A5083082888- Robot Manipulation and Learning
- Robotic Path Planning Algorithms
- Robotics and Sensor-Based Localization
- Reinforcement Learning in Robotics
- AI-based Problem Solving and Planning
- Robotic Locomotion and Control
- Soft Robotics and Applications
- Machine Learning and Algorithms
- Robotic Mechanisms and Dynamics
- Human Pose and Action Recognition
- Model Reduction and Neural Networks
- Modular Robots and Swarm Intelligence
- Advanced Vision and Imaging
- Tactile and Sensory Interactions
- Teleoperation and Haptic Systems
- Fault Detection and Control Systems
- Formal Methods in Verification
- 3D Shape Modeling and Analysis
- Human Motion and Animation
- Advanced Control Systems Optimization
- Prosthetics and Rehabilitation Robotics
- Social Robot Interaction and HRI
- Neural Networks and Applications
- Manufacturing Process and Optimization
- Muscle activation and electromyography studies
University of Michigan
2016-2025
Michigan United
2023
Robotics Research (United States)
2020-2022
Institute of Electrical and Electronics Engineers
2022
Gorgias Press (United States)
2022
Corvallis Environmental Center
2020
Worcester Polytechnic Institute
2013-2018
Normandie Université
2015
École Nationale Supérieure d'Ingénieurs de Caen
2015
University of California, Berkeley
2012-2013
This paper presents an overview of the inaugural Amazon Picking Challenge along with a summary survey conducted among 26 participating teams. The challenge goal was to design autonomous robot pick items from warehouse shelf. task is currently performed by human workers, and there hope that robots can someday help increase efficiency throughput while lowering cost. We report on 28-question posed teams learn about each team's background, mechanism design, perception apparatus, planning,...
We present a manipulation planning framework that allows robots to plan in the presence of constraints on end-effector pose, as well other common constraints. The has three main components: constraint representation, constraint-satisfaction strategies, and general algorithm. These components come together create an efficient probabilistically complete algorithm called Constrained BiDirectional Rapidly-exploring Random Tree (RRT) – CBiRRT2. underpinning our for pose is Task Space Regions...
In this paper we present a framework that allows human and robot to perform simultaneous manipulation tasks safely in close proximity. The proposed is based on early prediction of the human's motion. system, which builds previous work area gesture recognition, generates workspace occupancy by computing swept volume learned motion trajectories. planner then plans trajectories minimize penetration cost while interleaving planning execution. Multiple are computed parallel, one for each task...
We present the Constrained Bi-directional Rapidly-Exploring Random Tree (CBiRRT) algorithm for planning paths in configuration spaces with multiple constraints. This provides a general framework handling variety of constraints manipulation including torque limits, on pose an object held by robot, and following workspace surfaces. CBiRRT extends RRT (BiRRT) using projection techniques to explore space manifolds that correspond find bridges between them. Consequently, can solve many problems...
Robust perception is a vital capability for robotic manipulation in unstructured scenes. In this context, full pose estimation of relevant objects scene critical step towards the introduction robots into household environments. paper, we present an approach building metric 3D models using local descriptors from several images. Each model optimized to fit set calibrated training images, thus obtaining best possible alignment between and real object. Given new test image, match our stored...
We propose a framework, called Lightning, for planning paths in high-dimensional spaces that is able to learn from experience, with the aim of reducing computation time. This framework intended manipulation tasks arise applications ranging domestic assistance robot-assisted surgery. Our consists two main modules, which run parallel: planning-from-scratch module, and module retrieves repairs stored path library. After generated new query, library manager decides whether store based on time...
Deformable object manipulation (DOM) is an emerging research problem in robotics. The ability to manipulate deformable objects endows robots with higher autonomy and promises new applications the industrial, services, health-care sectors. However, compared rigid manipulation, of considerably more complex still open problem. Addressing DOM challenges demands breakthroughs almost all aspects robotics: hardware design, sensing, (deformation) modeling, planning, control. In this article, we...
This paper combines grasp analysis and manipulation planning techniques to perform fast in complex scenes. In much previous work on grasping, the object being grasped is assumed be only environment. Hence quality metrics grasping strategies developed do not well when close obstacles many good grasps are infeasible. We introduce a framework for finding valid cluttered environments that metric with information about local environment around robot's kinematics. encode these factors...
In this paper, we present efficient solutions for planning motions of dual-arm manipulation and re-grasping tasks. Motion such tasks on humanoid robots with a high number degrees freedom (DoF) requires computationally approaches to determine the robot's full joint configuration at given grasping position, i.e. solving Inverse Kinematics (IK) problem one or both hands robot. context, investigate inverse kinematics motion by combining gradient-descent approach in pre-computed reachability...
We present a method to manipulate deformable objects that does not require modeling and simulating deformation. Our is based on the concept of diminishing rigidity, which we use quickly compute an approximation Jacobian object. This used drive points within object towards set targets. However, this alone insufficient avoid stretching beyond its allowed length gripper collision with obstacles. Thus key part our approach incorporating techniques excessive stretching. experiments show how...
We present an approach to path planning for manipulators that uses Workspace Goal Regions (WGRs) specify goal end-effector poses. Instead of specifying a discrete set goals in the manipulator's configuration space, we more intuitively as volumes workspace. show WGRs provide common framework describing regions are useful grasping and manipulation. also describe two randomized algorithms capable with WGRs. The first is extension RRT-JT interleaves exploration using Rapidly-exploring Random...
We present the hardware design, software architecture, and core algorithms of Herb 2.0, a bimanual mobile manipulator developed at Personal Robotics Lab Carnegie Mellon University, Pittsburgh, PA. have 2.0 to perform useful tasks for with people in human environments. exploit two key paradigms environments: that they structure robot can learn, adapt exploit, demand general-purpose capability robotic systems. In this paper, we reveal some everyday environments been able harness manipulation...
To enable safe and efficient human-robot collaboration in shared workspaces, it is important for the robot to predict how a human will move when performing task. While predicting motion tasks not known priori very challenging, we argue that single-arm reaching motions collaborative settings (which are especially relevant manufacturing) indeed predictable. Two hypotheses underlie our approach such motions: First, trajectory performs optimal with respect an unknown cost function, second,...
Soft Pneumatic Actuators (SPAs) have recently become popular for use as fingers in robotic hands because of their inherent compliance, low cost, and ease construction. We seek to overcome two key limitations which limit SPAs' abilities grasp manipulate objects: 1) Current SPAs lack position or force sensor feedback, prevents controlling them precisely (e.g. achieve a hand preshape apply specified pushing force), 2) the tip SPA is compliant has high friction against common surfaces, causing...
To enable safe and efficient human-robot collaboration in shared workspaces, it is important for the robot to predict how a human will move when performing task. While predicting motion tasks not known priori very challenging, we argue that single-arm reaching motions collaborative settings (which are especially relevant manufacturing) indeed predictable. Two hypotheses underlie our approach such motions: First, trajectory performs optimal with respect an unknown cost function, second,...
We describe Team WPI-CMU's approach to the DARPA Robotics Challenge (DRC), focusing on our strategy avoid failures that required physical human intervention. implemented safety features in controller detect potential catastrophic failures, stop current behavior, and allow remote intervention by a supervisor. Our methods operator interface worked: we avoided catastrophe operators could safely recover from difficult situations. were only team DRC Finals attempted all tasks, scored points...
We present an optimization-based approach to grasping and path planning for mobile manipulators. focus on pick-and-place operations, where a given object must be moved from its start configuration goal by the robot. Given only configurations of model robot scene, our algorithm finds grasp trajectory that will bring configuration. The consists two phases: optimization planning. In phase, optimal are found in using co-evolutionary algorithm. is connecting phase Rapidly-Exploring Random Trees...
We present an algorithm for efficiently generating collision-free force-closure grasps dexterous hands in cluttered environments. Computing a grasp is complicated by the high dimensionality of hand configuration space, and cost validating candidate collision-checking testing force-closure. When object placed new scene, we use novel function to focus our search good regions pose space given preshape. The proposed fast compute encapsulates aspects object, ensuing grasp. low-cost produced are...
For successful deployment, personal robots must adapt to ever-changing indoor environments. While dealing with novel objects is a largely unsolved challenge in AI, it easy for people. In this paper we present framework robot supervision through Amazon Mechanical Turk. Unlike traditional models of teleoperation, people provide semantic information about the world and subjective judgements. The then autonomously utilizes additional enhance its capabilities. can be collected on demand large...
This paper explores how Cloud Computing can facilitate grasping with shape uncertainty. We consider the most common robot gripper: a pair of thin parallel jaws, and class objects that be modeled as extruded polygons. model conservative push-grasps enhance object alignment. The grasp planning algorithm takes input an approximate outline Gaussian uncertainty around each vertex center mass. define quality metric based on lower bound probability achieving force closure. present...
We plan for rope manipulation with an unreliable model by learning where to trust the and how recover when stuck.