Suppr超能文献

比较多连杆仿鱼机器人中不同电机控制方案的转弯性能。

Comparing the turn performance of different motor control schemes in multilink fish-inspired robots.

机构信息

University of Akron, Akron Ohio, United States of America.

出版信息

Bioinspir Biomim. 2021 Apr 12;16(3). doi: 10.1088/1748-3190/abe7cc.

Abstract

Fish robots have many possible applications in exploration, industry, research, and continue to increase in design complexity, control, and the behaviors they can complete. Maneuverability is an important metric of fish robot performance, with several strategies being implemented. By far the most common control scheme for fish robot maneuvers is an offset control scheme, wherein the robot's steady swimming is controlled by sinusoidal function and turns are generated biasing bending to one side or another. An early bio-inspired turn control scheme is based on the C-start escape response observed in live fish. We developed a control scheme that is based on the kinematics of routine maneuvers in live fish that we call the 'pulse', which is a pattern of increasing and decreasing curvature that propagates down the body. This pattern of curvature is consistent across a wide range of turn types and can be described with a limited number of variables. We compared the performance of turns using each of these three control schemes across a range of durations and bending amplitudes. We found that C-start and offset turns had the highest heading changes for a given set of inputs, whereas the bio-inspired pulse turns had the highest linear accelerations for a given set of inputs. However, pulses shift the conceptualization of swimming away from it being a continuous behavior towards it being an intermittent behavior that is built by combining individual bending events. Our bio-inspired pulse control scheme has the potential to increase the behavioral flexibility of bio-inspired robotic fish and solve some of the problems associated with integrating different swimming behaviors, despite lower maximal turning performance.

摘要

鱼类机器人在探索、工业、研究等领域有许多潜在的应用,其设计复杂性、控制和能够完成的行为也在不断增加。机动性是鱼类机器人性能的一个重要指标,有几种策略被应用。到目前为止,鱼类机器人机动的最常见控制方案是偏移控制方案,其中机器人的稳定游动由正弦函数控制,转弯则通过偏向一侧或另一侧的弯曲来产生。一个早期的仿生转弯控制方案是基于活体鱼中观察到的 C 型逃跑反应。我们开发了一种控制方案,该方案基于活体鱼常规机动的运动学,我们称之为“脉冲”,这是一种曲率逐渐增加和减小的模式,沿着身体传播。这种曲率模式在广泛的转弯类型中是一致的,可以用有限的几个变量来描述。我们比较了这三种控制方案在不同持续时间和弯曲幅度下转弯的性能。我们发现,对于给定的输入集,C 型和偏移转弯的航向变化最大,而仿生脉冲转弯的线性加速度最高。然而,脉冲将游泳的概念从连续行为转变为由单个弯曲事件组合而成的间歇行为。尽管最大转弯性能较低,但我们的仿生脉冲控制方案有可能提高仿生机器鱼的行为灵活性,并解决与整合不同游泳行为相关的一些问题。

文献AI研究员

20分钟写一篇综述,助力文献阅读效率提升50倍。

立即体验

用中文搜PubMed

大模型驱动的PubMed中文搜索引擎

马上搜索

文档翻译

学术文献翻译模型,支持多种主流文档格式。

立即体验