Whole-Body Geometric Retargeting for Humanoid Robots

Darvish, Kourosh; Tirupachuri, Yeshasvi; Romualdi, Giulio; Rapetti, Lorenzo; Ferigo, Diego; Chavez, Francisco Javier Andrade; Pucci, Daniele

doi:10.1109/humanoids43949.2019.9035059

Cited by 38 publications

(33 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Among the various approaches to human motion retargeting (see, e.g., [25], [26], [27], [28]), Whole-Body Geometric Retargeting (WBGR) is a recent method easily adaptable to different robot models and human subjects [29]. Assuming a degree of topological similarity between the human's and robot's mechanical structures, WBGR uses m correspondences between frames associated with m human and robot links at a reference configuration.…”

Section: B Whole-body Geometric Retargetingmentioning

confidence: 99%

“…1) Kinematically-feasible base motion retargeting: WBGR in [29] does not address the base motion retargeting, i.e., the retargeted base position and orientation may not be compatible with the robot kinematics, thus possibly leading to a robot moving forward faster than what its walking pace entails. In other words, a swaying effect arises when dynamic motions are retargeted to robot models structurally different from the human subject [29]. While in CG generating models which fit the collected data is a viable workaround [23] [18], base motion retargeting for actual robots requires special attention.…”

Section: B Retargetingmentioning

confidence: 99%

See 1 more Smart Citation

ADHERENT: Learning Human-like Trajectory Generators for Whole-body Control of Humanoid Robots

Viceconte

Camoriano

Romualdi

et al. 2022

IEEE Robot. Autom. Lett.

Self Cite

View full text Add to dashboard Cite

Human-like trajectory generation and footstep planning represent challenging problems in humanoid robotics. Recently, research in computer graphics investigated machinelearning methods for character animation based on training human-like models directly on motion capture data. Such methods proved effective in virtual environments, mainly focusing on trajectory visualization. This paper presents ADHERENT, a system architecture integrating machine-learning methods used in computer graphics with whole-body control methods employed in robotics to generate and stabilize human-like trajectories for humanoid robots. Leveraging human motion capture locomotion data, ADHERENT yields a general footstep planner, including forward, sideways, and backward walking trajectories that blend smoothly from one to another. Furthermore, at the joint configuration level, ADHERENT computes data-driven wholebody postural reference trajectories coherent with the generated footsteps, thus increasing the human likeness of the resulting robot motion. Extensive validations of the proposed architecture are presented with both simulations and real experiments on the iCub humanoid robot, thus demonstrating ADHERENT to be robust to varying step sizes and walking speeds.

show abstract

Section: B Whole-body Geometric Retargetingmentioning

confidence: 99%

Section: B Retargetingmentioning

confidence: 99%

ADHERENT: Learning Human-like Trajectory Generators for Whole-body Control of Humanoid Robots

Viceconte

Camoriano

Romualdi

et al. 2022

IEEE Robot. Autom. Lett.

Self Cite

View full text Add to dashboard Cite

show abstract

“…To operate such systems, several approaches have been proposed that range from teleoperation to autonomous control. In teleoperation, an operator, usually equipped with a motion capture system, generates movements for a follower robot to accomplish a desired task [1], [5]. Teleoperation systems have been a topic of research for a long time, with a particular focus on their stability and transparency [6]- [8].…”

Section: Introductionmentioning

confidence: 99%

A Probabilistic Shared-Control Framework for Mobile Robots

Gholami

Garate

Momi

et al. 2020

2020 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)

View full text Add to dashboard Cite

Full teleoperation of mobile robots during the execution of complex tasks not only demands high cognitive and physical effort but also generates less optimal trajectories compared to autonomous controllers. However, the use of the latter in cluttered and dynamically varying environments is still an open and challenging topic. This is due to several factors such as sensory measurement failures and rapid changes in task requirements. Shared-control approaches have been introduced to overcome these issues. However, these either present a strong decoupling that makes them still sensitive to unexpected events, or highly complex interfaces only accessible to expert users. In this work, we focus on the development of a novel and intuitive shared-control framework for target detection and control of mobile robots. The proposed framework merges the information coming from a teleoperation device with a stochastic evaluation of the desired goal to generate autonomous trajectories while keeping a human-in-control approach. This allows the operator to react in case of goal changes, sensor failures, or unexpected disturbances. The proposed approach is validated through several experiments both in simulation and in a real environment where the users try to reach a chosen goal in the presence of obstacles and unexpected disturbances. Operators receive both visual feedback of the environment and voice feedback of the goal estimation status while teleoperating a mobile robot through a control-pad. Results of the proposed method are compared to pure teleoperation proving a better time-efficiency and easiness-of-use of the presented approach.

show abstract

“…Some assembly tasks such as snap-joining induce high acceleration during the engagement or disengagement of two parts [1]. A combined motion tracking approach for both humans and robots is necessary to develop a concept of safe human-robot physical collaboration [2]. In assembly tasks such as snap-joining, basic motions such as pick-reach -join -apply force -move are the typical activities to be applied.…”

Section: Introductionmentioning

confidence: 99%

Real-Time Motion Tracking for Humans and Robots in a Collaborative Assembly Task

Tuli

Manns

2019

The 6th International Electronic Conference on Sensors and Applications

View full text Add to dashboard Cite

Human-robot collaboration combines the extended capabilities of humans and robots to create a more inclusive and human-centered production system in the future. However, human safety is the primary concern for manufacturing industries. Therefore, real-time motion tracking is necessary to identify if the human worker body parts enter the restricted working space solely dedicated to the robot. Tracking these motions using decentralized and different tracking systems requires a generic model controller and consistent motion exchanging formats. In this work, our task is to investigate a concept for a unified real-time motion tracking for human-robot collaboration. In this regard, a low cost and game-based motion tracking system, e.g., HTC Vive, is utilized to capture human motion by mapping into a digital human model in the Unity3D environment. In this context, the human model is described using a biomechanical model that comprises joint segments defined by position and orientation. Concerning robot motion tracking, a unified robot description format is used to describe the kinematic trees. Finally, a concept of assembly operation that involves snap joining is simulated to analyze the performance of the system in real-time capability. The distribution of joint variables in spatial-space and time-space is analyzed. The results suggest that real-time tracking in human-robot collaborative assembly environments can be considered to maximize the safety of the human worker. However, the accuracy and reliability of the system regarding system disturbances need to be justified.

show abstract

Whole-Body Geometric Retargeting for Humanoid Robots

Cited by 38 publications

References 32 publications

ADHERENT: Learning Human-like Trajectory Generators for Whole-body Control of Humanoid Robots

ADHERENT: Learning Human-like Trajectory Generators for Whole-body Control of Humanoid Robots

A Probabilistic Shared-Control Framework for Mobile Robots

Real-Time Motion Tracking for Humans and Robots in a Collaborative Assembly Task

Contact Info

Product

Resources

About