|
Journal of Zhejiang University SCIENCE A
ISSN 1673-565X(Print), 1862-1775(Online), Monthly
2022 Vol.23 No.1 P.14-26
A biomimetic robot crawling bidirectionally with load inspired by rock-climbing fish
Abstract: Using a unique adhesive locomotion system, the rock-climbing fish (Beaufortia kweichowensis) adheres to submerged surfaces and crawls both forwards and backwards in torrential streams. To emulate this mechanism, we present a biomimetic robot inspired by the locomotion model of the rock-climbing fish. The prototype contains two anisotropic adhesive components with linkages connected to a linear actuator. Each anisotropic adhesive component consists of one commercial sucker and two retractable bioinspired fin components. The fin components mimic the abduction and adduction of pectoral and pelvic fins through the retractable part to move up and down. The robot prototype was tested on vertical and inverted surfaces, and worked successfully. These results demonstrate that this novel system represents a new locomotion solution for surface movement without detachment from the substrate.
Key words: Fish kinematics; Adhesive locomotion mechanism; Fin rays motion; Climbing model; Bio-inspired robot
1福州大学电气工程与自动化学院,中国福州市,350108
2福州大学5G+工业互联网研究院,中国福州市,350108
3福州大学工业自动化控制技术与信息处理福建省高校重点实验室,中国福州市,350108
摘要:针对人–多机器人协同系统提出一种基于行为控制框架的带记忆强化学习任务管理器(RLTS)。由于重复的人工干预,现有人–多机器人协同系统决策时间成本高、任务跟踪误差大,限制了多机器人系统的自主性。此外,基于零空间行为控制框架的任务管理器依赖手动制定优先级切换规则,难以在多机器人和多任务情况下实现最优行为优先级调整策略。提出一种带记忆强化学习任务管理器,基于零空间行为控制框架融合深度Q-网络和长短时记忆神经网络知识库,实现任务冲突时最优行为优先级调整策略以及降低人为干预频率。当机器人在紧急情况下置信度不足时,所提带记忆强化学习任务管理器会记忆人类干预历史,在遭遇相同人工干预情况时重新加载历史控制信号。仿真结果验证了该方法的有效性。最后,通过一组受外界噪声和干扰的移动机器人实验,验证了所提带记忆强化学习任务管理器在不确定现实环境中的有效性。
关键词组:
References:
Open peer comments: Debate/Discuss/Question/Opinion
<1>
DOI:
10.1631/jzus.A2100280
CLC number:
Download Full Text:
Downloaded:
2905
Download summary:
<Click Here>Downloaded:
871Clicked:
4474
Cited:
0
On-line Access:
2024-08-27
Received:
2023-10-17
Revision Accepted:
2024-05-08
Crosschecked:
0000-00-00