IROS 2020 论文列表

Many kinodynamic motion planners have been developed that guarantee probabilistic completeness and asymptotic optimality for systems for which steering functions are available. Recently, some planners have been developed that achieve these properties of completeness and optimality without requiring a steering function. However, these planners have not taken strong advantage of heuristic guidance to speed their search. This paper introduces Region Informed Optimal Trees (RIOT), a sampling-based, asymptotically optimal motion planner for systems without steering functions. RIOT's search is guided by a low-dimensional abstraction of the state space that is updated during planning for better guidance. Simulation results suggest RIOT is adaptable, scalable, and more effective on difficult problems than previous work.

In this paper, we present a modular robotic system to tackle the problem of generating and performing antipodal robotic grasps for unknown objects from the n-channel image of the scene. We propose a novel Generative Residual Convolutional Neural Network (GR-ConvNet) model that can generate robust antipodal grasps from n-channel input at real-time speeds (~20ms). We evaluate the proposed model architecture on standard datasets and a diverse set of household objects. We achieved state-of-the-art accuracy of 97.7% and 94.6% on Cornell and Jacquard grasping datasets, respectively. We also demonstrate a grasp success rate of 95.4% and 93% on household and adversarial objects, respectively, using a 7 DoF robotic arm.

We address the problem of adapting robot trajectories to improve safety, comfort, and efficiency in humanrobot collaborative tasks. To this end, we propose CoMOTO, a trajectory optimization framework that utilizes stochastic motion prediction to anticipate the human's motion and adapt the robot's joint trajectory accordingly. We design a multiobjective cost function that simultaneously optimizes for i) separation distance, ii) visibility of the end-effector, iii) legibility, iv) efficiency, and v) smoothness. We evaluate CoMOTO against three existing methods for robot trajectory generation when in close proximity to humans. Our experimental results indicate that our approach consistently outperforms existing methods over a combined set of safety, comfort, and efficiency metrics.

Circulating tumor cells (CTCs) are the primary cause of tumor metastasis after surgery. Metastatic tumor recurrence is the leading reason of cancer death. It is prerequisite to develop a platform for CTCs separation to predict the cancer cell transfer in important organs. Herein, a novel acoustic microfluidic device was designed to capture the "true" CTCs from the whole blood sample. The blood got from the mice with breast tumors removed. There are some CTCs that have escaped from the solid tumor contained in these blood samples, instead of artificially mixing individual tumor cells into normal blood. In addition, the predictions of tumor prognosis are made based on the number of CTCs captured by the acoustofluidic device. Finally, the prediction has been confirmed through long-term observation of mice with tumor excised. The acoustofluidic device can efficiently capture CTCs and predict the tumor metastasis, which can help clinicians plan follow-up treatment for patients who have had their tumors surgically removed.

As warehouses, storage facilities and factories become more expanded and equipped with smart devices, there is a substantial need for rapid, intelligent and autonomous detection of unusual and potentially hazardous situations, also called anomalies. In particular for Autonomous Guided Vehicles (AGVs) that drive around these premises independently, unforeseen obstructions along their path-e.g. a cardboard box in the middle of a corridor or bumps in the floor-and sudden or unexpected actions executed by personnel-e.g. someone walking in a restricted area-make it hard for AGVs to navigate safely. We therefore propose a novel approach to detect such anomalies in an unsupervised manner by measuring Bayesian surprise: whenever an event is observed that does not align with the agent's prior knowledge of the world, this event is deemed surprising and could indicate an anomaly. This paper lays out the details on how to learn both the prior and posterior models of an AGV that drives around a warehouse and observes the environment through an RGBD camera. In the experiments we show that our Bayesian surprise approach outperforms a baseline that is traditionally used to detect anomalies in sequences of images.

Creating robots with emotional personalities will transform the usability of robots in the real-world. As previous emotive social robots are mostly based on statically stable robots whose mobility is limited, this paper develops an animation to real-world pipeline that enables dynamic bipedal robots that can twist, wiggle, and walk to behave with emotions. First, an animation method is introduced to design emotive motions for the virtual robot's character. Second, a dynamics optimizer is used to convert the animated motion to dynamically feasible motion. Third, real-time standing and walking controllers and an automaton are developed to bring the virtual character to life. This framework is deployed on a bipedal robot Cassie and validated in experiments. To the best of our knowledge, this paper is one of the first to present an animatronic dynamic legged robot that is able to perform motions with desired emotional attributes. We term robots that use dynamic motions to convey emotions as Dynamic Relatable Robotic Characters.

This paper presents a dynamic constraint formulation to provide protective virtual fixtures of 3D anatomical structures from polygon mesh representations. The proposed approach can anisotropically limit the tool motion of surgical robots without any assumption of the local anatomical shape close to the tool. Using a bounded search strategy and Principle Directed tree, the proposed system can run efficiently at 180 Hz for a mesh object containing 989,376 triangles and 493,460 vertices. The proposed algorithm has been validated in both simulation and skull cutting experiments. The skull cutting experiment setup uses a novel piezoelectric bone cutting tool designed for the da Vinci research kit. The result shows that the virtual fixture assisted teleoperation has statistically significant improvements in the cutting path accuracy and penetration depth control. The code has been made publicly available at https://github.com/mli0603/PolygonMeshVirtualFixture.

This paper presents electro-hydrostatic actuation as a valid substitute of electro-mechanical devices for powered knee prostheses. The work covers the design of a test rig exploiting linear electro-hydrostatic actuation. Typical control laws for prosthesis actuators are discussed, implemented and validated experimentally. Particularly, this work focuses on position and admittance control syntheses enhanced with feed-forward friction compensation. Finally, the efficiency of the test rig is characterized experimentally and compared to that of classical electro-mechanical designs. It is demonstrated that the electro-hydrostatic prototype is able to fulfill its targets from a control perspective, while also having the potential to outperform electro-mechanical actuation in efficiency.

This paper studies the contact stability and contact safety of a robotic intravascular cardiac catheter under blood flow disturbances while in contact with tissue surface. A probabilistic blood flow disturbance model, where the blood flow drag forces on the catheter body are approximated using a quasi-static model, is introduced. Using this blood flow disturbance model, probabilistic contact stability and contact safety metrics, employing a sample based representation of the blood flow velocity distribution, are proposed. Finally, the contact stability and contact safety of a MRI-actuated robotic catheter are analyzed using these models in a specific example scenario under left pulmonary inferior vein (LIV) blood flow disturbances.

Humans exhibit outstanding learning, planning and adaptation capabilities while performing different types of industrial tasks. Given some knowledge about the task requirements, humans are able to plan their limbs motion in anticipation of the execution of specific skills. For example, when an operator needs to drill a hole on a surface, the posture of her limbs varies to guarantee a stable configuration that is compatible with the drilling task specifications, e.g. exerting a force orthogonal to the surface. Therefore, we are interested in analyzing the human arms motion patterns in industrial activities. To do so, we build our analysis on the so-called manipulability ellipsoid, which captures a posture-dependent ability to perform motion and exert forces along different task directions. Through thorough analysis of the human movement manipulability, we found that the ellipsoid shape is task dependent and often provides more information about the human motion than classical manipulability indices. Moreover, we show how manipulability patterns can be transferred to robots by learning a probabilistic model and employing a manipulability tracking controller that acts on the task planning and execution according to predefined control hierarchies.

A fundamental challenge in the field of modular and collective robots is balancing the trade-off between unit-level simplicity, which allows scalability, and unit-level functionality, which allows meaningful behaviors of the collective. At the same time, a challenge in the field of soft robotics is creating untethered systems, especially at a large scale with many controlled degrees of freedom (DOF). As a contribution toward addressing these challenges, here we present an untethered, soft cellular robot unit. A single unit is simple and one DOF, yet can increase its volume by 8x and apply substantial forces to the environment, can modulate its surface friction, and can switch its unit-to-unit cohesion while agnostic to unit-to-unit orientation. As a soft robot, it is robust and can achieve untethered operation of its DOF. We present the design of the unit, a volumetric actuator with a perforated strain-limiting fabric skin embedded with magnets surrounding an elastomeric membrane, which in turn encompasses a low-cost micro-pump, battery, and control electronics. We model and test this unit and show simple demonstrations of three-unit configurations that lift, crawl, and perform plate manipulation. Our untethered, soft cellular robot unit lays the foundation for new robust soft robotic collectives that have the potential to apply human-scale forces to the world.

Soft robots are capable of inherently safer interactions with their environment than rigid robots since they can mechanically deform in response to unanticipated stimuli. However, their complex mechanics can make planning and control difficult, particularly with tasks such as locomotion. In this work, we present a mobile and untethered underwater crawling soft robot, PATRICK, paired with a testbed that demonstrates closed-loop locomotion planning. PATRICK is inspired by the brittle star, with five flexible legs actuated by a total of 20 shape-memory alloy (SMA) wires, providing a rich variety of possible motions via its large input space. We propose a motion planning infrastructure based on a simple set of PATRICK's motion primitives, and provide experiments showing that the planner can command the robot to locomote to a goal state. These experiments contribute the first examples of closed-loop, state-space goal seeking of an underwater, untethered, soft crawling robot, and make progress towards full autonomy of soft mobile robotic systems.

We present the first demonstration of a battery-free untethered wirelessly powered sub-gram jumping robot on an insect-scale. In order to operate the insect-sized robot autonomously, the limitation in battery use emphasizes the need for a wireless power transmission system as an onboard power solution. We designed a wireless power transmission system based on inductive coupling to power the Shape Memory Alloy (SMA), which serves as an elastic energy storage element and actuator for the jumping robot. The assembled mechanical structures, onboard power and electronics yield a 2 mm (high) × 24 mm (long) × 12 mm (wide) robot with ~a weight of 216 mg. The experiments show that our jumping robot wirelessly lift-off up to 5.75 times its body length and repeats the jump around 7 times per minute. To date, out of the several untethered sub-gram insect-scale jumping robots with onboard power, this is the first wirelessly powered robot with the highest jumping performance. The novelty in this work, which addresses the engineering challenges in insect-scale jumping robots, is an untethered wirelessly powered design that achieves dynamic jumping maneuvers, and has self-righting ability.

Functional nanomaterials possess exceptional multi-physical (e.g., mechanical, electrical and optical) properties compared with their bulk counterparts. To facilitate both synthesis and device applications of these nanomaterials, it is highly desired to characterize their multi-physical properties with high accuracy and efficiency. The nanomanipulation techniques under scanning electron microscopy (SEM) has enabled the testing of mechanical and electrical properties of various nanomaterials. However, the seamless integration of mechanical, electrical, and optical testing techniques into an SEM for triple-field-coupled characterization of single nanostructures is still unexplored. In this work, we report the first SEM-based nanomanipulation system for high-resolution mechano-optoelectronic testing of single semiconductor InGaN/GaN nanowires (NWs). A custom-made optical measurement setup was integrated onto a four-probe nanomanipulator inside an SEM, with two optical microfibers actuated by the nanomanipulator for NW excitation and emission measurement. A conductive tungsten nanoprobe and a conductive atomic force microscopy (AFM) cantilever probe were integrated onto the nanomanipulator for electrical nanoprobing of single NWs for electroluminescence (EL) measurement. The AFM probe also served as a force sensor for quantifying the contact force applied to the NW during nanoprobing. Using this unique system, we examined, for the first time, the effect of mechanical compression applied to an InGaN/GaN NW on its optoelectronic properties.

Radar sensors have become an important part of the perception sensor suite due to their long range and their ability to work in adverse weather conditions. However, several shortcomings such as large amounts of noise and extreme sparsity of the point cloud result in them not being used to their full potential. In this paper, we present a novel Recursive Least Squares (RLS) based approach to estimate the instantaneous velocity of dynamic objects in real-time that is capable of handling large amounts of noise in the input data stream. We also present an end-to-end pipeline to track extended objects in real-time that uses the computed velocity estimates for data association and track initialisation. The approaches are evaluated using several real-world inspired driving scenarios that test the limits of these algorithms. It is also experimentally proven that our approaches run in real-time with frame execution time not exceeding 30 ms even in dense traffic scenarios, thus allowing for their direct implementation on autonomous vehicles.

Robot-assisted vitreoretinal surgery can filter surgeons' hand tremors and provide safe, accurate tool manipulation. In this paper, we report the design, optimization, and evaluation of a novel tilt mechanism for a new Steady-Hand Eye Robot (SHER). The new tilt mechanism features a four-bar linkage design and has a compact structure. Its kinematic configuration is optimized to minimize the required linear range of motion (LRM) for implementing a virtual remote center-of-motion (V-RCM) while tilting a surgical tool. Due to the different optimization constraints for the robots at the left and right sides of the human head, two configurations of this tilt mechanism are proposed. Experimental results show that the optimized tilt mechanism requires a significantly smaller LRM (e.g. 5.08 mm along Z direction and 8.77 mm along Y direction for left side robot) as compared to the slider-crank tilt mechanism used in the previous SHER (32.39 mm along Z direction and 21.10 mm along Y direction). The feasibility of the proposed tilt mechanism is verified in a mock bilateral robot-assisted vitreoretinal surgery. The ergonomically acceptable robot postures needed to access the surgical field is also determined.

Recent breakthroughs both in reinforcement learning and trajectory optimization have made significant advances towards real world robotic system deployment. Reinforcement learning (RL) can be applied to many problems without needing any modeling or intuition about the system, at the cost of high sample complexity and the inability to prove any metrics about the learned policies. Trajectory optimization (TO) on the other hand allows for stability and robustness analyses on generated motions and trajectories, but is only as good as the often over-simplified derived model, and may have prohibitively expensive computation times for real-time control, for example in contact rich environments. This paper seeks to combine the benefits from these two areas while mitigating their drawbacks by (1) decreasing RL sample complexity by using existing knowledge of the problem with real-time optimal control, and (2) allowing online policy deployment at any point in the training process by using the TO (MPC) as a baseline or worst-case scenario action, while continuously improving the combined learned-optimized policy with deep RL. This method is evaluated on tasks of successively navigating a car model to a series of goal destinations over slippery terrains as fast as possible, in which drifting will allow the system to more quickly change directions while maintaining high speeds.

This paper introduces an obstacle-crossing strategy, and the self-reconfiguration algorithm for a new class of modular robots called the rolling sphere, which can fit obstacles represented by cubes of different sizes due to the chain connection of multiple spheres. For the self-reconfiguration of the rolling spheres, a large gradient is obtained by classifying its action types and hierarchically minimizing the distance between the initial configuration and the final configuration. The most direct use of this large gradient is the fast crossing of various obstacles, by jointing multiple self-reconfigurations according to the OctoMap of the obstacles. It is verified in simulation that the self-reconfiguration takes full advantage of the parallel movement of multiple modules to reduce the total time steps, and the obstacle-crossing strategy can adapt to a variety of obstacles.

The recent development of in-pipe robots (IPR) with locomotion and inspection functions provides a new possibility to water distribution pipe maintenance - to rehabilitate pipe defects internally. Yet only a limited number of Rehabilitation in-pipe robots (R-IPR) have been proposed. One primary concern that impedes the development of Rehabilitation in-pipe robots is the excessive amount of contamination generated during the rehabilitation process. Correspondingly, we propose a novel concept: Contamination-Less in-pipe Rehabilitation (CLR) and develop the CLR in-pipe robot as an innovative solution. The proposed robot contains three modules for pipe-surface sealing, pipe-wall cleaning, and in-pipe manipulation. This paper centers on the comprehensive design of the manipulator module. First, the manipulator features a high-DoF configuration to deploy the other two modules simultaneously. Second, the configuration adopts a nested-outer-inner architecture to ensure the seal always encloses the pipe-wall cleaning device. The holistic and detailed design process of the manipulator, including design concept, kinematics, load requirements, design for manufacturing, and simulated deployment, are presented. Eventually, the fully implemented robot accomplished the first Contamination-Less in-pipe Rehabilitation.