ICRA 2019 论文列表

Soft robotics has yielded numerous examples of soft grippers that utilize compliance to achieve impressive grasping performances with great simplicity, adaptability, and robustness. Designing soft grippers with substantial grasping strength while remaining compliant and gentle is one of the most important challenges in this field. In this paper, we present a light-weight, vacuum-driven soft robotic gripper made of an origami “magic-ball” and a flexible thin membrane. We also describe the design and fabrication method to rapidly manufacture the gripper with different combinations of low-cost materials for diverse applications. Grasping experiments demonstrate that our gripper can lift a large variety of objects, including delicate foods, heavy bottles, and other miscellaneous items. The grasp force on 3D-printed objects is also characterized through mechanical load tests. The results reveal that our soft gripper can produce significant grasp force on various shapes using negative pneumatic pressure (vacuum). This new gripper holds the potential for many practical applications that require safe, strong, and simple grasping.

In adult laparoscopy, robot-aided surgery is a reality in thousands of operating rooms worldwide, owing to the increased dexterity provided by the robotic tools. Many robots and robot control techniques have been developed to aid in more challenging scenarios, such as pediatric surgery and microsurgery. However, the prevalence of case-specific solutions, particularly those focused on non-redundant robots, reduces the reproducibility of the initial results in more challenging scenarios. In this paper, we propose a general framework for the control of surgical robotics in constrained workspaces under teleoperation, regardless of the robot geometry. Our technique is divided into a slave-side constrained optimization algorithm, which provides virtual fixtures, and with Cartesian impedance on the master side to provide force feedback. Experiments with two robotic systems, one redundant and one non-redundant, show that smooth teleoperation can be achieved in adult laparoscopy and infant surgery.

This paper presents a novel framework for simultaneously implementing localization and segmentation, which are two of the most important vision-based tasks for robotics. While the goals and techniques used for them were considered to be different previously, we show that by making use of the intermediate results of the two modules, their performance can be enhanced at the same time. Our framework is able to handle both the instantaneous motion and long-term changes of instances in localization with the help of the segmentation result, which also benefits from the refined 3D pose information. We conduct experiments on various datasets, and prove that our framework works effectively on improving the precision and robustness of the two tasks and outperforms existing localization and segmentation algorithms.

This paper presents a unified motion planning approach for an Intervention Autonomous Underwater Vehicle (I-AUV) in a cluttered environment with localization uncertainty. With the uncertainty being propagated by an information filter, a trajectory optimization problem closed by a Linear-Quadratic-Gaussian controller is formulated for a coupled design of optimal trajectory, localization, and control. Due to the presence of obstacles or complexity of the cluttered environment, a set of feasible initial I-AUV trajectories covering multiple homotopy classes are required by optimization solvers. Parameterized through polynomials, the initial base trajectories are from solving quasi-quadratic optimization problems that are linearly constrained by waypoints from RRTconnect, while the initial trajectories of the manipulator are generated by a null space saturation controller. Simulations on an I-AUV with a 3 DOF manipulator in cluttered underwater environments demonstrated that initial trajectories are generated efficiently and that optimal and collision-free I-AUV trajectories with low state uncertainty are obtained.

Modern robotic systems are very complex and need to be tested in simulations with detailed sensor noise models to effectively verify robotic behavior. Depth imagery in particular comes with significant noise in the form of scene-dependent pixel-wise dropouts and distortions. Unfortunately, many depth camera simulations contain limited noise models, or can only support generating realistic depth images of simple scenes, which limits their usefulness in effectively testing perception algorithms. We propose a data driven approach to generate more realistic noise for complex simulated environments by using a convolutional neural network (CNN) to predict which pixels of a simulated noise-free depth image will not have returns (no-depth-return pixels, or NDP). We choose to focus on NDP here, as these dropouts are the most common and dramatic form of depth image noise. To train this network, we use reconstructed real-world scenes from the Label Fusion dataset to provide ground truth depth for each noisy depth image used to scan the scene. We use the resulting noise-free and noisy depth image pairs as labeled examples and train the network to predict which pixels of the noise-free image will be NDP. When used to post-process a simulation of a depth sensor, this system produces realistic depth images, even in cluttered scenes. To demonstrate that our approach successfully closes the reality gap for depth imagery, we show that the popular ICP algorithm for object pose estimation fails more realistically on our CNN-corrupted simulated depth images than on uncorrupted depth images and unsupervised domain adaptation baselines.

This study presents an impedance controller for series elastic actuators (SEAs), using the singular perturbation (SP) theory and time-delay estimation (TDE) technique. While the SP theory attenuates the requirement for states to be measured, the TDE technique eliminates the requirement for identifying system parameters. Through a numerical analysis and experimental validation, we demonstrate that the proposed controller produces satisfactory tracking performance while-at the same time-pursues wider operational bandwidth and lower driving-point impedance.

Compliant robotic grippers are more robust to uncertainties in grasping and manipulation tasks, especially when paired with tactile and proprioceptive feedback. Although considerable progress has been made towards achieving proprioceptive soft robotic grippers, current efforts require complex driving hardware or fabrication techniques. In this paper, we present a simple scalable soft robotic gripper integrated with high-deformation strain and pressure sensors. The gripper is composed of structurally-compliant handed shearing auxetic structures actuated by electric motors. Coupling deformable sensors with the compliant grippers enables gripper proprioception and object classification. With this sensorized system, we are able to identify objects' size to within 33% of actual radius and sort objects as hard/soft with 78% accuracy.

Correlation filters have provided exceptional results in the field of visual object tracking in the past few years. However, these methods typically learn a single filter to be robust to many different appearance changes, which can be challenging. We propose a simple solution to this problem by utilizing an ensemble method of base trackers trained on different temporal windows of the video history. The proposed tracker, called MTCF, exhibits the following features: i) it can be trained using gradient-based convex optimization; ii) it is robust to short-term and long-term changes in visual appearance. MTCF performs on par with or outperforms state-of-the-art trackers on the OTB and the VOT benchmark datasets. We present an extensive analysis of the performance of MTCF on these benchmark datasets.

Human Robot interaction is a fundamental perquisite for any robot performing a physical task in collaboration with a human. The presence of disturbances arising from the partially known tasks payloads, the unexpected interaction forces in general, and the uncertainty in the interpretation of the human intention in terms of motions and forces can pose significant challenges and eventually compromise the execution of the collaborative task. This work presents a novel, intrinsically adaptable multimodal (force, motion and verbal) interaction framework for human-robot collaboration (HRC) that leverages on an online self-tuning stiffness regulation principle to provide adaptation to interaction/payload forces and reject disturbances arising by unexpected interaction loads. Besides the presented method, it enables the rejection of unnecessary motion commands (e.g. oscillations generated by the human operator) to reach the robot co-worker through the filtering of the human generated motions, that are outside the range (in terms of speed and acceleration) of the envisioned manipulation manoeuvres. Finally, a verbal interaction channel allows the operator to convey securely his high level intentions and to control the states of the task execution. We evaluated and demonstrated the effectiveness of the proposed multimodal interaction framework in a high weight carrying human-robot collaboration task using the humanoid robot COMAN +.

Robust motion planning is a well-studied problem in the robotics literature, yet current algorithms struggle to operate scalably and safely in the presence of other moving agents, such as humans. This paper introduces a novel framework for robot navigation that accounts for high-order system dynamics and maintains safety in the presence of external disturbances, other robots, and humans. Our approach precomputes a tracking error margin for each robot, generates confidence-aware human motion predictions, and coordinates multiple robots with a sequential priority ordering, effectively enabling scalable safe trajectory planning and execution. We demonstrate our approach in hardware with two robots and two humans, and showcase scalability in a larger simulation.

Snake-like robots are commonly used in Minimally Invasive Surgery as they are able to reach areas deep inside the human body. These robots have instruments that are deployed out of the robot's head and controlled via tendons, which connect the instrument to motors at the proximal end. In most currently available systems the instruments are lacking a rolling motion of the end-effector.In this paper, we present a new instrument prototype for a snake-like robot that can perform a stable in-place rolling motion. The prototype has a diameter of 4mm, uses 13 tendons and has 6 degrees of freedom. The robot can bend and roll to high angles, and strongly improves the dexterity compared to an instrument without rolling capabilities. In the evaluation we show that the rolling-tip gripper can rotate about 165° and is capable of applying forces up to 6.5N.

Linear Series Elastic Actuators exhibit a restricted design space. This inevitably leads to design trade-offs translating into robot performance limitations. These prevent robots from eventually reaching human comparable soft but also powerful physical interaction performance.This work presents a novel fixed passive rolling flexure design principle enabling the realization of a wide range of progressive torque-deflection characteristics. The proposed principle displays low hysteresis and can be manufactured in single 2D components. The paper derives the analytic foundation for the rolling flexure principle and is supported by numerical finite element analysis. The theory is validated by experimental results obtained on two laboratory prototypes.

This paper presents a robust controller for uncertain robot manipulators subject to disturbances which are composed of sinusoids. The controller employs the disturbance observer based controller which can effectively estimate and compensate the effect of plant uncertainties and the disturbances. Assuming that the frequencies of sinusoids are known, we embed the internal model of disturbances into the proposed controller so that the design parameters of the controller can be chosen without using the magnitude of disturbance or its time derivative. A rigorous stability analysis shows that the closed-loop system under the proposed controller behaves like the nominal closed-loop system free of disturbances. Simulation results for a 2-DOF manipulator show the effectiveness of the proposed controller.

Extrusion-based additive manufacturing (AM), also known as fused deposition modeling (FDM) extrudes filaments through a heated nozzle and builds a part layer-by-layer. Using a smaller diameter nozzle can achieve better surface finish. However, there is a trade-off between surface finish and build times as using a small diameter nozzle leads to smaller layer thickness and long build times. Traditional FDM printers create a part with planar layers, and this restricts control over fiber orientations. This paper presents a robotic cell for multi-resolution AM. The cell consists of two 6 degrees of freedom (DOF) robot manipulators capable of printing non-planar and/or planar layers. We describe algorithms for decomposing parts into multi-resolution layers and generating collision-free trajectories for the robot manipulators. We validate our approach by printing five parts with multi-resolution.

In this work, reinforcement learning is used to develop a position controller for an underactuated nature-inspired Unmanned Aerial Vehicle (UAV). This particular configuration of UAVs achieves lift by spinning its entire body contrary to standard multi-rotors or fixed-wing aircraft. Deep Deterministic Policy Gradients (DDPG) with Ape-X Distributed Prioritized Experience Replay was used to train neural network function approximators that were implemented as the final control policy. The reinforcement learning agent was trained in simulations and directly ported over to real-life hardware. Position control tests were performed on the learned control policy and compared to a baseline PID controller. The learned controller was found to exhibit better control over the inherent oscillations that arise from the non-linear dynamics of the platform.

Inherent compliance plays an enabling role in soft robots, which rely on it to mechanically conform to the environment. However, it also limits the payload of the robots. Various variable stiffness approaches have been adopted to limit compliance and provide structural stability, but most of them can only achieve stiffening of discrete fixed regions which means compliance cannot be precisely adjusted for different needs. This paper offers an approach to enhance the payload with finely adjusted compliance for different needs. We have developed a manipulator that incorporates a novel variable stiffness mechanism and a sliding layer mechanism. The variable stiffness mechanism can achieve a 6.4 stiffness changing ratio with a miniaturized size (10 mm diameter for the testing prototype) through interlocking jamming layers with a honeycomb core. The sliding layer mechanism can actively shift the position of the stiffening regions through sliding of jamming layers. A model to predict the robot shape is derived with verifications via an experiment. The stiffening capacity of the variable stiffness mechanism is also empirically evaluated. A case study of a potential application in laparoscopic surgeries is showcased. The payload of the manipulator is investigated, and the prototype shows up to 57.8 percentage decrease of the vertical deflection due to an external load after reconfigurations.

Augmented reality (AR) is a promising technology where the surgeon can see the medical abnormality in the context of the patient. It makes the anatomy of interest visible to the surgeon which otherwise is not visible. It can result in better surgical precision and therefore, potentially better surgical outcomes and faster recovery times. Despite these benefits, the current AR systems suffer from two major challenges; first, incorrect depth perception and, second, the lack of suitable evaluation systems. Therefore, in the current paper we addressed both of these problems. We proposed a color depth encoding (CDE) technique to estimate the distance between the tumor and the tissue surface using a surgical instrument. We mapped the distance between the tumor and the tissue surface to the blue-red color spectrum. For evaluation and interaction with our AR technique, we propose to use a virtual surgical instrument method using the CAD model of the instrument. The users were asked to reach the judged distance in the surgical field using the virtual tool. Realistic tool movement was simulated by collecting the forward kinematics joint encoder data. The results showed significant improvement in depth estimation, time for task completion and confidence, using our CDE technique with and without stereo versus other two cases, that are, Stereo-No CDE and No Stereo-No CDE.

Driving is a complex task that requires the perception of the surrounding environment, decision making and control of the vehicle. Human drivers predict how surrounding objects move and decide an appropriate driving behavior. As with human drivers, autonomous driving vehicles should consider the condition of the surrounding environment and behave naturally so as not to disturb the traffic flow. We propose a reward function for learning how natural the driving is based on the hypothesis that the movement of surrounding vehicles becomes unpredictable when the ego vehicle takes an unnatural driving behavior. The reward function is based on the prediction error of a deep predictive network that models the transition of the surrounding environment. Occupancy grid image is used to perceive the surrounding environment and the predictions up to two seconds are used to calculate the reward function. We evaluated the reward function using both simulated and the real world data. We trained the prediction network using real driving data and trained a reinforcement learning agent based on the reward function. Then we compared the speed planned by the agent and a human driver, which showed a correlation of 0.52. We also confirmed the benefit of taking prediction into account by observing the behavior of the agent in a specific traffic scenario.

Insertion is a challenging haptic and visual control problem with significant practical value for manufacturing. Existing approaches in the model-based robotics community can be highly effective when task geometry is known, but are complex and cumbersome to implement, and must be tailored to each individual problem by a qualified engineer. Within the learning community there is a long history of insertion research, but existing approaches are either too sample-inefficient to run on real robots, or assume access to high-level object features, e.g. socket pose. In this paper we show that relatively minor modifications to an off-the-shelf Deep-RL algorithm (DDPG), combined with a small number of human demonstrations, allows the robot to quickly learn to solve these tasks efficiently and robustly. Our approach requires no modeling or simulation, no parameterized search or alignment behaviors, no vision system aside from raw images, and no reward shaping. We evaluate our approach on a narrow-clearance peg-insertion task and a deformable clip-insertion task, both of which include variability in the socket position. Our results show that these tasks can be solved reliably on the real robot in less than 10 minutes of interaction time, and that the resulting policies are robust to variance in the socket position and orientation.