ICRA 2018 论文列表

Effective communication between humans often embeds both temporal and spatial context. While spatial context captures the geographic settings of objects in the environment, temporal context describes their changes over time. In this paper, we propose temporal spatial inverse semantics (TeSIS) to extend the inverse semantics approach to also consider the temporal context for robots communicating with humans. Inverse semantics generates natural language requests while taking into account how well the human listeners would interpret those requests given the current spatial context. Compared to inverse semantics, our approach incorporates also temporal context by referring to spatial context information in the past. To achieve this, we extend the sentence structure in inverse semantics to generate sentences that can refer to not only the current but also previous states of the environment. A new metric based on the extended sentence structure is developed by breaking a single sentence into multiple independent sentences that refer to environment states at different times. Using this approach, we are able to generate sentences such as “Please pick up the cup beside the oven that was on the dining table”. To evaluate our approach, we randomly generate scenarios in an experimental domain. Each scenario includes the description of the current and several immediate previous states. Natural language sentences are then generated for these scenarios using both inverse semantics that uses only the spatial context and our approach. Amazon MTurk is used to compare the sentences generated and results show that TeSIS achieves better accuracy, sometimes by a significant margin, than the baseline.

Using Programming by Demonstration to teach robot learners generalisable skills relies on having effective human teachers. This paper aims to address two problems commonly observed in demonstration data sets that arise due to poor teaching strategies; undemonstrated states and ambiguous demonstrations. Overcoming these issues through the use of visual feedback and simple heuristic rules is investigated as a potential way of guiding novice users to more effectively teach robot learners to generalise a task. The proposed method intends to offer the user a more transparent understanding of the robot learner's model state during the teaching phase, to create a more interactive and robust teaching process. Results from a single-factor, three-phase repeated measures study with n=30 participants, comparing the proposed feedback and heuristic rules set against an unguided condition, show a statistically significant (F(2,58)=7.952,p=0.001) improvement of user teaching efficiency of approximately 180% when using the proposed feedback visualisation.

A novel method for replaying behavior of a mobile robot from its memory of past experiences is presented in this paper. The method is a version of a particle filter on episode (PFoE), which applies a particle filter on the memory so as to efficiently find some similar situations with the current one. Though the original PFoE was proposed as a reinforcement learning method, we once removed the reward system from the original one so as to apply it to task teaching. In the experiment, we gave several kinds of motion to a micromouse type robot with the proposed method through a gamepad. The robot replayed the behaviors robustly with sensor feedback after several number of repetitive teaching.

When performing multi-robot tasks, sensory feedback is crucial in reducing uncertainty for correct execution. Yet the utilization of sensors should be planned as an integral part of the task planning, taken into account several factors such as the tolerance of different inferred properties of the scene and interaction with different agents. In this paper we handle this complex problem in a principled, yet efficient way. We use surrogate predictors based on open-loop simulation to estimate and bound the probability of success for specific tasks. We reason about such task-specific uncertainty approximants and their effectiveness. We show how they can be incorporated into a multi-robot planner, and demonstrate results with a team of robots performing assembly tasks.

In autonomous robotics there are many situations that require solving a motion planning problem to complete a task. A Task Space (T-Space), composed of parameters that define the task being performed, can be a more effective planning space for these problems, however, planning within a T-Space is often computationally challenging. In this paper, we present a novel method to analyse the relationship between T-Space parameters and the pose of manipulator bodies to create a dependency matrix. We then use this information to decompose the motion planning problem into sequential lower complexity sub-problems. We call this approach Task Space Motion Planning Decomposition (TSMPD). This paper introduces TSMPD and quantifies the improvement to planning efficiency on a challenging maze navigation problem and weld path planning problem.

Due to the distributed nature of cooperative simultaneous localization and mapping (CSLAM), detecting inter-robot loop closures necessitates sharing sensory data with other robots. A naïve approach to data sharing can easily lead to a waste of mission-critical resources. This paper investigates the logistical aspects of CSLAM. Particularly, we present a general resource-efficient communication planning framework that takes into account both the total amount of exchanged data and the induced division of labor between the participating robots. Compared to other state-of-the-art approaches, our framework is able to verify the same set of potential inter-robot loop closures while exchanging considerably less data and influencing the induced workloads. We develop a fast algorithm for finding globally optimal communication policies, and present theoretical analysis to characterize the necessary and sufficient conditions under which simpler strategies are optimal. The proposed framework is extensively evaluated with data from the KITTI odometry benchmark datasets.

We present a system to infer and execute a human-readable program from a real-world demonstration. The system consists of a series of neural networks to perform perception, program generation, and program execution. Leveraging convolutional pose machines, the perception network reliably detects the bounding cuboids of objects in real images even when severely occluded, after training only on synthetic images using domain randomization. To increase the applicability of the perception network to new scenarios, the network is formulated to predict in image space rather than in world space. Additional networks detect relationships between objects, generate plans, and determine actions to reproduce a real-world demonstration. The networks are trained entirely in simulation, and the system is tested in the real world on the pick-and-place problem of stacking colored cubes using a Baxter robot.

In this study, we addressed a path planning problem of a mobile robot to construct highly accurate 3D models of an unknown environment. Most studies have focused on exploration approaches, which find the most informative viewpoint or trajectories by analyzing a volumetric map. However, the completion of a volumetric map does not necessarily describe the completion of a 3D model. A highly complicated structure sometimes cannot be represented as a volumetric model. We propose a novel exploration algorithm that considers not only a volumetric map but also reconstructed surfaces. Unlike previous approaches, we evaluate the model completeness according to the quality of the reconstructed surfaces and extract low-confidence surfaces. The surface information is used to guide the computation of the exploration path. Experimental results showed that the proposed algorithm performed better than other state-of-the-art exploration methods and especially improved the completeness and confidence of the 3D models.

Surveying 3D scenes is a common task in robotics. Systems can do so autonomously by iteratively obtaining measurements. This process of planning observations to improve the model of a scene is called Next Best View (NBV) planning. NBV planning approaches often use either volumetric (e.g., voxel grids) or surface (e.g., triangulated meshes) representations. Volumetric approaches generalise well between scenes as they do not depend on surface geometry but do not scale to high-resolution models of large scenes. Surface representations can obtain high-resolution models at any scale but often require tuning of unintuitive parameters or multiple survey stages. This paper presents a scene-model-free NBV planning approach with a density representation. The Surface Edge Explorer (SEE) uses the density of current measurements to detect and explore observed surface boundaries. This approach is shown experimentally to provide better surface coverage in lower computation time than the evaluated state-of-the-art volumetric approaches while moving equivalent distances.

Active lower limb prosthetics rely on the detection of gait mode to direct controller response. The majority of systems require feedback from the prosthetic and/or inertial measurement units (IMUs). Reliance on movement delays classification, reducing the range of patient activities and terrain traversed. Neuromuscular interfaces using electromyography (EMG) enable real-time monitoring by registering user intent, however EMG has known robustness issues out-of-clinic that have impeded its translation. Furthermore, supervised training of gait classifiers can require large subject-specific amputee data sets which are difficult to obtain. Mechanomyography (MMG) has shown less dependence on environmental conditions than EMG yet has seen limited use in this realm. In this investigation we introduce an MMG gait classifier targeting improved control of prosthetic (robotic) legs. We compare the accuracy of subject specific classifiers to those trained using subject-independent pooling. Additionally, we quantify the effect of introducing a small amount of data from individual test subjects to the training pool. Experiments were performed on 12 participants and 5 gait modes. A support vector machine (SVM) classifier achieved 65% accuracy with subject-specific data, 92% with pooled training data, and 94% with pooled plus limited user-specific data. The results show the promise of MMG gait classifiers with increased robustness and reduced subject-specific training in prosthetic control.

Recently, the demand to manipulate deformable objects such as a string and cloth by robots is growing. The reason is that it has the possibility of making our lives more convenient in many domains. The manipulation of deformable objects, however, is more difficult than that of rigid objects, because deformable objects have diversity of shape and behavior. Therefore, our research group has been focusing on the string shape operation. This paper describes planning method of string untying operation based on knot theory and algorithms to generate the motion of a manipulator. The novel contribution of our planning method is automatic selection of optimal shape operation based on cost function. At final, the results of string untying experiments are reported.

We present an approach for achieving a natural, efficient gait on bipedal robots using straightened legs and toe-off. Our algorithm avoids complex height planning by allowing a whole-body controller to determine the straightest possible leg configuration at run-time. The controller solutions are biased towards a straight leg configuration by projecting leg joint angle objectives into the null-space of the other quadratic program motion objectives. To allow the legs to remain straight throughout the gait, toe-off was utilized to increase the kinematic reachability of the legs. The toe-off motion is achieved through underconstraining the foot position, allowing it to emerge naturally. We applied this approach of under-specifying the motion objectives to the Atlas humanoid, allowing it to walk over a variety of terrain. We present both experimental and simulation results and discuss performance limitations and potential improvements.

Jamming of granular media has been shown to possess the property of stiffness variation, transitioning from a soft to a quasi-solid state depending on the packing density of the granules. Recently, a gradual stiffness change for bending has been reported by using compliant, cubic shaped granules. Here we demonstrate that the same method and material also exhibits a gradual stiffness change for compression. As a potential application of “compliant jamming”, a bio-inspired robotic platform with jamming membranes as end effectors is designed and tasked with climbing straight vertical shafts. First, the benefit of varying the bending stiffness is investigated in the climbing task by measuring the friction force that the end effectors are applying to the shaft walls, especially if the walls are irregularly shaped. Then the role of compressive stiffness variation is explored by analyzing the performance of the robot in the climbing task, showing multi-modal properties of jamming membranes: (i) enabling a pressure sensor to detect the shaft walls, acting (ii) as grippers that actively use the irregularities of the walls to climb up by state-switching the granular material and (iii) as force dissipators that can dissipate internal forces caused by closed kinematic chains.

Stiffness modulators are devices that promote a novel means of actuation; they provide stiffness modulation without deliberately doing mechanical work. These type of compliant actuators may be used for human augmentation to complement co-contracted antagonistic muscles and as such reduce muscle activity and metabolic energy cost. Despite the theoretical appeal of this concept, its implementation remains elusive in practical applications. This is particularly true for human augmentation which requires a portable stiffness modulator. In this paper, we present a compact, lightweight, and self-contained stiffness modulator. Using this device, we demonstrate stiffness augmentation of the human knee joint in a sit to stand task. The experimental results indicate that the proposed device is able to assist a human by reducing muscle activity while drawing minimal battery power.

We present a novel systematic design framework for general under-actuated tendon-driven (UATD) robotic systems to exhibit desired behaviors both during the free motion and the contact task. For this, we propose stiffness decomposition, which enables us to completely decompose the configuration space of the UATD robotic systems into the actuated space (with full actuation via active tendons) and the un-actuated space (with no actuation, only with passive compliance and contact wrench). The behavior in the actuated space is then fully-controllable, thus, the attainment of the desired behaviors, particularly those during the contact task, hinges upon that in the un-actuated space. For this, relying on the stiffness decomposition, we optimize the design parameters (e.g., tendon routing, pulley radius, passive compliance, etc.) to ensure the deformation in the un-actuated space as directional (e.g., for adaptive grasping) and minimized (e.g., pushing with posture maintained) for different contact wrench sets as possible, while also rendering the free motion to be as compliant and backdrivable as possible. The presented framework is then applied to design a UATD robotic finger and experimentally verified with the robot able to mimic the behavior of human index finger both during the free motion and pinch-pushing.

Human performers have developed impressive acrobatic techniques over thousands of years of practicing the gymnastic arts. At the same time, robots have started to become more mobile and autonomous, and can begin to imitate these stunts in dramatic and informative ways. We present a simple two degree of freedom robot that uses a gravity-driven pendulum launch and produces a variety of somersaulting stunts. The robot uses an IMU and a laser range-finder to estimate its state mid-flight and actuates to change its motion both on and and off the pendulum. We discuss the dynamics of this behavior in a framework of acrobatic capability and present experimental results.

Dynamic environments are challenging for visual SLAM as moving objects can impair camera pose tracking and cause corruptions to be integrated into the map. In this paper, we propose a method for robust dense RGB-D SLAM in dynamic environments which detects moving objects and simultaneously reconstructs the background structure. While most methods employ implicit robust penalisers or outlier filtering techniques in order to handle moving objects, our approach is to simultaneously estimate the camera motion as well as a probabilistic static/dynamic segmentation of the current RGB-D image pair. This segmentation is then used for weighted dense RGB-D fusion to estimate a 3D model of only the static parts of the environment. By leveraging the 3D model for frame-to-model alignment, as well as static/dynamic segmentation, camera motion estimation has reduced overall drift - as well as being more robust to the presence of dynamics in the scene. Demonstrations are presented which compare the proposed method to related state-of-the-art approaches using both static and dynamic sequences. The proposed method achieves similar performance in static environments and improved accuracy and robustness in dynamic scenes.

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.

Autonomous vehicles promise safer roads, energy savings, and more efficient use of existing infrastructure, among many other benefits. Although the effect of autonomous vehicles has been studied in the limits (near-zero or full penetration), the transition range requires new formulations, mathematical modeling, and control analysis. In this article, we study the ability of small numbers of autonomous vehicles to stabilize a single-lane system of human-driven vehicles. We formalize the problem in terms of linear string stability, derive optimality conditions from frequency-domain analysis, and pose the resulting nonlinear optimization problem. In particular, we introduce two conditions which simultaneously stabilize traffic while imposing a safety constraint on the autonomous vehicle and limiting degradation of performance. With this optimal linear controller in a system with typical human driver behavior, we can numerically determine that only a 6% uniform penetration of autonomously controlled vehicles (i.e. one per string of up to 16 human-driven vehicles) is necessary to stabilize traffic across all traffic conditions.