ICRA 2020 论文列表

In industrial assembly tasks, the position of an object grasped by the robot has to be known with high precision in order to insert or place it. In real applications, this problem is commonly solved by jigs that are specially produced for each part. However, they significantly limit flexibility and are prohibitive when the target parts change often, so a flexible method to localize parts with high accuracy after grasping is desired. To solve this problem, we propose a method that can estimate the position of an object in the robot's hand to sub-millimeter precision, and can improve its estimate incrementally, using only minimal calibration and a force sensor. Our method is applicable to any robotic gripper and any rigid object that the gripper can hold, and requires only a force sensor. We demonstrate that the method can determine the position of an object to a precision of under 1 mm without using any part-specific jigs or equipment.

Path planning of an object which is allowed to be in contact with other objects during assembly process is a significant challenge due to the variety of permitted or forbidden collisions between the distinct parts of the objects to be assembled. In order to put objects together in real-life scenarios, parts of assembled objects may be required to flex, whereas other parts may have to fit exactly. Consequently, existing collision checking and distance computation algorithms have to be modified to enable path planning of objects that can be in contact during the assembly process. In this paper, we analyze an improved broad phase proximity query algorithm to enable such contact-based assembly tasks we call CHAT (Contact-based Hierarchy for Assembly Tasks). We demonstrate that, compared to existing approaches, our proposed method is more than an order of magnitude faster for collision queries and up to three times faster for distance queries when the two objects contain a large number of parts (with some parts containing thousands or tens of thousands of triangles). Due to the nature of the algorithm, we expect the performance improvements to increase as the number of parts in an object becomes larger.

While many robotic tasks, like manipulation and locomotion, are fundamentally based in making and breaking contact with the environment, state-of-the-art control policies struggle to deal with the hybrid nature of multi-contact motion. Such controllers often rely heavily upon heuristics or, due to the combinatoric structure in the dynamics, are unsuitable for real-time control. Principled deployment of tactile sensors offers a promising mechanism for stable and robust control, but modern approaches often use this data in an ad hoc manner, for instance to guide guarded moves. In this work, by exploiting the complementarity structure of contact dynamics, we propose a control framework which can close the loop on rich, tactile sensors. Critically, this framework is non-combinatoric, enabling optimization algorithms to automatically synthesize provably stable control policies. We demonstrate this approach on three different underactuated, multi-contact robotics problems.

Legged systems need to optimize contact force in order to maintain contacts. For this, the controller needs to have the knowledge of the surface geometry and how slippery the terrain is. We can use a vision system to realize the terrain, but the accuracy of the vision system degrades in harsh weather, and it cannot visualize the terrain if it is covered with water or grass. Also, the degree of friction cannot be directly visualized. In this paper, we propose an online method to estimate the surface information via haptic exploration. We also introduce a probabilistic criterion to measure the quality of the estimation. The method is validated on both simulation and a real robot platform.

Contact force quality is one of the most critical factors for safe and effective lesion formation during cardiac ablation. The contact force and contact stability plays important roles in determining the lesion size and creating a gap-free lesion. In this paper, the contact stability of a novel magnetic resonance imaging (MRI)-actuated robotic catheter under tissue surface motion is studied. The robotic catheter is modeled using a pseudo-rigid-body model, and the contact model under surface constraint is provided. Two contact force control schemes to improve the contact stability of the catheter under heart surface motions are proposed and their performance are evaluated in simulation.

We consider the problem of creating tighter-fitting bounding volumes (more specifically rectangular swept spheres) when constructing bounding volume hierarchies (BVHs) for complex 3D geometries given in the form of unstructured triangle meshes/soups with the aim of speeding up our IPS Path Planner for rigid bodies, where the triangles often have very different sizes. Currently, the underlying collision and distance computation module (IPS CDC) does not take into account the sizes of the triangles when it constructs BVHs using a top-down strategy. To split triangles in a BVH node into two BVH nodes, IPS CDC has to compute both the split axis and the split position. In this work, we use the principal axes of the tensor of inertia as the potential split axes and the center of mass as the split position, where the computations of both the tensor of inertia and the center of mass require knowledge of the areas of the triangles. We show that our method improves performance (up to 20 % faster) of our IPS Path Planner when it is used to plan collision-free disassembly paths for three different test cases taken from manufacturing industries.

We propose a sampling-based trajectory optimization methodology for constrained problems. We extend recent works on stochastic search to deal with box control constraints, as well as nonlinear state constraints for discrete dynamical systems. Regarding the former, our strategy is to optimize over truncated parameterized distributions on control inputs. Furthermore, we show how non-smooth penalty functions can be incorporated into our framework to handle state constraints. Simulations on cartpole and quadcopter show that our approach outperforms previous methods on constrained sampling-based optimization, in terms of quality of solutions and convergence speed.

In this paper, we propose a novel filtering-based method that fuses events from a dynamic vision sensor (DVS), images, and inertial measurements to estimate camera poses. A DVS is a bio-inspired sensor that generates events triggered by brightness changes. It can cover the drawbacks of a conventional camera by virtual of its independent pixels and high dynamic range. Specifically, we focus on optical flow obtained from both a stream of events and intensity images in which the former is much like a differential quantity, whereas the latter is a pixel difference in a much longer time interval than events. This nature characteristic motivates us to model optical flow estimated from events directly, but feature tracks for images in the filter design. An inequality constraint is considered in our method since the inverse scene-depth is larger than zero by its definition. Furthermore, we evaluate our proposed method in the benchmark DVS dataset and a dataset collected by the authors. The results reveal that the presented algorithm has reduced the position error by 49.9% on average and comparable accuracy only using events when compared to the state-of-the-art filtering-based estimator.

Connectivity maintenance is an essential aspect to consider while controlling a multi-robot system. In general, a multi-robot system should be connected to obtain a certain common objective. Connectivity must be kept regardless of the control strategy or the objective of the multi-robot system. Two main methods exist for connectivity maintenance: keep the initial connections (local connectivity) or allow modifications to the initial connections, but always keeping the overall system connected (global connectivity). In this paper we present a method that allows, at the same time, to maintain global connectivity and to implement the desired control strategy (e.g., consensus, formation control, coverage), all in an optimized fashion. For this purpose, we defined and implemented a Control Barrier Function that can incorporate constraints and objectives. We provide a mathematical proof of the method, and we demonstrate its versatility with simulations of different applications.

We present a congestion-aware routing solution for indoor evacuation, which produces real-time individual-customized evacuation routes among multiple destinations while keeping tracks of all evacuees’ locations. A population density map, obtained on-the-fly by aggregating locations of evacuees from user-end Augmented Reality (AR) devices, is used to model the congestion distribution inside a building. To efficiently search the evacuation route among all destinations, a variant of A⋆ algorithm is devised to obtain the optimal solution in a single pass. In a series of simulated studies, we show that the proposed algorithm is more computationally optimized compared to classic path planning algorithms; it generates a more time-efficient evacuation route for each individual that minimizes the overall congestion. A complete system using AR devices is implemented for a pilot study in real-world environments, demonstrating the efficacy of the proposed approach.

To control a free-floating robotic system with uncertain parameters in OOS tasks with high accuracy, a fast parameter identification method, previously developed by the authors, is enhanced further and used concurrently with a controller. The method provides accurate parameter estimates, without any prior knowledge of any system dynamic properties. This control scheme compensates for the accumulated angular momentum on the reaction wheels (RWs), which acts as a disturbance to the robotic servicer base. While any controller using parameter information can be used, a transposed Jacobian controller, modified to include RW angular momentum disturbance rejection, is employed here. Threedimensional simulations demonstrate the method's validity.

Soft materials and compliant actuation concepts have generated new design and control approaches in areas from robotics to wearable devices. Despite the potential of soft robotic systems, most designs currently use hard pumps, valves, and electromagnetic actuators. In this work, we take a step towards fully soft robots by developing a new compliant electromagnetic actuator architecture using gallium-indium liquid metal conductors, as well as compliant permanent magnetic and compliant iron composites. Properties of the new materials are first characterized and then co-fabricated to create an exemplary biologically-inspired soft actuator with pulsing or grasping motions, similar to Xenia soft corals. As current is applied to the liquid metal coil, the compliant permanent magnetic tips on passive silicone arms are attracted or repelled. The dynamics of the robotic actuator are characterized using stochastic system identification techniques and then operated at the resonant frequency of 7 Hz to generate high-stroke (>6 mm) motions.

Magnetic tracking algorithms can be used to determine the position and orientation of magnetic markers or devices. These techniques are particularly interesting for biomedical applications such as teleoperated surgical robots or the control of upper limb prostheses. The performance of different algorithms used for magnetic tracking was compared in the past. However, in most cases, those algorithms were required to track a single magnet.Here we investigated the performance of three localization algorithms in tracking up to 9 magnets: two optimization-based (Levenberg-Marquardt algorithm, LMA, and Trust Region Reflective algorithm, TRRA) and one recursion-based (Unscented Kalman Filter, UKF). The tracking accuracy of the algorithms and their computation time were investigated through simulations.The accuracy of the three algorithms, when tracking up to six magnets, was similar, leading to estimation errors varying from 0.06 ± 0.02 mm to 2.26 ± 0.07 mm within a 100 mm × 54 mm × 100 mm workspace, at the highest sampling frequency. In all cases, computation times under 300 ms for the UKF and 45 ms for the LMA/TRRA were obtained. The TRRA showed the best tracking performance overall.These outcomes are of interest for a wide range of robotics applications that require remote tracking.

Fingernail imaging has been proven to be effective in prior works [1], [2] for estimating the 3D fingertip forces with a maximum RMS estimation error of 7%. In the current research, fingernail imaging is used to perform unconstrained grasp force measurement on multiple fingers to study human grasping. Moreover, two robotic arms with mounted cameras and a visual tracking system have been devised to keep the human fingers in the camera frame during the experiments. Experimental tests have been conducted for six human subjects under both constrained and unconstrained grasping conditions, and the results indicate a significant difference in force collaboration among the fingers between the two grasping conditions. Another interesting result according to the experiments is that in comparison to constrained grasping, unconstrained grasp forces are more evenly distributed over the fingers and there is less force variation (more steadiness) in each finger force. These results validate the importance of measuring grasp forces in an unconstrained manner in order to study how humans naturally grasp objects.

Generally capable Spatial AI systems must build persistent scene representations where geometric models are combined with meaningful semantic labels. The many approaches to labelling scenes can be divided into two clear groups: view-based which estimate labels from the input view-wise data and then incrementally fuse them into the scene model as it is built; and map-based which label the generated scene model. However, there has so far been no attempt to quantitatively compare view-based and map-based labelling. Here, we present an experimental framework and comparison which uses real-time height map fusion as an accessible platform for a fair comparison, opening up the route to further systematic research in this area.

We exploit Markov games as a framework for collision-free navigation of human-centered robots. Unlike the classical methods which formulate robot navigation as a single-agent Markov decision process with a static environment, our framework of Markov games adopts a multi-agent formulation with one primary agent representing the robot and the remaining auxiliary agents form a dynamic or even competing environment. Such a framework allows us to develop a path-following type adversarial training strategy to learn a robust decentralized collision avoidance policy. Through thorough experiments on both simulated and real-world mobile robots, we show that the learnt policy outperforms the state-of-the-art algorithms in both sample complexity and runtime robustness.

Collision avoidance is a key technology towards safe human-robot interaction, especially on-line and fastreacting motions are required. Skins with proximity sensors mounted on the robot's outer shell provide an interesting approach to occlusion-free and low-latency perception. However, collision avoidance algorithms which make extensive use of these properties for fast-reacting motions have not yet been fully investigated. We present an improved collision avoidance algorithm for proximity sensing skins by formulating a quadratic optimization problem with inequality constraints to compute instantaneous optimal joint velocities. Compared to common repulsive force methods, our algorithm confines the approach velocity to obstacles and keeps motions pointing away from obstacles unrestricted. Since with repulsive motions the robot only moves in one direction, opposite to obstacles, our approach has better exploitation of the redundancy space to maintain the task motion and gets stuck less likely in local minima. Furthermore, our method incorporates an active behaviour for avoiding obstacles and evaluates all potentially colliding obstacles for the whole arm, rather than just the single nearest obstacle. We demonstrate the effectiveness of our method with simulations and on real robot manipulators in comparison with commonly used repulsive force methods and our prior proposed approach.

Master-slave systems for endovascular catheterization have brought major clinical benefits including reduced radiation doses to the operators, improved precision and stability of the instruments, as well as reduced procedural duration. Emerging deep reinforcement learning (RL) technologies could potentially automate more complex endovascular tasks with enhanced success rates, more consistent motion and reduced fatigue and cognitive workload of the operators. However, the complexity of the pulsatile flows within the vasculature and non-linear behavior of the instruments hinder the use of model-based approaches for RL. This paper describes model-free generative adversarial imitation learning to automate a standard arterial catherization task. The automation policies have been trained in a pre-clinical setting. Detailed validation results show high success rates after skill transfer to a different vascular anatomical model. The quality of the catheter motions also shows less mean and maximum contact forces compared to manual-based approaches.

This paper presents a novel filter architecture that allows a team of vehicles to collaboratively localize using Terrain Relative Navigation (TRN). The work explores several causes of measurement correlation that preclude the use of traditional estimators, and proposes an estimator structure that eliminates one source of measurement correlation while properly incorporating others through the use of Covariance Intersection. The result is a consistent estimator that is able to augment proven TRN techniques with multi-robot information, significantly improving localization for vehicles in uninformative terrain. The approach is demonstrated using field data from an Autonomous Underwater Vehicle (AUV) navigating with TRN in Monterey Bay and simulated inter-vehicle range measurements. In addition, a Monte Carlo simulation was used to quantify the algorithm's performance on one example mission. Monte Carlo results show that a vehicle operating in uninformative terrain has 62% lower localization error when fusing range measurements to two converged AUVs than it would using standard TRN.