Dongheui Lee

Zhifan Ni, Esteve Valls Mascaró, Hyemin Ahn et al.

Understanding the human-object interactions (HOIs) from a video is essential to fully comprehend a visual scene. This line of research has been addressed by detecting HOIs from images and lately from videos. However, the video-based HOI anticipation task in the third-person view remains understudied. In this paper, we design a framework to detect current HOIs and anticipate future HOIs in videos. We propose to leverage human gaze information since people often fixate on an object before interacting with it. These gaze features together with the scene contexts and the visual appearances of human-object pairs are fused through a spatio-temporal transformer. To evaluate the model in the HOI anticipation task in a multi-person scenario, we propose a set of person-wise multi-label metrics. Our model is trained and validated on the VidHOI dataset, which contains videos capturing daily life and is currently the largest video HOI dataset. Experimental results in the HOI detection task show that our approach improves the baseline by a great margin of 36.3% relatively. Moreover, we conduct an extensive ablation study to demonstrate the effectiveness of our modifications and extensions to the spatio-temporal transformer. Our code is publicly available on https://github.com/nizhf/hoi-prediction-gaze-transformer.

Esteve Valls Mascaro, Hyemin Ahn, Dongheui Lee

To anticipate how a human would act in the future, it is essential to understand the human intention since it guides the human towards a certain goal. In this paper, we propose a hierarchical architecture which assumes a sequence of human action (low-level) can be driven from the human intention (high-level). Based on this, we deal with Long-Term Action Anticipation task in egocentric videos. Our framework first extracts two level of human information over the N observed videos human actions through a Hierarchical Multi-task MLP Mixer (H3M). Then, we condition the uncertainty of the future through an Intention-Conditioned Variational Auto-Encoder (I-CVAE) that generates K stable predictions of the next Z=20 actions that the observed human might perform. By leveraging human intention as high-level information, we claim that our model is able to anticipate more time-consistent actions in the long-term, thus improving the results over baseline methods in EGO4D Challenge. This work ranked first in both CVPR@2022 and ECVV@2022 EGO4D LTA Challenge by providing more plausible anticipated sequences, improving the anticipation of nouns and overall actions. Webpage: https://evm7.github.io/icvae-page/

Hyemin Ahn, Esteve Valls Mascaro, Dongheui Lee

After many researchers observed fruitfulness from the recent diffusion probabilistic model, its effectiveness in image generation is actively studied these days. In this paper, our objective is to evaluate the potential of diffusion probabilistic models for 3D human motion-related tasks. To this end, this paper presents a study of employing diffusion probabilistic models to predict future 3D human motion(s) from the previously observed motion. Based on the Human 3.6M and HumanEva-I datasets, our results show that diffusion probabilistic models are competitive for both single (deterministic) and multiple (stochastic) 3D motion prediction tasks, after finishing a single training process. In addition, we find out that diffusion probabilistic models can offer an attractive compromise, since they can strike the right balance between the likelihood and diversity of the predicted future motions. Our code is publicly available on the project website: https://sites.google.com/view/diffusion-motion-prediction.

Manuel Stoiber, Mariam Elsayed, Anne E. Reichert et al.

In many applications of advanced robotic manipulation, six degrees of freedom (6DoF) object pose estimates are continuously required. In this work, we develop a multi-modality tracker that fuses information from visual appearance and geometry to estimate object poses. The algorithm extends our previous method ICG, which uses geometry, to additionally consider surface appearance. In general, object surfaces contain local characteristics from text, graphics, and patterns, as well as global differences from distinct materials and colors. To incorporate this visual information, two modalities are developed. For local characteristics, keypoint features are used to minimize distances between points from keyframes and the current image. For global differences, a novel region approach is developed that considers multiple regions on the object surface. In addition, it allows the modeling of external geometries. Experiments on the YCB-Video and OPT datasets demonstrate that our approach ICG+ performs best on both datasets, outperforming both conventional and deep learning-based methods. At the same time, the algorithm is highly efficient and runs at more than 300 Hz. The source code of our tracker is publicly available.

Esteve Valls Mascaro, Shuo Ma, Hyemin Ahn et al.

Comprehending human motion is a fundamental challenge for developing Human-Robot Collaborative applications. Computer vision researchers have addressed this field by only focusing on reducing error in predictions, but not taking into account the requirements to facilitate its implementation in robots. In this paper, we propose a new model based on Transformer that simultaneously deals with the real time 3D human motion forecasting in the short and long term. Our 2-Channel Transformer (2CH-TR) is able to efficiently exploit the spatio-temporal information of a shortly observed sequence (400ms) and generates a competitive accuracy against the current state-of-the-art. 2CH-TR stands out for the efficient performance of the Transformer, being lighter and faster than its competitors. In addition, our model is tested in conditions where the human motion is severely occluded, demonstrating its robustness in reconstructing and predicting 3D human motion in a highly noisy environment. Our experiment results show that the proposed 2CH-TR outperforms the ST-Transformer, which is another state-of-the-art model based on the Transformer, in terms of reconstruction and prediction under the same conditions of input prefix. Our model reduces in 8.89% the mean squared error of ST-Transformer in short-term prediction, and 2.57% in long-term prediction in Human3.6M dataset with 400ms input prefix. Webpage: https://evm7.github.io/2CHTR-page/

David Paulius, Alejandro Agostini, Dongheui Lee

Following work on joint object-action representations, functional object-oriented networks (FOON) were introduced as a knowledge graph representation for robots. A FOON contains symbolic concepts useful to a robot's understanding of tasks and its environment for object-level planning. Prior to this work, little has been done to show how plans acquired from FOON can be executed by a robot, as the concepts in a FOON are too abstract for execution. We thereby introduce the idea of exploiting object-level knowledge as a FOON for task planning and execution. Our approach automatically transforms FOON into PDDL and leverages off-the-shelf planners, action contexts, and robot skills in a hierarchical planning pipeline to generate executable task plans. We demonstrate our entire approach on long-horizon tasks in CoppeliaSim and show how learned action contexts can be extended to never-before-seen scenarios.

Esteve Valls Mascaro, Hyemin Ahn, Dongheui Lee

The synthesis of human motion has traditionally been addressed through task-dependent models that focus on specific challenges, such as predicting future motions or filling in intermediate poses conditioned on known key-poses. In this paper, we present a novel task-independent model called UNIMASK-M, which can effectively address these challenges using a unified architecture. Our model obtains comparable or better performance than the state-of-the-art in each field. Inspired by Vision Transformers (ViTs), our UNIMASK-M model decomposes a human pose into body parts to leverage the spatio-temporal relationships existing in human motion. Moreover, we reformulate various pose-conditioned motion synthesis tasks as a reconstruction problem with different masking patterns given as input. By explicitly informing our model about the masked joints, our UNIMASK-M becomes more robust to occlusions. Experimental results show that our model successfully forecasts human motion on the Human3.6M dataset. Moreover, it achieves state-of-the-art results in motion inbetweening on the LaFAN1 dataset, particularly in long transition periods. More information can be found on the project website https://evm7.github.io/UNIMASKM-page/

Victor Kowalski, Chengxi Li, Dongheui Lee

When using reinforcement learning (RL) for contact-rich robotic manipulation, vision can provide task-relevant information that accelerates learning beyond what proprioception alone can achieve. However, vision-enabled policies tend to overfit to the visual conditions seen during training, limiting their robustness and transferability. We present a human-in-the-loop RL framework that employs teacher-student distillation to achieve robust performance across multiple task variants, trained entirely in the real world without requiring domain randomization or data augmentation. A vision-enabled teacher distills its knowledge into a vision-free student that relies solely on pose, twist, and wrench sensing, combining fast training with strong task generalization. On the real-world NIST assembly benchmark board, our approach achieves 95\% overall success after approximately 50 minutes of training on 3 representative tasks, including robust generalization to 8 unseen task variants. Fine-tuning with distillation achieves full success on the most challenging task. We demonstrate that the resulting policies outperform baselines in both robustness and adaptability.

Yashuai Yan, Esteve Valls Mascaro, Dongheui Lee

This paper introduces a novel deep-learning approach for human-to-robot motion retargeting, enabling robots to mimic human poses accurately. Contrary to prior deep-learning-based works, our method does not require paired human-to-robot data, which facilitates its translation to new robots. First, we construct a shared latent space between humans and robots via adaptive contrastive learning that takes advantage of a proposed cross-domain similarity metric between the human and robot poses. Additionally, we propose a consistency term to build a common latent space that captures the similarity of the poses with precision while allowing direct robot motion control from the latent space. For instance, we can generate in-between motion through simple linear interpolation between two projected human poses. We conduct a comprehensive evaluation of robot control from diverse modalities (i.e., texts, RGB videos, and key poses), which facilitates robot control for non-expert users. Our model outperforms existing works regarding human-to-robot retargeting in terms of efficiency and precision. Finally, we implemented our method in a real robot with self-collision avoidance through a whole-body controller to showcase the effectiveness of our approach. More information on our website https://evm7.github.io/UnsH2R/

Esteve Valls Mascaro, Daniel Sliwowski, Dongheui Lee

Robots are becoming increasingly integrated into our lives, assisting us in various tasks. To ensure effective collaboration between humans and robots, it is essential that they understand our intentions and anticipate our actions. In this paper, we propose a Human-Object Interaction (HOI) anticipation framework for collaborative robots. We propose an efficient and robust transformer-based model to detect and anticipate HOIs from videos. This enhanced anticipation empowers robots to proactively assist humans, resulting in more efficient and intuitive collaborations. Our model outperforms state-of-the-art results in HOI detection and anticipation in VidHOI dataset with an increase of 1.76% and 1.04% in mAP respectively while being 15.4 times faster. We showcase the effectiveness of our approach through experimental results in a real robot, demonstrating that the robot's ability to anticipate HOIs is key for better Human-Robot Interaction. More information can be found on our project webpage: https://evm7.github.io/HOI4ABOT_page/

Esteve Valls Mascaro, Dongheui Lee

As humanoid robots transition from labs to real-world environments, it is essential to democratize robot control for non-expert users. Recent human-robot imitation algorithms focus on following a reference human motion with high precision, but they are susceptible to the quality of the reference motion and require the human operator to simplify its movements to match the robot's capabilities. Instead, we consider that the robot should understand and adapt the reference motion to its own abilities, facilitating the operator's task. For that, we introduce a deep-learning model that anticipates the robot's performance when imitating a given reference. Then, our system can generate multiple references given a high-level task command, assign a score to each of them, and select the best reference to achieve the desired robot behavior. Our Self-AWare model (SAW) ranks potential robot behaviors based on various criteria, such as fall likelihood, adherence to the reference motion, and smoothness. We integrate advanced motion generation, robot control, and SAW in one unique system, ensuring optimal robot behavior for any task command. For instance, SAW can anticipate falls with 99.29% accuracy. For more information check our project page: https://evm7.github.io/Self-AWare

Dongmyoung Lee, Chengxi Li, Dongheui Lee

Dexterous teleoperation via Mixed Reality (MR)-based interfaces offers a scalable paradigm for transferring human manipulation skills to dexterous robot hands. However, conventional retargeting approaches that minimize kinematic dissimilarity (e.g., joint angle or fingertip position error) often fail in contact-rich rotational manipulation, such as cap opening, key turning, and bolt screwing. This failure stems from the embodiment gap: mismatched link lengths, joint axes/limits, and fingertip geometry can cause direct pose imitation to induce tangential fingertip sliding rather than stable object rotation, resulting in screw axis drift, contact slip, and grasp instability. To address this, we propose DexTwist, a functional twist-retargeting framework for MR-based dexterous teleoperation. DexTwist detects a tripod pinch, estimates the operator's intended screw axis and twist magnitude, and applies a real-time residual joint-space refinement that tracks turning progress while regularizing the robot tripod geometry. The refinement minimizes a virtual-object objective defined by turning angle, screw axis consistency, fingertip closure, and tripod stability. Simulation and real-world experiments show that DexTwist improves turning angle tracking and screw axis stability compared with a vector-based retargeting baseline.

Sergej Stanovcic, Daniel Sliwowski, Dongheui Lee

Annotating long-horizon robotic demonstrations with precise temporal action boundaries is crucial for training and evaluating action segmentation and manipulation policy learning methods. Existing annotation tools, however, are often limited: they are designed primarily for vision-only data, do not natively support synchronized visualization of robot-specific time-series signals (e.g., gripper state or force/torque), or require substantial effort to adapt to different dataset formats. In this paper, we introduce ATLAS, an annotation tool tailored for long-horizon robotic action segmentation. ATLAS provides time-synchronized visualization of multi-modal robotic data, including multi-view video and proprioceptive signals, and supports annotation of action boundaries, action labels, and task outcomes. The tool natively handles widely used robotics dataset formats such as ROS bags and the Reinforcement Learning Dataset (RLDS) format, and provides direct support for specific datasets such as REASSEMBLE. ATLAS can be easily extended to new formats via a modular dataset abstraction layer. Its keyboard-centric interface minimizes annotation effort and improves efficiency. In experiments on a contact-rich assembly task, ATLAS reduced the average per-action annotation time by at least 6% compared to ELAN, while the inclusion of time-series data improved temporal alignment with expert annotations by more than 2.8% and decreased boundary error fivefold compared to vision-only annotation tools.

Yashuai Yan, Esteve Valls Mascaro, Tobias Egle et al.

Humanoid robots have the potential to mimic human motions with high visual fidelity, yet translating these motions into practical, physical execution remains a significant challenge. Existing techniques in the graphics community often prioritize visual fidelity over physics-based feasibility, posing a significant challenge for deploying bipedal systems in practical applications. This paper addresses these issues through bounded residual reinforcement learning to produce physics-based high-quality motion imitation onto legged humanoid robots that enhance motion resemblance while successfully following the reference human trajectory. Our framework, Imitation to Control Humanoid Robots Through Bounded Residual Reinforcement Learning (I-CTRL), reformulates motion imitation as a constrained refinement over non-physics-based retargeted motions. I-CTRL excels in motion imitation with simple and unique rewards that generalize across five robots. Moreover, our framework introduces an automatic priority scheduler to manage large-scale motion datasets when efficiently training a unified RL policy across diverse motions. The proposed approach signifies a crucial step forward in advancing the control of bipedal robots, emphasizing the importance of aligning visual and physical realism for successful motion imitation.

Daniel Sliwowski, Dongheui Lee

The introduction of robots into everyday scenarios necessitates algorithms capable of monitoring the execution of tasks. In this paper, we propose ConditionNET, an approach for learning the preconditions and effects of actions in a fully data-driven manner. We develop an efficient vision-language model and introduce additional optimization objectives during training to optimize for consistent feature representations. ConditionNET explicitly models the dependencies between actions, preconditions, and effects, leading to improved performance. We evaluate our model on two robotic datasets, one of which we collected for this paper, containing 406 successful and 138 failed teleoperated demonstrations of a Franka Emika Panda robot performing tasks like pouring and cleaning the counter. We show in our experiments that ConditionNET outperforms all baselines on both anomaly detection and phase prediction tasks. Furthermore, we implement an action monitoring system on a real robot to demonstrate the practical applicability of the learned preconditions and effects. Our results highlight the potential of ConditionNET for enhancing the reliability and adaptability of robots in real-world environments. The data is available on the project website: https://dsliwowski1.github.io/ConditionNET_page.

Esteve Valls Mascaro, Yashuai Yan, Dongheui Lee

Integrating robots into populated environments is a complex challenge that requires an understanding of human social dynamics. In this work, we propose to model social motion forecasting in a shared human-robot representation space, which facilitates us to synthesize robot motions that interact with humans in social scenarios despite not observing any robot in the motion training. We develop a transformer-based architecture called ECHO, which operates in the aforementioned shared space to predict the future motions of the agents encountered in social scenarios. Contrary to prior works, we reformulate the social motion problem as the refinement of the predicted individual motions based on the surrounding agents, which facilitates the training while allowing for single-motion forecasting when only one human is in the scene. We evaluate our model in multi-person and human-robot motion forecasting tasks and obtain state-of-the-art performance by a large margin while being efficient and performing in real-time. Additionally, our qualitative results showcase the effectiveness of our approach in generating human-robot interaction behaviors that can be controlled via text commands. Webpage: https://evm7.github.io/ECHO/

Daniel Sliwowski, Dongheui Lee

Temporal action segmentation (TAS) has long been a key area of research in both robotics and computer vision. In robotics, algorithms have primarily focused on leveraging proprioceptive information to determine skill boundaries, with recent approaches in surgical robotics incorporating vision. In contrast, computer vision typically relies on exteroceptive sensors, such as cameras. Existing multimodal TAS models in robotics integrate feature fusion within the model, making it difficult to reuse learned features across different models. Meanwhile, pretrained vision-only feature extractors commonly used in computer vision struggle in scenarios with limited object visibility. In this work, we address these challenges by proposing M2R2, a multimodal feature extractor tailored for TAS, which combines information from both proprioceptive and exteroceptive sensors. We introduce a novel pretraining strategy that enables the reuse of learned features across multiple TAS models. Our method achieves state-of-the-art performance on the REASSEMBLE dataset, a challenging multimodal robotic assembly dataset, outperforming existing robotic action segmentation models by 46.6%. Additionally, we conduct an extensive ablation study to evaluate the contribution of different modalities in robotic TAS tasks.

Dario Spoljaric, Yashuai Yan, Dongheui Lee

Reinforcement-learned locomotion enables legged robots to perform highly dynamic motions but often accompanies time-consuming manual tuning of joint stiffness. This paper introduces a novel control paradigm that integrates variable stiffness into the action space alongside joint positions, enabling grouped stiffness control such as per-joint stiffness (PJS), per-leg stiffness (PLS) and hybrid joint-leg stiffness (HJLS). We show that variable stiffness policies, with grouping in per-leg stiffness (PLS), outperform position-based control in velocity tracking and push recovery. In contrast, HJLS excels in energy efficiency. Despite the fact that our policy is trained on flat floor only, our method showcases robust walking behaviour on diverse outdoor terrains, indicating robust sim-to-real transfer. Our approach simplifies design by eliminating per-joint stiffness tuning while keeping competitive results with various metrics.

David Paulius, Alejandro Agostini, Yu Sun et al.

Following work on joint object-action representations, the functional object-oriented network (FOON) was introduced as a knowledge graph representation for robots. Taking the form of a bipartite graph, a FOON contains symbolic or high-level information that would be pertinent to a robot's understanding of its environment and tasks in a way that mirrors human understanding of actions. In this work, we outline a road-map for future development of FOON and its application in robotic systems for task planning as well as knowledge acquisition from demonstration. We propose preliminary ideas to show how a FOON can be created in a real-world scenario with a robot and human teacher in a way that can jointly augment existing knowledge in a FOON and teach a robot the skills it needs to replicate the demonstrated actions and solve a given manipulation problem.

Hyemin Ahn, Obin Kwon, Kyoungdo Kim et al.

This paper emphasizes the importance of a robot's ability to refer to its task history, especially when it executes a series of pick-and-place manipulations by following language instructions given one by one. The advantage of referring to the manipulation history can be categorized into two folds: (1) the language instructions omitting details but using expressions referring to the past can be interpreted, and (2) the visual information of objects occluded by previous manipulations can be inferred. For this, we introduce a history-dependent manipulation task which objective is to visually ground a series of language instructions for proper pick-and-place manipulations by referring to the past. We also suggest a relevant dataset and model which can be a baseline, and show that our model trained with the proposed dataset can also be applied to the real world based on the CycleGAN. Our dataset and code are publicly available on the project website: https://sites.google.com/view/history-dependent-manipulation.

Alejandro Agostini, Dongheui Lee

The definition of symbolic descriptions that consistently represent relevant geometrical aspects in manipulation tasks is a challenging problem that has received little attention in the robotic community. This definition is usually done from an observer perspective of a finite set of object relations and orientations that only satisfy geometrical constraints to execute experiments in laboratory conditions. This restricts the possible changes with manipulation actions in the object configuration space to those compatible with that particular external reference definitions, which greatly limits the spectrum of possible manipulations. To tackle these limitations we propose an object-centered representation that permits characterizing a much wider set of possible changes in configuration spaces than the traditional observer perspective counterpart. Based on this representation, we define universal planning operators for picking and placing actions that permits generating plans with geometric and force consistency in manipulation tasks. This object-centered description is directly obtained from the poses and bounding boxes of objects using a novel learning mechanisms that permits generating signal-symbols relations without the need of handcrafting these relations for each particular scenario.

Anil Armagan, Guillermo Garcia-Hernando, Seungryul Baek et al.

We study how well different types of approaches generalise in the task of 3D hand pose estimation under single hand scenarios and hand-object interaction. We show that the accuracy of state-of-the-art methods can drop, and that they fail mostly on poses absent from the training set. Unfortunately, since the space of hand poses is highly dimensional, it is inherently not feasible to cover the whole space densely, despite recent efforts in collecting large-scale training datasets. This sampling problem is even more severe when hands are interacting with objects and/or inputs are RGB rather than depth images, as RGB images also vary with lighting conditions and colors. To address these issues, we designed a public challenge (HANDS'19) to evaluate the abilities of current 3D hand pose estimators (HPEs) to interpolate and extrapolate the poses of a training set. More exactly, HANDS'19 is designed (a) to evaluate the influence of both depth and color modalities on 3D hand pose estimation, under the presence or absence of objects; (b) to assess the generalisation abilities w.r.t. four main axes: shapes, articulations, viewpoints, and objects; (c) to explore the use of a synthetic hand model to fill the gaps of current datasets. Through the challenge, the overall accuracy has dramatically improved over the baseline, especially on extrapolation tasks, from 27mm to 13mm mean joint error. Our analyses highlight the impacts of: Data pre-processing, ensemble approaches, the use of a parametric 3D hand model (MANO), and different HPE methods/backbones.

Matteo Saveriano, Dongheui Lee

Efficient skill acquisition, representation, and on-line adaptation to different scenarios has become of fundamental importance for assistive robotic applications. In the past decade, dynamical systems (DS) have arisen as a flexible and robust tool to represent learned skills and to generate motion trajectories. This work presents a novel approach to incrementally modify the dynamics of a generic autonomous DS when new demonstrations of a task are provided. A control input is learned from demonstrations to modify the trajectory of the system while preserving the stability properties of the reshaped DS. Learning is performed incrementally through Gaussian process regression, increasing the robot's knowledge of the skill every time a new demonstration is provided. The effectiveness of the proposed approach is demonstrated with experiments on a publicly available dataset of complex motions.

Matteo Saveriano, Felix Franzel, Dongheui Lee

In this paper, we focus on generating complex robotic trajectories by merging sequential motion primitives. A robotic trajectory is a time series of positions and orientations ending at a desired target. Hence, we first discuss the generation of converging pose trajectories via dynamical systems, providing a rigorous stability analysis. Then, we present approaches to merge motion primitives which represent both the position and the orientation part of the motion. Developed approaches preserve the shape of each learned movement and allow for continuous transitions among succeeding motion primitives. Presented methodologies are theoretically described and experimentally evaluated, showing that it is possible to generate a smooth pose trajectory out of multiple motion primitives.

Matteo Saveriano, Dongheui Lee

Stable dynamical systems are a flexible tool to plan robotic motions in real-time. In the robotic literature, dynamical system motions are typically planned without considering possible limitations in the robot's workspace. This work presents a novel approach to learn workspace constraints from human demonstrations and to generate motion trajectories for the robot that lie in the constrained workspace. Training data are incrementally clustered into different linear subspaces and used to fit a low dimensional representation of each subspace. By considering the learned constraint subspaces as zeroing barrier functions, we are able to design a control input that keeps the system trajectory within the learned bounds. This control input is effectively combined with the original system dynamics preserving eventual asymptotic properties of the unconstrained system. Simulations and experiments on a real robot show the effectiveness of the proposed approach.

Ziyuan Liu, Dongheui Lee, Wolfgang Sepp

The challenge of markerless human motion tracking is the high dimensionality of the search space. Thus, efficient exploration in the search space is of great significance. In this paper, a motion capturing algorithm is proposed for upper body motion tracking. The proposed system tracks human motion based on monocular silhouette-matching, and it is built on the top of a hierarchical particle filter, within which a novel deterministic resampling strategy (DRS) is applied. The proposed system is evaluated quantitatively with the ground truth data measured by an inertial sensor system. In addition, we compare the DRS with the stratified resampling strategy (SRS). It is shown in experiments that DRS outperforms SRS with the same amount of particles. Moreover, a new 3D articulated human upper body model with the name 3D cardbox model is created and is proven to work successfully for motion tracking. Experiments show that the proposed system can robustly track upper body motion without self-occlusion. Motions towards the camera can also be well tracked.

Pietro Falco, Shuang Lu, Ciro Natale et al.

In this work, we introduce the problem of cross-modal visuo-tactile object recognition with robotic active exploration. With this term, we mean that the robot observes a set of objects with visual perception and, later on, it is able to recognize such objects only with tactile exploration, without having touched any object before. Using a machine learning terminology, in our application we have a visual training set and a tactile test set, or vice versa. To tackle this problem, we propose an approach constituted by four steps: finding a visuo-tactile common representation, defining a suitable set of features, transferring the features across the domains, and classifying the objects. We show the results of our approach using a set of 15 objects, collecting 40 visual examples and five tactile examples for each object. The proposed approach achieves an accuracy of 94.7%, which is comparable with the accuracy of the monomodal case, i.e., when using visual data both as training set and test set. Moreover, it performs well compared to the human ability, which we have roughly estimated carrying out an experiment with ten participants.

Pietro Falco, Matteo Saveriano, Eka Gibran Hasany et al.

In this paper, we present a descriptor for human whole-body actions based on motion coordination. We exploit the principle, well known in neuromechanics, that humans move their joints in a coordinated fashion. Our coordination-based descriptor (CODE) is computed by two main steps. The first step is to identify the most informative joints which characterize the motion. The second step enriches the descriptor considering minimum and maximum joint velocities and the correlations between the most informative joints. In order to compute the distances between action descriptors, we propose a novel correlation-based similarity measure. The performance of CODE is tested on two public datasets, namely HDM05 and Berkeley MHAD, and compared with state-of-the-art approaches, showing recognition results.

Pietro Falco, Abdallah Attawia, Matteo Saveriano et al.

In this letter, we present an approach for learning in-hand manipulation skills with a low-cost, underactuated prosthetic hand in the presence of irreversible events. Our approach combines reinforcement learning based on visual perception with low-level reactive control based on tactile perception, which aims to avoid slipping. The objective of the reinforcement learning level consists not only in fulfilling the in-hand manipulation goal, but also in minimizing the intervention of the tactile reactive control. This way, the occurrence of object slipping during the learning procedure, which we consider an irreversible event, is significantly reduced. When an irreversible event occurs, the learning process is considered failed. We show the performance in two tasks, which consist in reorienting a cup and a bottle only using the fingers. The experimental results show that the proposed architecture allows reaching the goal in the Cartesian space and reduces significantly the occurrence of object slipping during the learning procedure. Moreover, without the proposed synergy between reactive control and reinforcement learning it was not possible to avoid irreversible events and, therefore, to learn the task.

Sang-ik An, Dongheui Lee

A prioritized inverse kinematics (PIK) solution can be considered as a (regulation or output tracking) control law of a dynamical system with prioritized multiple outputs. We propose a method that guarantees that a joint trajectory generated from a class of PIK solutions exists uniquely in a nonsingular configuration space. We start by assuming that desired task trajectories stay in nonsingular task spaces and find conditions for task trajectories to stay in a neighborhood of desired task trajectories in which we can guarantee existence and uniqueness of a joint trajectory in a nonsingular configuration space. Based on this result, we find a sufficient condition for task convergence and analyze various stability notions such as stability, uniform stability, uniform asymptotic stability, and exponential stability in both continuous and discrete times. We discuss why the number of tasks is limited in discrete time and show how preconditioning can be used in order to overcome this limitation.

Sang-ik An, Dongheui Lee

In this paper, we study various theoretical properties of a class of prioritized inverse kinematics (PIK) solutions that can be considered as a class of (output regulation or tracking) control laws of a dynamical system with prioritized multiple outputs. We first develop tools to investigate nonsmoothness of PIK solutions and find a sufficient condition for nonsmoothness. It implies that existence and uniqueness of a joint trajectory satisfying a PIK solution cannot be guaranteed by the classical theorems. So, we construct an alternative existence and uniqueness theorem that uses structural information of PIK solutions. Then, we narrow the class of PIK solutions down to the case that all tasks are designed to follow some desired task trajectories and discover a few properties related to task convergence. The study goes further to analyze stability of equilibrium points of the differential equation whose right hand side is a PIK solution when all tasks are designed to reach some desired task positions. Finally, we furnish an example with a two-link manipulator that shows how our findings can be used to analyze the behavior of a joint trajectory generated from a PIK solution.

Shile Li, Dongheui Lee

Recently, 3D input data based hand pose estimation methods have shown state-of-the-art performance, because 3D data capture more spatial information than the depth image. Whereas 3D voxel-based methods need a large amount of memory, PointNet based methods need tedious preprocessing steps such as K-nearest neighbour search for each point. In this paper, we present a novel deep learning hand pose estimation method for an unordered point cloud. Our method takes 1024 3D points as input and does not require additional information. We use Permutation Equivariant Layer (PEL) as the basic element, where a residual network version of PEL is proposed for the hand pose estimation task. Furthermore, we propose a voting based scheme to merge information from individual points to the final pose output. In addition to the pose estimation task, the voting-based scheme can also provide point cloud segmentation result without ground-truth for segmentation. We evaluate our method on both NYU dataset and the Hands2017Challenge dataset. Our method outperforms recent state-of-the-art methods, where our pose accuracy is currently the best for the Hands2017Challenge dataset.

Min Jun Kim, Jianjie Lin, Konstantin Kondak et al.

This paper presents an approach to damp out the oscillatory motion of the pendulum-like hanging platform on which a robotic manipulator is mounted. To this end, moving masses were installed on top of the platform. In this paper, asymptotic stability of the platform (which implies oscillation damping) is achieved by designing reference acceleration of the moving masses properly. A main feature of this work is that we can achieve asymptotic stability of not only the platform, but also the moving masses, which may be challenging due to the under-actuation nature. The proposed scheme is validated by the simulation studies.

Jan Wöhlke, Shile Li, Dongheui Lee

Since the emergence of large annotated datasets, state-of-the-art hand pose estimation methods have been mostly based on discriminative learning. Recently, a hybrid approach has embedded a kinematic layer into the deep learning structure in such a way that the pose estimates obey the physical constraints of human hand kinematics. However, the existing approach relies on a single person's hand shape parameters, which are fixed constants. Therefore, the existing hybrid method has problems to generalize to new, unseen hands. In this work, we extend the kinematic layer to make the hand shape parameters learnable. In this way, the learnt network can generalize towards arbitrary hand shapes. Furthermore, inspired by the idea of Spatial Transformer Networks, we apply a cascade of appearance normalization networks to decrease the variance in the input data. The input images are shifted, rotated, and globally scaled to a similar appearance. The effectiveness and limitations of our proposed approach are extensively evaluated on the Hands 2017 challenge dataset and the NYU dataset.

Shanxin Yuan, Guillermo Garcia-Hernando, Bjorn Stenger et al.

In this paper, we strive to answer two questions: What is the current state of 3D hand pose estimation from depth images? And, what are the next challenges that need to be tackled? Following the successful Hands In the Million Challenge (HIM2017), we investigate the top 10 state-of-the-art methods on three tasks: single frame 3D pose estimation, 3D hand tracking, and hand pose estimation during object interaction. We analyze the performance of different CNN structures with regard to hand shape, joint visibility, view point and articulation distributions. Our findings include: (1) isolated 3D hand pose estimation achieves low mean errors (10 mm) in the view point range of [70, 120] degrees, but it is far from being solved for extreme view points; (2) 3D volumetric representations outperform 2D CNNs, better capturing the spatial structure of the depth data; (3) Discriminative methods still generalize poorly to unseen hand shapes; (4) While joint occlusions pose a challenge for most methods, explicit modeling of structure constraints can significantly narrow the gap between errors on visible and occluded joints.

Simon Kaltenbacher, Nicholas H. Kirk, Dongheui Lee

Estimating the internal state of a robotic system is complex: this is performed from multiple heterogeneous sensor inputs and knowledge sources. Discretization of such inputs is done to capture saliences, represented as symbolic information, which often presents structure and recurrence. As these sequences are used to reason over complex scenarios, a more compact representation would aid exactness of technical cognitive reasoning capabilities, which are today constrained by computational complexity issues and fallback to representational heuristics or human intervention. Such problems need to be addressed to ensure timely and meaningful human-robot interaction. Our work is towards understanding the variability of learning informativeness when training on subsets of a given input dataset. This is in view of reducing the training size while retaining the majority of the symbolic learning potential. We prove the concept on human-written texts, and conjecture this work will reduce training data size of sequential instructions, while preserving semantic relations, when gathering information from large remote sources.