检索结果-内蒙古大学图书馆

model-Based Policy Search Using Monte Carlo Gradient Estimation With Real Systems Application

IEEE TRANSACTIONS ON ROBOTICS 2022年第6期38卷 3879-3898页

作者： Amadio, Fabio Dalla Libera, Alberto Antonello, Riccardo Nikovski, Daniel Carli, Ruggero Romeres, Diego Univ Padua Deptartment Informat Engn I-35131 Padua Italy Mitsubishi Elect Res Labs MERL Cambridge MA 02139 USA

In this article, we present a model-based reinforcement learning (MBRL) algorithm named Monte Carlo probabilistic inference for learning control (MC-PILCO). This algorithm relies on Gaussian processes (GPs) to model the system dynamics and on a Monte Carlo approach to estimate the policy gradient. This defines a framework in which we ablate the choice of the components, which are the selection of the cost function, the optimization of policies using dropout, and an improved data efficiency through the use of structured kernels in the GP models. The combination of the aforementioned aspects affects dramatically the performance of MC-PILCO. Numerical comparisons in a simulated cart-pole environment show that MC-PILCO exhibits better data efficiency and control performance w.r.t. state-of-the-art GP-based MBRL algorithms. Finally, we apply MC-PILCO to real systems, considering, in particular, systems with partially measurable states. We discuss the importance of modeling both the measurement system and the state estimators during policy optimization. The effectiveness of the proposed solutions has been tested in simulation and on two real systems, which are the Furuta pendulum and the ball-and-plate rig.

关键词： Optimization Particle measurements Atmospheric measurements Prediction algorithms Kernel Costs Computational modeling Dynamics learning and adaptive systems model learning for control robot learning

来源：评论

学校读者我要写书评

暂无评论

Fast Simulation-Based Order Sequence Optimization Assisted by Pre-Trained Bayesian Recurrent Neural Network

引用

IEEE ROBOTICS AND AUTOMATION LETTERS 2022年第3期7卷 7818-7825页

作者： Suemitsu, Issei Bhamgara, Hanoz Kaiwan Utsugi, Kei Hashizume, Jiro Ito, Kiyoto Hitachi Ltd Ctr Technol Innovat Res & Dev Grp Kokubunji Tokyo 1858601 Japan

This paper presents a fast optimization method for the picking order sequence of automated order picking systems in logistics warehouses. In this order sequencing problem (OSP), the fulfillment sequence of the given picking order set is determined to optimize the performance measures such as makespan and deadlock occurrence. Simulation is generally necessary to evaluate these measures for complex automated systems. However, their order sequence cannot be optimized quickly due to the long calculation time. It may make the system productivity and flexibility lower than expected because its picking schedules cannot be updated frequently. We, therefore, propose a fast optimization method to solve these simulation-based OSPs by taking a pretrained surrogate-assisted optimization approach. Firstly, we utilized a Bayesian recurrent neural network (BRNN) as a surrogate model to accurately learn the relationship between picking order sequence and performances. Secondly, we developed the surrogate-assisted optimization method based on simulated annealing (SA) and BRNN. Numerical experiments show that the surrogate model can evaluate about 10000 times faster than the simulation. The proposed method also obtains an optimized solution 8.9 times faster than simulation-based optimization by the original SA.

关键词： Logistics Planning Scheduling and Coordination Optimization and Optimal control Deep learning Methods model learning for control

来源：评论

学校读者我要写书评

暂无评论

Auto-Tuned Motion Scaling in Teleoperation Based on Human Reaction model Identification

引用

IEEE ROBOTICS AND AUTOMATION LETTERS 2022年第1期7卷 318-325页

作者： Hsia, Shao-Kang Chuang, Yi-Hang Chen, Cheng-Wei Natl Taiwan Univ Dept Elect Engn Taipei 10617 Taiwan

Motion scaling is an essential technique in robotic surgical systems adopting the leader-follower configuration. By properly reducing the scaling factor, the surgeon can magnify the motion resolution that a human cannot achieve. However, manually tuning the scaling factor distracts the surgeon during the operation. Hence, adaptive methods were introduced to adjust the scaling factor autonomously, despite increasing the system's complexity and leaving more parameters to be designed. We propose a novel framework enabling a systematic design of the motion scaling auto-tuner to address this problem. First, the leader-follower configurated teleoperation is modeled as a human-in-loop control system. Then, we attain the motion scaling auto-tuner by model-matching based filter design. The proposed method is also integrated with virtual fixture techniques, which improve the safety of surgical tasks via haptic feedback. Finally, experiments are conducted for performance evaluation and comparison. The task completion time and other evaluation metrics are effectively improved with the systematic design framework.

关键词： Robots Surgery Task analysis Systematics Haptic interfaces Dynamics Collision avoidance Medical robots and systems telerobotics and teleoperation model learning for control

来源：评论

学校读者我要写书评

暂无评论

Trajectory Optimization and model Predictive control for Functional Electrical Stimulation-controlled Reaching

引用

IEEE ROBOTICS AND AUTOMATION LETTERS 2022年第2期7卷 3093-3098页

作者： Wolf, Derek N. Schearer, Eric M. Vanderbilt Univ Dept Mech Engn Nashville TN 37206 USA Cleveland State Univ Ctr Human Machine Syst Cleveland Funct Elect Stimulat Ctr Cleveland OH 44115 USA Metrohlth Med Ctr Dept Phys Med & Rehabil Cleveland OH 44115 USA

Functional electrical stimulation (FES) offers promise as a technology to restore reaching motions to individuals with spinal cord injuries. To date, the level of reaching necessary for everyday use has not been achieved due to the complexity and limitations of the arm and muscles of an individual with a spinal cord injury. To improve the performance of FES-driven reaching controllers, we developed a trajectory optimization and model predictive control scheme that incorporates knowledge of the person-specific muscle capabilities and arm dynamics. Our controller achieved 3D reaching motions with an average accuracy of 8.5 cm and demonstrated an ability to reach targets throughout the workspace. With improvements to the model, this control scheme has the potential to unlock many daily reaching tasks for individuals with spinal cord injuries.

关键词： model learning for control motion and path planning motion control prosthetics and exoskeletons rehabilitation robotics

来源：评论

学校读者我要写书评

暂无评论

control of Rough Terrain Vehicles Using Deep Reinforcement learning

引用

IEEE ROBOTICS AND AUTOMATION LETTERS 2022年第1期7卷 390-397页

作者： Wiberg, Viktor Wallin, Erik Nordfjell, Tomas Servin, Martin Umea Univ Dept Phys S-90338 Umea Sweden Swedish Univ Agr Sci S-75007 Uppsala Sweden

We explore the potential to control terrain vehicles using deep reinforcement in scenarios where human operators and traditional control methods are inadequate. This letter presents a controller that perceives, plans, and successfully controls a 16-tonne forestry vehicle with two frame articulation joints, six wheels, and their actively articulated suspensions to traverse rough terrain. The carefully shaped reward signal promotes safe, environmental, and efficient driving, which leads to the emergence of unprecedented driving skills. We test learned skills in a virtual environment, including terrains reconstructed from high-density laser scans of forest sites. The controller displays the ability to handle obstructing obstacles, slopes up to 27 degrees, and a variety of natural terrains, all with limited wheel slip, smooth, and upright traversal with intelligent use of the active suspensions. The results confirm that deep reinforcement learning has the potential to enhance control of vehicles with complex dynamics and high-dimensional observation data compared to human operators or traditional control methods, especially in rough terrain.

关键词： Deep learning methods reinforcement learning autonomous vehicle navigation model learning for control robotics and automation in agriculture and forestry

来源：评论

学校读者我要写书评

暂无评论

learning-Based Approach for a Soft Assistive Robotic Arm to Achieve Simultaneous Position and Force control

引用

IEEE ROBOTICS AND AUTOMATION LETTERS 2022年第3期7卷 8315-8322页

作者： Tang, Zhiqiang Wang, Peiyi Xin, Wenci Laschi, Cecilia Natl Univ Singapore Dept Mech Engn Singapore 117575 Singapore Beijing Jiaotong Univ Robot Res Ctr Beijing 100044 Peoples R China

Soft robotics have demonstrated great advantages in assisting elderly/disabled people during daily tasks, owing to their highly dexterous motions and safe human-robot interactions. However, simultaneously controlling the position and force of soft robots is still a challenging task due to soft actuators' nonlinearity, system uncertainty, and high-dimensional control space. Classical control methods are usually based on first-principle/analytical models that are difficult to derive for soft robots without making significant simplifications. To overcome such control challenges, the central concept of this work is to introduce a learning-based data-driven approach. The approach employs a probabilistic model to explicitly capture system nonlinearity and uncertainty. Besides, nonparametric local learning methods are investigated to deal with redundant high-dimensional control space. The approach is applied in a soft robotic arm interacting with a manikin to simulate the bathing task. Experimental results demonstrate that the soft robotic arm could be well controlled to track the desired position and force simultaneously (maximum error of position and force is 6 mm and 0.037 N). Meanwhile, our method outperforms another typical data-driven approach (maximum error of position and force is 10 mm and 0.058 N). The results indicate that our approach is helpful for soft robots because of the physical interactions needed in assistive tasks.

关键词： modeling control and learning for soft robots model learning for control optimization and optimal control

来源：评论

学校读者我要写书评

暂无评论

On-Line learning for Planning and control of Underactuated Robots With Uncertain Dynamics

引用

IEEE ROBOTICS AND AUTOMATION LETTERS 2022年第1期7卷 358-365页

作者： Turrisi, Giulio Capotondi, Marco Gaz, Claudio Modugno, Valerio Oriolo, Giuseppe De Luca, Alessandro Sapienza Univ Roma Dipartimento Ingn Informat Automat & Gest I-00185 Rome Italy

We present an iterative approach for planning and controlling motions of underactuated robots with uncertain dynamics. At its core, there is a learning process which estimates the perturbations induced by the model uncertainty on the active and passive degrees of freedom. The generic iteration of the algorithm makes use of the learned data in both the planning phase, which is based on optimization, and the control phase, where partial feedback linearization of the active dofs is performed on the model updated on-line. The performance of the proposed approach is shown by comparative simulations and experiments on a Pendubot executing various types of swing-up maneuvers. Very few iterations are typically needed to generate dynamically feasible trajectories and the tracking control that guarantees their accurate execution, even in the presence of large model uncertainties.

关键词： Underactuated robot model learning for control optimization and optimal control

来源：评论

学校读者我要写书评

暂无评论

Sample Efficient Dynamics learning for Symmetrical Legged Robots: Leveraging Physics Invariance and Geometric Symmetries

Sample Efficient Dynamics Learning for Symmetrical Legged Ro...

引用

IEEE International Conference on Robotics and Automation (ICRA)

作者： Lee, Jee-eun Lee, Jaemin Bandyopadhyay, Tirthankar Sentis, Luis Univ Texas Austin Dept Aerosp Engn & Engn Mech Human Ctr Robot Lab Austin TX 78712 USA CALTECH Dept Mech & Civil Engn Pasadena CA 91125 USA Univ Texas Austin Fac Dept Aerosp Engn & Engn Mech Austin TX USA CSIRO Robot & Autonomous Syst Grp Data61 Pullenvale Qld 4069 Australia

ISBN: (纸本)9798350323658

model generalization of the underlying dynamics is critical for achieving data efficiency when learning for robot control. This paper proposes a novel approach for learning dynamics leveraging the symmetry in the underlying robotic system, which allows for robust extrapolation from fewer samples. Existing frameworks that represent all data in vector space fail to consider the structured information of the robot, such as leg symmetry, rotational symmetry, and physics invariance. As a result, these schemes require vast amounts of training data to learn the system's redundant elements because they are learned independently. Instead, we propose considering the geometric prior by representing the system in symmetrical object groups and designing neural network architecture to assess invariance and equivariance between the objects. Finally, we demonstrate the effectiveness of our approach by comparing the generalization to unseen data of the proposed model and the existing models. We also implement a controller of a climbing robot based on learned inverse dynamics models. The results show that our method generates accurate control inputs that help the robot reach the desired state while requiring less training data than existing methods.

关键词： model learning for control Representation learning Group-equivalent Neural Networks

来源：评论

学校读者我要写书评

暂无评论

Neural Predictor for Flight control With Payload

引用

IEEE ROBOTICS AND AUTOMATION LETTERS 2025年第7期10卷 7055-7062页

作者： Jin, Ao Li, Chenhao Wang, Qinyi Liu, Ya Huang, Panfeng Zhang, Fan Northwestern Polytech Univ Res Ctr Intelligent Robot Sch Astronaut Shaanxi Prov Innovat Team Intelligent Robot Techno Xian 710072 Peoples R China

Aerial robotics for transporting suspended payloads as the form of freely-floating manipulator are growinggreat interest in recent years. However, the force/torque caused by payload and residual dynamics will introduce unmodeled perturbations to the aerial robotics, which negatively affects the closed-loop performance. Different from estimation-like methods, this letter proposes Neural Predictor, a learning-based approach to model force/torque induced by payload and residual dynamics as a dynamical system. It yields a hybrid model that combines the first-principles dynamics with the learned dynamics. The hybrid model is then integrated into a MPC framework to improve closed-loop performance. Effectiveness of proposed framework is verified extensively in both numerical simulations and real-world flight experiments. The results indicate that our approach can capture force/torque caused by suspended payload and residual dynamics accurately, respond quickly to the changes of them and improve the closed-loop performance significantly. In particular, Neural Predictor outperforms a state-of-the-art learning-based estimator and has reduced the force and torque estimation errors by up to 66.15% and 33.33% while requiring less samples.

关键词： Payloads Aerodynamics Torque Quadrotors Force Linear systems Predictive models Nonlinear dynamical systems Autonomous aerial vehicles Training learning Based control model learning for control model Predictive control

来源：评论

学校读者我要写书评

暂无评论

learning Deep Nets for Gravitational Dynamics With Unknown Disturbance Through Physical Knowledge Distillation: Initial Feasibility Study

引用

IEEE ROBOTICS AND AUTOMATION LETTERS 2021年第2期6卷 2658-2665页

作者： Lin, Hongbin Gao, Qian Chu, Xiangyu Dou, Qi Deguet, Anton Kazanzides, Peter Au, K. W. Samuel Chinese Univ Hong Kong Dept Mech & Automat Engn Hong Kong Peoples R China Chinese Univ Hong Kong Sch Sci & Engn Shenzhen Inst Artificial Intelligence & Robot Soc Shenzhen Peoples R China Chinese Univ Hong Kong Dept Comp Sci & Engn Hong Kong Peoples R China Johns Hopkins Univ Dept Comp Sci Baltimore MD 21218 USA

learning high-performance deep neural networks for dynamic modeling of high Degree-Of-Freedom (DOF) robots remains challenging due to the sampling complexity. Typical unknown system disturbance caused by unmodeled dynamics (such as internal compliance, cables) further exacerbates the problem. In this letter, a novel framework characterized by both high data efficiency and disturbance-adapting capability is proposed to address the problem of modeling gravitational dynamics using deep nets in feedforward gravity compensation control for high-DOF master manipulators with unknown disturbance. In particular, Feedforward Deep Neural Networks (FDNNs) are learned from both prior knowledge of an existing analytical model and observation of the robot system by Knowledge Distillation (KD). Through extensive experiments in high-DOF master manipulators with significant disturbance, we show that our method surpasses a standard learning-from-Scratch (LfS) approach in terms of data efficiency and disturbance adaptation. Our initial feasibility study has demonstrated the potential of outperforming the analytical teacher model as the training data increases.

关键词： Dynamics medical robots and systems model learning for control

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：