We propose a model-free algorithm for learning efficient policies capable of returning table tennis balls by controlling robot joints at a rate of 100Hz. We demonstrate that evolutionary search (ES) methods acting on CNN-based policy architectures for non-visual inputs and convolving across time learn compact controllers leading to smooth motions. Furthermore, we show that with appropriately tuned curriculum learning on the task and rewards, policies are capable of developing multi-modal styles, specifically forehand and backhand stroke, whilst achieving 80\% return rate on a wide range of ball throws. We observe that multi-modality does not require any architectural priors, such as multi-head architectures or hierarchical policies.
MPI for Intelligent Systems（马克斯·普朗克智能系统研究所）；Technische Universität Darmstadt（达姆施塔特工业大学）；Facebook AI Research（脸书人工智能研究院）
Dynamic tasks like table tennis are relatively easy to learn for humans but pose significant challenges to robots. Such tasks require accurate control of fast movements and precise timing in the presence of imprecise state estimation of the flying ball and the robot. Reinforcement Learning (RL) has shown promise in learning of complex control tasks from data. However, applying step-based RL to dynamic tasks on real systems is safety-critical as RL requires exploring and failing safely for millions of time steps in high-speed regimes. In this paper, we demonstrate that safe learning of table te…
Machine learning, classification and prediction models have applications across a range of fields. Sport analytics is an increasingly popular application, but most existing work is focused on automated refereeing in mainstream sports and injury prevention. Research on other sports, such as table tennis, has only recently started gaining more traction. This paper proposes the use of machine learning to predict the outcome of table tennis single matches. We use player and match statistics as features and evaluate their relative importance in an ablation study. In terms of models, a number of pop…
Max Planck Institute for Evolutionary Anthropology（马克斯·普朗克进化人类学研究所）
This paper presents a table tennis stroke detection method from videos. The method relies on a two-stream Convolutional Neural Network processing in parallel the RGB Stream and its computed optical flow. The method has been developed as part of the MediaEval 2021 benchmark for the Sport task. Our contribution did not outperform the provided baseline on the test set but has performed the best among the other participants with regard to the mAP metric.
We introduce a novel method for collecting table tennis video data and perform stroke detection and classification. A diverse dataset containing video data of 11 basic strokes obtained from 14 professional table tennis players, summing up to a total of 22111 videos has been collected using the proposed setup. The temporal convolutional neural network model developed using 2D pose estimation performs multiclass classification of these 11 table tennis strokes with a validation accuracy of 99.37%. Moreover, the neural network generalizes well over the data of a player excluded from the training a…
Design of Recognition and Evaluation System for Table Tennis Players' Motor Skills Based on Artificial Intelligence
School of Electronics and Information, Northwestern Polytechnical University（西北工业大学电子信息学院）；School of Foreign Studies, Northwestern Polytechnical University（西北工业大学外国语学院）
In this work, the novel task of detecting and classifying table tennis strokes solely using the ball trajectory has been explored. A single camera setup positioned in the umpire's view has been employed to procure a dataset consisting of six stroke classes executed by four professional table tennis players. Ball tracking using YOLOv4, a traditional object detection model, and TrackNetv2, a temporal heatmap based model, have been implemented on our dataset and their performances have been benchmarked. A mathematical approach developed to extract temporal boundaries of strokes using the ball tra…
University of Bordeaux（波尔多大学）；CNRS（法国国家科学研究中心）；Bordeaux INP（波尔多国立高等物理、数学与信息学院）；University of La Rochelle（拉罗谢尔大学）
The paper addresses the problem of recognition of actions in video with low inter-class variability such as Table Tennis strokes. Two stream, "twin" convolutional neural networks are used with 3D convolutions both on RGB data and optical flow. Actions are recognized by classification of temporal windows. We introduce 3D attention modules and examine their impact on classification efficiency. In the context of the study of sportsmen performances, a corpus of the particular actions of table tennis strokes is considered. The use of attention blocks in the network speeds up the training step and i…
We propose a framework to enable multipurpose assistive mobile robots to autonomously wipe tables to clean spills and crumbs. This problem is challenging, as it requires planning wiping actions while reasoning over uncertain latent dynamics of crumbs and spills captured via high-dimensional visual observations. Simultaneously, we must guarantee constraints satisfaction to enable safe deployment in unstructured cluttered environments. To tackle this problem, we first propose a stochastic differential equation to model crumbs and spill dynamics and absorption with a robot wiper. Using this model…
Max Planck Institute for Evolutionary Anthropology（马克斯·普朗克进化人类学研究所）；Univ. Bordeaux（波尔多大学）；CNRS（法国国家科学研究中心）；Bordeaux INP（波尔多国立理工学院）；LaBRI（波尔多计算机科学研究所）；MIA, La Rochelle University（拉罗谢尔大学MIA研究所）；IMS, University of Bordeaux（波尔多大学IMS研究所）
This paper proposes a fusion method of modalities extracted from video through a three-stream network with spatio-temporal and temporal convolutions for fine-grained action classification in sport. It is applied to TTStroke-21 dataset which consists of untrimmed videos of table tennis games. The goal is to detect and classify table tennis strokes in the videos, the first step of a bigger scheme aiming at giving feedback to the players for improving their performance. The three modalities are raw RGB data, the computed optical flow and the estimated pose of the player. The network consists of t…
Max Planck Institute for Evolutionary Anthropology（马克斯·普朗克进化人类学研究所）；La Rochelle University（拉罗谢尔大学）；Univ. Bordeaux（波尔多大学）；CNRS（法国国家科学研究中心）；Bordeaux INP（波尔多国立高等理工学院）；LaBRI（波尔多计算机科学实验室）；University of Bordeaux（波尔多大学）
Sports video analysis is a prevalent research topic due to the variety of application areas, ranging from multimedia intelligent devices with user-tailored digests up to analysis of athletes' performance. The Sports Video task is part of the MediaEval 2021 benchmark. This task tackles fine-grained action detection and classification from videos. The focus is on recordings of table tennis games. Running since 2019, the task has offered a classification challenge from untrimmed video recorded in natural conditions with known temporal boundaries for each stroke. This year, the dataset is extended…
Max Planck Institute for Evolutionary Anthropology（马克斯·普朗克进化人类学研究所）；La Rochelle University（拉罗谢尔大学）；University of Bordeaux（波尔多大学）；CNRS（法国国家科学研究中心）；Bordeaux INP（波尔多国立理工学院）
Sports video analysis is a widespread research topic. Its applications are very diverse, like events detection during a match, video summary, or fine-grained movement analysis of athletes. As part of the MediaEval 2022 benchmarking initiative, this task aims at detecting and classifying subtle movements from sport videos. We focus on recordings of table tennis matches. Conducted since 2019, this task provides a classification challenge from untrimmed videos recorded under natural conditions with known temporal boundaries for each stroke. Since 2021, the task also provides a stroke detection ch…
Technical University of Darmstadt（达姆施塔特工业大学）；Mitsubishi Electric Research Laboratories（三菱电机研究实验室）
Robotic manipulation stands as a largely unsolved problem despite significant advances in robotics and machine learning in recent years. One of the key challenges in manipulation is the exploration of the dynamics of the environment when there is continuous contact between the objects being manipulated. This paper proposes a model-based active exploration approach that enables efficient learning in sparse-reward robotic manipulation tasks. The proposed method estimates an information gain objective using an ensemble of probabilistic models and deploys model predictive control (MPC) to plan act…
Reinforcement learning (RL) has achieved tremendous success as a general framework for learning how to make decisions. However, this success relies on the interactive hand-tuning of a reward function by RL experts. On the other hand, inverse reinforcement learning (IRL) seeks to learn a reward function from readily-obtained human demonstrations. Yet, IRL suffers from two major limitations: 1) reward ambiguity - there are an infinite number of possible reward functions that could explain an expert's demonstration and 2) heterogeneity - human experts adopt varying strategies and preferences, whi…
Manipulation tasks can often be decomposed into multiple subtasks performed in parallel, e.g., sliding an object to a goal pose while maintaining contact with a table. Individual subtasks can be achieved by task-axis controllers defined relative to the objects being manipulated, and a set of object-centric controllers can be combined in an hierarchy. In prior works, such combinations are defined manually or learned from demonstrations. By contrast, we propose using reinforcement learning to dynamically compose hierarchical object-centric controllers for manipulation tasks. Experiments in both …
Learning Nonlinear Causal Reductions to Explain Reinforcement Learning Policies
Max Planck Institute for Intelligent Systems（马克斯·普朗克智能系统研究所）；University of Alberta（阿尔伯塔大学）；Tübingen AI Center（图宾根人工智能中心）；ELLIS Institute（ELLIS研究所）；Technische Universität Braunschweig（不伦瑞克工业大学）
Max Planck Institute for Intelligent Systems（马克斯·普朗克智能系统研究所）
Reinforcement learning has shown great potential in solving complex tasks when large amounts of data can be generated with little effort. In robotics, one approach to generate training data builds on simulations based on dynamics models derived from first principles. However, for tasks that, for instance, involve complex soft robots, devising such models is substantially more challenging. Being able to train effectively in increasingly complicated scenarios with reinforcement learning enables to take advantage of complex systems such as soft robots. Here, we leverage the imbalance in complexit…
Institute for Mechatronic Systems, TU Darmstadt（达姆施塔特工业大学机电系统研究所）；Hessian Center for Artificial Intelligence (hessian.AI)（黑森人工智能中心）；Department of Computer Science, TU Darmstadt（达姆施塔特工业大学计算机科学系）；Department of Electrical Engineering, TU Darmstadt（达姆施塔特工业大学电气工程系）；ABB AG（ABB集团）
Transformers have revolutionized vision and natural language processing with their ability to scale with large datasets. But in robotic manipulation, data is both limited and expensive. Can manipulation still benefit from Transformers with the right problem formulation? We investigate this question with PerAct, a language-conditioned behavior-cloning agent for multi-task 6-DoF manipulation. PerAct encodes language goals and RGB-D voxel observations with a Perceiver Transformer, and outputs discretized actions by ``detecting the next best voxel action''. Unlike frameworks that operate on 2D ima…
