2D Velcro track, where it reliably climbs vertically, upside down and back to a flat surface, completing all the interior transitions in-between.

Programming robots to do manipulation tasks requires users to specify relevant perceptual landmarks, which include objects, parts of objects, or parts of the workspace. While many techniques have been developed for object detection, few are designed to detect arbitrary parts of objects or of the workspace. This paper presents CustomLandmarks, a flexible tool that lets non-roboticists build their own perceptual detectors for many kinds of landmarks. The system components include a simple 3D interface for specifying landmarks, a novel representation for landmarks, and an algorithm for locating landmarks in new scenes. We evaluate the system's detection performance through systematic experiments and by using the system to aid a PR2 robot with several manipulation tasks. Finally, we present a user study showing that novices to the system are able to understand and use CustomLandmarks quickly, creatively, and effectively.

We present a flexible ultrasonic motor that generates linear motion as a novel soft actuator. This motor is comprised of a single cube stator with a through-hole, and a flexible coil spring inserted into the hole. The coil spring, which has a diameter of 10 mm and length of 200 mm, is translated by a vibration generated in the stator. A preload between the stator and the coil spring is important for improving the motor performance. The coil spring inserted into the stator hole generates the preload by expanding in a radial direction. We build a prototype flexible ultrasonic motor, and examine its basic performance parameters such as thrust force, velocity, response time, resolution and flexibility. To demonstrate flexibility, we show the stable thrust force and velocity even when the coil spring is bent to a bending radius of 15 mm.

The demands from precision agriculture (PA) for high-quality information at the individual plant level require to re-think the approaches exploited to date for remote sensing as performed by unmanned aerial vehicles (UAVs). A swarm of collaborating UAVs may prove more efficient and economically viable compared to other solutions. To identify the merits and limitations of a swarm intelligence approach to remote sensing, we propose here a decentralised multi-agent system for a field coverage and weed mapping problem, which is efficient, intrinsically robust and scalable to different group sizes. The proposed solution is based on a reinforced random walk with inhibition of return, where the information available from other agents (UAVs) is exploited to bias the individual motion pattern. Experiments are performed to demonstrate the efficiency and scalability of the proposed approach under a variety of experimental conditions, accounting also for limited communication range and different routing protocols.

Learning complex robot manipulation policies for real-world objects is challenging, often requiring significant tuning within controlled environments. In this paper, we learn a manipulation model to execute tasks with multiple stages and variable structure, which typically are not suitable for most robot manipulation approaches. The model is learned from human demonstration using a tactile glove that measures both hand pose and contact forces. The tactile glove enables observation of visually latent changes in the scene, specifically the forces imposed to unlock the child-safety mechanisms of medicine bottles. From these observations, we learn an action planner through both a top-down stochastic grammar model (And-Or graph) to represent the compositional nature of the task sequence and a bottom-up discriminative model from the observed poses and forces. These two terms are combined during planning to select the next optimal action. We present a method for transferring this human-specific knowledge onto a robot platform and demonstrate that the robot can perform successful manipulations of unseen objects with similar task structure.

We present a novel motion planning algorithm for pouring a liquid body from a source to a target container. Our approach uses a receding-horizon optimization strategy that considers liquid dynamics and various other constraints. To handle liquid dynamics without costly fluid simulations, we use a neural network to infer a set of key liquid-related parameters from the observation of the current liquid configuration. To train the neural network, we generate a dataset of successful pouring examples using stochastic optimization in a problem-specific search space. These parameters are then used in the objective function for trajectory optimization. Our feedback motion planner achieves real-time performance, and we observe a high success rate in our simulated 2D and 3D liquid pouring benchmarks.

The gap between our ability to collect interesting data and our ability to analyze these data is growing at an unprecedented rate. Recent algorithmic attempts to fill this gap have employed unsupervised tools to discover structure in data. Some of the most successful approaches have used probabilistic models to uncover latent thematic structure in discrete data. Despite the success of these models on textual data, they have not generalized as well to image data, in part because of the spatial and temporal structure that may exist in an image stream. We introduce a novel unsupervised machine learning framework that incorporates the ability of convolutional autoencoders to discover features from images that directly encode spatial information, within a Bayesian nonparametric topic model that discovers meaningful latent patterns within discrete data. By using this hybrid framework, we overcome the fundamental dependency of traditional topic models on rigidly hand-coded data representations, while simultaneously encoding spatial dependency in our topics without adding model complexity. We apply this model to the motivating application of high-level scene understanding and mission summarization for exploratory marine robots. Our experiments on a seafloor dataset collected by a marine robot show that the proposed hybrid framework outperforms current state-of-the-art approaches on the task of unsupervised seafloor terrain characterization.

The use of flexure hinges offers significant benefits compared to conventional revolute joints, such as the use for compact monolithic designs of flexural structures, no friction losses and reduced assembly and manufacturing costs. The combination of this design concept with additive manufacturing (AM) techniques provides huge potential for kinematics design, because of the ability to manufacture complex kinematic structures in one part with integrated adjustable functional elements, i.e. flexure hinges. Considering that there is sparse knowledge of thin-walled structures fabricated using additive manufacturing in current literature, it is obvious that the design of flexure hinge mechanisms for soft robotic applications is challenging today. Therefore, this paper describes software for simulation of plane flexure hinge structures based on investigations on the material characteristics of laser sintered polyamide, especially fatigue strength. Furthermore, the software enables automated design of these flexural structures (e.g. manipulator structures) as STL files, representing the standard interface to all additive manufacturing technologies.

This paper introduces Salient Depth Partitioning (SDP), a depth-based region cropping algorithm devised to be easily adapted to existing detection algorithms. SDP is designed to give robots a better sense of visual attention, and to reduce the processing time of pedestrian detectors. In contrast to proposal generators, our algorithm generates sparse regions, to combat image degradation caused by robot motion, making them more suitable for real-world operation. Furthermore, SDP is able achieve real-time performance (77 frames per second) on a single processor without a GPU. Our algorithm requires no training, and is designed to work with any pedestrian detection algorithm, provided that the input is in the form of a calibrated RGB-D image. We tested our algorithm with four state-of-the-art pedestrian detectors (HOG and SVM [1], Aggregate Channel Features [2], Checkerboards [3], and R-CNN [4]), and show that it improves computation time by up to 30%, with no discernible change in accuracy.

This paper presents a novel technique that allows for both computationally fast and sufficiently plausible simulation of vehicles with non-deformable tracks. The method is based on an effect we have called Contact Surface Motion. A comparison with several other methods for simulation of tracked vehicle dynamics is presented with the aim to evaluate methods that are available off-the-shelf or with minimum effort in general-purpose robotics simulators. The proposed method is implemented as a plugin for the open-source physics-based simulator Gazebo using the Open Dynamics Engine.

This paper deals with the problem of detecting fallen people lying on the floor by means of a mobile robot equipped with a 3D depth sensor. In the proposed algorithm, inspired by semantic segmentation techniques, the 3D scene is over-segmented into small patches. Fallen people are then detected by means of two SVM classifiers: the first one labels each patch, while the second one captures the spatial relations between them. This novel approach showed to be robust and fast. Indeed, thanks to the use of small patches, fallen people in real cluttered scenes with objects side by side are correctly detected. Moreover, the algorithm can be executed on a mobile robot fitted with a standard laptop making it possible to exploit the 2D environmental map built by the robot and the multiple points of view obtained during the robot navigation. Additionally, this algorithm is robust to illumination changes since it does not rely on RGB data but on depth data. All the methods have been thoroughly validated on the IASLAB-RGBD Fallen Person Dataset, which is published online as a further contribution. It consists of several static and dynamic sequences with 15 different people and 2 different environments.

As agile robots learn to run and hop, we believe they will soon also perform large jumps and spins that require graceful landing. To land properly, they will need to be aware of their relative body orientation during free fall and make appropriate adjustments. In this paper, we examine the sensing and actuation necessary for a mechanism to alter its fall under a high rate of spin. We also present a simple new robot, the Binary Robotic Inertially Controlled bricK (BRICK). This device combines information from an accelerometer, a gyroscope, and a laser range-finder to track its orientation, and adjusts spin mid-flight by rapidly changing its moment of inertia between two states. An error analysis shows mechanical inaccuracies and uncertainty to be the largest source of variation. Nevertheless, our mechanism is able to fall from arbitrary heights, with arbitrary spin, and always land in a desired orientation.

We present a class of tendon-actuated soft robots, which promise to be low-cost and accessible to non-experts. The fabrication techniques we introduce are largely based on traditional techniques for fabricating plush toys, and so we term the robots created using our approach “plush robots.” A plush robot moves by driving internal winches that pull in (or let out) tendons routed through its skin. We provide a forward simulation model for predicting a plush robot's deformation behavior given some contractions of its internal winches. We also leverage this forward model for use in an interactive control scheme, in which the user provides a target pose for the robot, and optimal contractions of the robot's winches are automatically computed in real-time. We fabricate two examples to demonstrate the use of our system, and also discuss the design challenges inherent to plush robots.

This paper introduces a new concept of building soft pneumatic structures by assembling modular units of fabric-based rotary actuators (FRAs) and beams. Upon pressurization, the inner folds of FRA would expand, which causes the FRA module to unfold, generating angular displacement. Hence, FRAs would enable mobility function of the structure and its range of motion. The modular nature of the actuator units enables customized configuration of pneumatic structures, which can be modified and scaled by selecting the appropriate modules. FRAs are also designed to be bladder-less, that is, they are made without an additional layer of inner bladder. Thus, a simple fabrication process can be used to prepare the actuators. In this paper, we studied how the performance of the FRA modules changes with their dimensions and demonstrated how a soft gripper can be constructed using these modules. The kinematic response of the actuator segments of the gripper was analyzed and a pressure control algorithm was developed to regulate the pneumatic pressure of the actuator. The modular based soft robotic gripper alone weighs about 140g. Yet, based on the grip tests, it is able to lift heavier objects (up to 2.4kg), achieving a high payload-to-weight ratio of about 1714%, which is higher than the value reported by previously developed soft pneumatic grippers using elastomeric materials. Lastly, we also demonstrated that the gripper is capable of performing two essential grasping modes, which are power grasping and fingertip grasping, for objects of various shapes.

This paper presents FROST, an open-source MATLAB toolkit for modeling, trajectory optimization and simulation of hybrid dynamical systems with a particular focus in dynamic locomotion. The design objective of FROST is to provide a unified software environment for developing model-based control and motion planning algorithms for robotic systems whose dynamics is hybrid in nature. In particular, FROST uses directed graphs to describe the underlying discrete structure of hybrid system models, which renders it capable of representing a wide variety of robotic systems. Equipped with a custom symbolic math toolbox in MATLAB using Wolfram Mathematica, one can rapidly prototype the mathematical model of robot kinematics and dynamics and generate optimized code of symbolic expressions to boost the speed of optimization and simulation in FROST. In favor of agile and dynamic behaviors, we utilize virtual constraint based motion planning and feedback controllers for robotic systems to exploit the full-order dynamics of the model. Moreover, FROST provides a fast and tractable framework for planning optimal trajectories of hybrid dynamical systems using advanced direct collocation algorithms. FROST has been successfully used to synthesize dynamic walking in multiple bipedal robots. Case studies of such applications are considered in this paper, wherein different types of walking gaits are generated for two specific humanoid robots and validated in simulation.

Robotics in the last decades is moving towards bioinspired solutions in order to develop systems increasingly integrated with the human environment. Among them, legged robots fascinates more and more researchers thanks to their ability of moving in unstructured environment such as the ones typical of earthquakes, where in the near future robots are planned to be send to help humans while performing dangerous tasks. On the base of this findings, the authors propose a novel concept for a jumping humanoid leg, based on the key role played from the structural flexibility. The geometric and dynamic features of this leg have been selected thanks to a targeted set of numerical simulations. An extensive campaign of experimental tests useful for the validation of the numerical model here presented will be a matter of future works.

Navigation for autonomous underwater vehicles (AUVs) in mid-depth water column is notoriously challenging due to the lack of features for reference. In this paper, we propose a flow-aided cooperative navigation (FACON) strategy to improve the navigation performance of a team of AUVs when neither frequent surfacing nor persistent bottom-locking is available. Preloaded ocean current forecasts are referenced by each individual vehicle as it performs dead-reckoning with an inertial navigation system and measures relative current velocities during navigation. A marginalized particle filter is applied by each AUV to track its location, velocity, sensor biases, and local flow perturbation unresolved by the ocean forecast. Meanwhile, AUVs perform distributed cooperative localization based on relative pose measurements and asynchronized local communication when they enter the communication range of one another. Opportunistic information fusion among AUVs is realized through covariance intersection. The performance of the flow-aided navigation scheme of a single AUV is analyzed within a simulated experiment based on field test data. The feasibility of FACON is discussed through simulation in a turbulent, multi-gyre flow field.

Most state estimation procedures in mobile robotics require information about the locations of the individual sensors on the platform. In this paper, we study the motion-based multi-sensor extrinsic calibration problem and point out an overlooked defect of traditional least squares estimation in this context. We present a novel calibration approach based-on the Gauss-Helmert estimation paradigm, together with a formulation for multi-sensor motion constraint. Our approach estimates not only the extrinsic parameters but also the pose observation errors, thus recovering the underlying sensor movements that exactly fulfill the motion constraints. Compared to traditional least squares approaches that estimate only the parameters, our approach is statistically optimal, thus is more accurate and robust. We implemented our approach and tested it on real robot. The experiments show that our approach is able to accurately determine the extrinsic configuration of each sensor and can largely improve the accuracy when the noise level is high.

Motor synergies are an important concept in human motor control. Through the co-activation of multiple muscles, complex motion involving many degrees-of-freedom can be generated. However, leveraging this concept in robotics typically entails using human data that may be incompatible for the kinematics of the robot. In this paper, our goal is to enable a robot to identify synergies for low-dimensional control using trial-and-error only. We discuss how synergies can be learned through latent space policy search and introduce an extension of the algorithm for the re-use of previously learned synergies for exploration. The application of the algorithm on a bimanual manipulation task for the Baxter robot shows that performance can be increased by reusing learned synergies intra-task when learning to lift objects. But the reuse of synergies between two tasks with different objects did not lead to a significant improvement.