Learning to Play Table Tennis From Scratch Using Muscular Robots

Büchler, Dieter; Guist, Simon; Calandra, Roberto; Berenz, Vincent; Schoelkopf, Bernhard; Peters, Jan

doi:10.1109/tro.2022.3176207

Cited by 35 publications

(18 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…One of the most popular areas of study in this department is sport-related activities and accomplishment evaluation techniques. Sports players' motions have been collected and analyzed in a variety of ways, including cameras, sensors, and a combination of the two [ 24 ]. There have been a lot of studies performed on the subject of sport-related analysis and estimation.…”

Section: Introductionmentioning

confidence: 99%

Artificial Intelligence Technologies and Their Application for Reform and Development of Table Tennis Training in Complex Environments

Han

Zhang

2022

Computational Intelligence and Neuroscience

View full text Add to dashboard Cite

The development of information technology has been deployed in almost all sectors and is making life easier. In this development, the sports sector has seen tremendous expansion. In traditional table tennis training, the coach and the players have to meet daily to take appropriate training for the game. This process is time-consuming in a complex environment, and it will have a significant impact on the reformation and development of table tennis training. The utilization of improved technologies can overcome this challenge by performing training of the game online with an intelligent wireless system with advanced training mechanisms. Carrying the traditional and heavier intelligent wireless devices will be difficult for the players. In this research, the fine-grained evaluation (FGE) system is incorporated into the deep learning model to analyze the player’s body postures during the training and event sessions and make them develop after each session through online training in any circumstance. The proposed FGE was compared with the traditional statistical model, and it was observed that the proposed FGE had obtained higher precision and recall values of 70% and 98.9% than the statistical model.

show abstract

Section: Introductionmentioning

confidence: 99%

Artificial Intelligence Technologies and Their Application for Reform and Development of Table Tennis Training in Complex Environments

Han

Zhang

2022

Computational Intelligence and Neuroscience

View full text Add to dashboard Cite

show abstract

“…Hwangbo et al [16] and Lee et al [19] trained a neural network policy and then transferred it to a legged robot, which enabled the robot to overcome challenging terrain. Büchler et al [10] used a reinforcement learning method for enabling a PAM-driven robot arm to play table tennis. Alternative approaches include guided policy search, [21,20], which aims at finding a global feedback policy.…”

Section: A Related Workmentioning

confidence: 99%

“…The disturbance d[k] also contains interactions between the different degrees of freedom. We can convert (10) into the lifted state-space:…”

Section: B Formulationmentioning

confidence: 99%

See 1 more Smart Citation

A Learning-based Iterative Control Framework for Controlling a Robot Arm with Pneumatic Artificial Muscles

Ma¹,

B�chler²,

Sch�lkopf³

et al. 2022

Robotics: Science and Systems XVIII

View full text Add to dashboard Cite

In this work, we propose a new learning-based iterative control (IC) framework that enables a complex softrobotic arm to track trajectories accurately. Compared to traditional iterative learning control (ILC), which operates on a single fixed reference trajectory, we use a deep learning approach to generalize across various reference trajectories. The resulting nonlinear mapping computes feedforward actions and is used in a two degrees of freedom control design. Our method incorporates prior knowledge about the system dynamics and by learning only feedforward actions, it mitigates the risk of instability. We demonstrate a low sample complexity and an excellent tracking performance in real-world experiments. The experiments are carried out on a custom-made robot arm with four degrees of freedom that is actuated with pneumatic artificial muscles. The experiments include high acceleration and high velocity motion.

show abstract

“…While failure due to a high torque is a major problem for rigid robots, the body of a soft robot can absorb the vibrations caused by the high torque, and the softness of its body can reduce the impact if its motion is close to the prescribed limits. Therefore, soft robots and reinforcement learning may be compatible (Büchler et al, 2020).…”

Section: Soft Robot and Modelingmentioning

confidence: 99%

Characterization of continuum robot arms under reinforcement learning and derived improvements

et al. 2022

View full text Add to dashboard Cite

In robotics, soft continuum robot arms are a promising prospect owing to their redundancy and passivity; however, no comprehensive study exists that examines their characteristics compared to rigid manipulators. In this study, we examined the advantages of a continuum robot arm as compared to a typical and rigid seven-degree-of-freedom (7-DoF) robot manipulator in terms of performing various tasks through reinforcement learning. We conducted simulations for tasks with different characteristics that require control over position and force. Common tasks in robot manipulators, such as reaching, crank rotation, object throwing, and peg-in-hole were considered. The initial conditions of the robot and environment were randomized, aiming for evaluations including robustness. The results indicate that the continuum robot arm excels in the crank-rotation task, which is characterized by uncertainty in environmental conditions and cumulative rewards. However, the rigid robot arm learned better motions for the peg-in-hole task than the other tasks, which requires fine motion control of the end-effector. In the throwing task, the continuum robot arm scored well owing to its good handling of anisotropy. Moreover, we developed a reinforcement-learning method based on the comprehensive experimental results. The proposed method successfully improved the motion learning of a continuum robot arm by adding a technique to regulate the initial state of the robot. To the best of our knowledge, ours is the first reinforcement-learning experiment with multiple tasks on a single continuum robot arm and is the first report of a comparison between a single continuum robot arm and rigid manipulator on a wide range of tasks. This simulation study can make a significant contribution to the design of continuum arms and specification of their applications, and development of control and reinforcement learning methods.

show abstract

Learning to Play Table Tennis From Scratch Using Muscular Robots

Cited by 35 publications

References 24 publications

Artificial Intelligence Technologies and Their Application for Reform and Development of Table Tennis Training in Complex Environments

Artificial Intelligence Technologies and Their Application for Reform and Development of Table Tennis Training in Complex Environments

A Learning-based Iterative Control Framework for Controlling a Robot Arm with Pneumatic Artificial Muscles

Characterization of continuum robot arms under reinforcement learning and derived improvements

Contact Info

Product

Resources

About