全软体动物运动控制对仿生机器人设计的启发  

Soft-body animal motion control and its inspiration fordesigning biomimetic robots

作　　者：龚哲峰 郑能干[3] 陆豪健 GONG ZheFeng;ZHENG NengGan;LU HaoJian(School of Brain Science and Brain Medicine,Zhejiang University,Hangzhou 310058,China;Department of Neurology,the Fourth Affiliated Hospital,Zhejiang University School of Medicine,Yiwu 322000,China;Qiushi Academy for Advanced Studies,Zhejiang University,Hangzhou 310007,China;School of Control Science and Engineering,Zhejiang University,Hangzhou 310007,China)

机构地区：[1]浙江大学脑科学与脑医学学院,杭州310058 [2]浙江大学医学院附属第四医院神经内科,义乌322000 [3]浙江大学求是高等研究院,杭州310007 [4]浙江大学控制学院,杭州310007

出　　处：《中国科学：生命科学》2024年第2期308-324,共17页Scientia Sinica(Vitae)

基　　金：浙江省重点研发项目(批准号:2022C01022);国家重点研发计划(批准号:2020YFB1313501);之江实验室项目(批准号:2020KB0AC02);国家自然科学基金(批准号:61972347,T2293720,T2293721);浙江省自然科学基金(批准号:LR19F020005);中央高校基本业务费(批准号:226-2022-00051);浙江省属基本科研业务专项资金(批准号:2021XZZX021)资助。

摘　　要：如何实现灵活准确的运动控制是仿生软体机器人研究中的一个重要难题.但对全软体动物而言,经过长期进化选择,其神经系统运动控制功能和骨骼肌肉系统的运动执行功能之间的优越配合,可以产生灵活多样、收放自如的运动,且具有良好的环境适应性.因此,对软体动物的运动控制机制进行深度仿生研究可以为提升软体机器人的性能提供新的思路.本文选择水母、水蛭、线虫、果蝇幼虫等无脊椎全软体动物作为分析对象,介绍其运动行为的神经肌肉基础,并结合相关神经力学模型介绍全软体动物运动神经控制机制的特性.通过对比不同的仿生机器人与全软体动物原型,提出了借鉴神经力学模型对仿生机器人设计进行改进的思路.最后,分析了现有全软体动物运动的神经力学模型的不足,提出了系统性研究全软体动物运动控制机制和神经力学建模的思路,并展望了软体机器人进行运动控制仿生的方式以及高性能软体机器人的潜在应用场景.Motion control is of great challenge for bionic soft robotics. For soft-body animals, their motions can be realized smoothly anddextrously under various modes. After millions of years’ evolution, their nervous system and motor system are highly compatible toenable excellent motion adaptativity to different environmental conditions. Mimicking the motion control of soft-body animals couldprovide new clues to improve the performance of soft robots’ motion control. Here we investigate several soft-body invertebrateanimals such as jellyfish, C. elegans, Leech, and Drosophila larva and introduce their behavioral motion modes, the neuromuscularbasis, as well as related neuromechanical models for the neuromuscular mechanisms underlying motion control. Through thecomparative analysis between the robots and the animal prototype, we put forward the idea of improving the design of biomimeticrobots based on the neuromechanical models. Finally, we analyzed the shortcomings of the existing neuromechanical models of softbody animals, proposed the future directions in systematic studies on motion control mechanism and neuromechanical modeling insoft-body anmials, and also predicted future applications of the soft robotics.

关 键 词：全软体动物 运动控制 神经力学模型 仿生软体机器人 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]