【视频摘要】《中国科学：信息科学》呈献 | 北理工陈杰院士团队研究进展

更多精彩视频登陆网站

www.koushare.com

文章视频摘要

文章摘要

An optimization-based shared-control framework was proposed which contains explicit modeling of control objectives. This may help system designers to develop more flexible shared control controllers for multi-robot systems. Components of this framework were standardized and a class of collaboration costs is characterized that can render the system input-to-state stable under proper conditions.

正文

In recent years, researchers have been developing tools to allow human operators to work with multiple robots. To this end, results for autonomous systems can be helpful, e.g. controllability analysis, null-space approach, containment control. On the other hand, results from teleoperation systems show that the shared control method is helpful in producing safe and efficient humanrobot-interaction systems.
Among existing studies on shared control, most approaches can be seen as variants under the policy blending framework (1),
u(t) = (1 − λ(t))u_T(t) + λ(t)Ps(u_H(t)), (1) 

in which human input u_H(t) is first projected onto a safe/permissible space by Ps, then blended with an autonomous control signal u_T(t) by a weighting function λ(t). However, this kind of methods does not explicitly address the control objectives of u_H and u_T, and a blend of actions as in (1) may be unsatisfactory for both sides. With such a potential defect, this method may not always be suitable for multi-agent systems, since control objectives for such systems, like formation adjustments and group separations, can be complex.
To solve this problem, the method in is extended and an optimization based shared-control framework is proposed. In this framework, control objectives are expressed as cost functions and actions are computed by minimizing a blended cost. The framework is formulated using the model predictive control (MPC) method, which allows theoretical analysis on the stability of the closed-loop system. An experiment of one human operator controlling a four-robot system is also presented.
The shared control framework. In general, a shared control system tries to balance between a local task and a collaboration task. The local task considered here is to stabilize the system around the origin and the collaboration task is to comply with human control intentions.
For a system described by
x(k + 1) = f(x(k), u(k)), (2)
the proposed controller framework consists of three components, a task model, a human model, and a task balancing algorithm. These components are formulated as follows.

......

查看全文，请点击下方“阅读原文”！

SCIENCE CHINA Information Sciences (Sci China Inf Sci), cosponsored by the Chinese Academy of Sciences and the National Natural Science Foundation of China, and published by Science China Press, is committed to publishing high-quality, original results of both basic and applied research in all areas of information sciences, including computer science and technology; control science and engineering; information and communication engineering; microelectronics and solid state electronics, etc. SCIENCE CHINA Information Sciences is published monthly in both print and electronic forms. It is indexed by Science Citation Index Expanded (SCIE), Engineering Index (EI), Journal Citation Reports/Science Edition (JCR), etc.
蔻享科学平台秉承传播科学、共享科学、服务科学的理念，为广大科研工作者呈现最新科研动态。有发表的最新成果，大家可以录制视频摘要（时长不超过10分钟）发送至蔻享，蔻享免费为大家宣传。欢迎投稿！

◆【视频摘要】亚波长波导模式转换的瞬态过程研究：光学芯片中模式转换和太赫兹集成器件的新思路

◆【科学综述】硼烯及其在能源与催化领域的研究进展

为满足更多科研工作者的需求，蔻享平台开通了各科研领域的微信交流群。

进群请添加微信：18256943123

小编拉您入群哟！

未经许可，禁止转载本页内容

转载授权请微信联系：18256943123