师说 | 林天麟教授：科研中探险，不确定是有趣的源泉

专访

林天麟教授是香港中文大学（深圳）助理教授、电机电子工程师学会（IEEE）高级会员、深圳市人工智能与机器人研究院（AIRS）智能机器人研究中心主任。2020年，林天麟教授的研究论文“FreeBOT：能任意连接、自由组合的模块化自重构机器人——设计与实现”获得了IROS 2020的机器人机构设计最佳论文奖。近期，林教授及其团队在ICRA 2021发表了七篇论文。

林教授分别于2006年和2010年在香港中文大学获得一级荣誉学士学位和博士学位。其研究方向包括多机器人系统，新型移动机器人及人机协作等。他已发表2部专著，50多篇国际期刊论文及国际会议论文，获60多项授权专利。大部份论文均发表于机器人和自动化领域的顶级国际期刊及国际会议，如TRO，T-MECH，RA-L，JFR，ICRA和IROS等。他于2011年获得IEEE/ASME T-MECH最佳论文奖。他的研究成果还被包括Reuters，Discovery Channel和IEEE Spectrum在内的众多国际知名媒体报导。

Q：恭喜您的研究论文“FreeBOT：能任意连接、自由组合的模块化自重构机器人——设计与实现”获得了 IROS 2020 的机器人机构设计最佳论文奖，这是中国内地近十年来第二次获得 IROS 最佳论文奖项，对此您有想要分享的感言吗？

林天麟：首先要感谢IROS评委们对我们工作的认可，也要感谢 AIRS对我们科研团队的支持，更重要的是感谢科研团队每一位成员的付出，很高兴能看到团队的努力获得专家们的肯定。除了这个奖项外，我们也有多篇相关文章发表在这次IROS会议中。这些成果对我们这个刚组建不久的科研团队来说是一个很大的鼓舞，印证了我们选择的研究方向以及研究成果是具国际影响力的。在未来，我们团队会继续聚焦在自重构机器人及多机器人系统的研究，希望能持续为社会贡献更多更先进的技术及人才。

IEEE Spectrum 报道该研究成果

Q：您的FreeBOT机器人被誉为现实版“超能陆战队”，能够自主自由地组合成多种形状以满足不同的任务需求，具有实现自由组合机器人的巨大潜力。能否请您介绍一下 FreeBOT 的设计理念和技术特点？

林天麟：FreeBOT是一种模块化自重构机器人，模块化自重构机器人是由多台相同的小机器人构成，而小机器人之间更是可以自行通过物理上的连接去形成不同形状及功能的更大个体。理论上能成为一个任意变形的通用机器人。这种机器人有着极高的适应能力、自恢复能力及扩充能力等，特别适合应用在环境及任务不确定的工作上，如星际探索、灾区搜救等工作。

然而，要实现任意变形的通用机器人还有很多难关需要克服，需要长时间坚持投入才能一步一步攻克。其中一个难题是现有模块化自重构机器人的设计对连接条件非常苛刻，每一次连接的难度犹如进行太空站对接一样困难。使改变构形的任务变得寸步为艰。而FreeBOT的设计就针对这个问题进行了突破，利用了磁场及磁导球面任意点连接的方法，实现模块之间能瞬间任意连接，使机器人能实现更迅速、更灵活的变形。这种任意连接的设计主要是受到一种黏菌的启发。该黏菌是原生生物，其表面并没有固定的连接构造，但当有需要时，它们能聚集起来并依附在一起形成不同的形状及功能去应对环境的变化。因此，也启发了我们思考能否做出自由连接的设计应用在模块化自重构机器人上。

FreeBOT

Q：我们留意到从您在博士生期间做的仿生机器人爬树机器人TreeBot，到现在的模块化自重构机器人FreeBOT，一直都专注于在机器人设计上不断创新，请问您当初为什么会进入到这个领域，并一直坚持做机器人设计的？

林天麟：做喜欢的事情不能说是坚持，我只是到现在仍保持着研发机器人的兴趣。

我从小就对大自然中会到处走动的生物很感兴趣，喜欢观察不同生物的生活习性、如何应对大自然的各种挑战等，赞叹生物的各种神奇。因此，我当时最喜欢和成绩最好的是生物课。另一方面，我也很喜欢动手创造东西，经常动手制作一些有趣的玩意，很自然地就会去想我是否可按自己的想法去创造一些新的生物。往深层想，我认为生物的创造周期可能需要很漫长的时间才见效，而且创造生物也会有伦理道德的问题。再想想，其实创造有智慧会动的东西不一定只有生物方法，机器人也算是一种通过非生物手段的人造生命。而且机器人的开发周期更短，更快见成果。因此，年少性急的我就立志要往机器人的方向发展。高中选科时毅然放弃选读成绩一直最好的生物科而选择攻读数理科。到报考大学时，我专门挑选开设机器人研究方面专业的大学。经过了一番努力，有幸顺利地在心仪的大学修读喜欢的专业，更有幸的是遇到机器人领域的名师徐扬生院士，他目前担任香港中文大学（深圳）校长。从此，我的机器人研究生涯正式开始了，实现了年少时创造机器人的梦想。

林天麟教授

Q：您从事机器人研究这么多年来，觉得其中最有趣的是什么？

林天麟：科研就如同探险一样，你知道想要什么，但你不会知道最终会得到什么。

科研存在各种不确定性，而不确定性才是有趣的源泉。在研究的过程中，心情可能长时间都像坐过山车一样起起伏伏。面对一个科研问题，你思前想后，好不容易突然灵光一闪，想出了一个方法，当下觉得自己真是个天才！兴奋得睡不着觉。当验证方法的时候才发现原来行不通，心情又会瞬间跌落谷底，感觉自己原来是个大笨蛋……接下来，要不承认自己就是个大笨蛋，从此放弃；要不说服自己成功地排除了一个不可行的方法，离成功又近了一步，然后再重新上路思考新的方法，周而复始，直至成功。

当年做爬树机器人的这段时期是我科研研验中让我经历种种起伏时间最长最深刻的一段。最初的爬树机器人的设计与大家现在看到的版本可以说是大相径庭。研发过程中出现了超过十个完全不同的结构及方法，而且最终证明是失败的版本。回看这些“坐过山车”的过程，可以说是有笑有泪，百感交集。未到最后一刻，都不会知道成功的结构和设计是什么样的。

林天麟教授介绍爬树机器人

Q：您在AIRS所领导的智能机器人研究中心在过去一年里可谓硕果累累，不仅获批一项国家自然科学基金项目，还在多个顶刊和顶会发表多篇论文。可否分享您作为中心主任对中心发展的一些想法？

林天麟：智能机器人研究中心的目标是联合国际顶尖专家共同研发机器人前沿技术，推动新型机器人技术在全新领域的应用，面向社会普及机器人技术和人工智能知识，致力于打造以深圳市为核心的国际一流机器人技术研发及应用推广枢纽。中心自成立以来，持续引进顶尖人才并联同国际顶尖专家展开机器人前沿技术的研发工作，目前已渐见成果。希望随着持间的推移，我们的工作能续步落地，惠及社会，秉承 AIRS "for Society"的精神。

英文版请见下文

See below for English version

Q：Congratulations on receiving the IROS 2020 Best Paper Award on Robot Mechanisms and Design for your research paper "FreeBOT: A Freeform Modular Self-reconfigurable Robot with Arbitrary Connection Point – Design and Implementation". Is there anything you’d like to share with us?

Prof. Tin Lun Lam: Firstly, I’d like to thank the IROS committees for the recognition of our work, and AIRS for the support, and more importantly, I’d like to thank every member of our team for their dedication and hard work. Besides the award, our team also published a number of papers in the IROS 2020 conference, which is a great encouragement to us, a newly formed research team. It also gives us the confidence that the research direction we are working on is making an impact internationally. In the future, our team will continue to focus on the research of freeform robots and multi-robot systems, and we are ready to make more contributions to the society by developing advanced technologies and nurturing top-notch talents.

IEEE Spectrum reports on the paper

Q：The FreeBOT, hailed as a real-life "Big Hero 6", can be extended to different configurations to meet more functional requirements and accomplish multiple tasks, which demonstrates great potential to realize a freeform robotic system. Could you please introduce the design concept and technical features of FreeBOT?

Prof. Tin Lun Lam: FreeBOT is a kind of modular self-reconfigurable robot (MSRR). MSRR system consists of many homogeneous modules, which can connect with each other physically to arrange themselves into different configurations according to task requirements. Theoretically, it can be a universal robot that can be transformed at will. This kind of robot is recoverable, scaleable and has strong adaptability, which can be applied in scenarios with complex environments and tasks, such as interstellar exploration, rescue, etc.

However, there are still many challenges to overcome in order to realize a universal robot with the ability to self-configure, and it will take a long and persistent commitment to overcome them step by step. One of these challenges is that the connection between modules is time-consuming and as complicated as performing a space station docking, which makes it extremely difficult to rearrange into different configurations. The design of FreeBOT has tackled this problem. With the ferromagnetic shell and an internal magnet, FreeBOT can be connected freely at any point and to rearrange in an effective way with fewer physical constraints. The design of this arbitrary connection was inspired by slime molds. The slime mold is a protist that does not have a fixed connection structure on its surface, but when needed, it can gather and cling together to form different shapes and functions in response to changes of environment. This has inspired us to think about the possibility of designing arbitrary connection points for modular, self-reconfigurable robots.

FreeBOT

Q: We noticed that from the bionic robot TreeBot, a tree-climbing robot you made during your PhD, to this modular, self-reconfiguring robot FreeBOT, you have always focused on continuous innovation in robot design. Why did you choose to focus on robot design?

Prof. Tin Lun Lam: I made that choice out of passion. I’m simply enthusiastic about developing robots.

Ever since I was a child, I have always been interested in amazing creatures of nature. I loved to observe and always marvel at habits of different animals, and how they deal with the various challenges of nature. Therefore, I enjoyed and did best in Biology at that time. Besides, I also liked to make things and interesting handicrafts by myself. Naturally, I started to wonder if I could create brand new life in my own way. But on second thought, creating new life takes a long time and might involve some ethical issues; in fact, biology is not the only approach to create intelligent, moving things-robotics can also create artificial life within a much shorter period. So I decided to pursue a career in robotics. In high school, I chose to study mathematics and science instead of biology which was always my forte. When I applied for university, I also chose those that offer robotics courses. Fortunately, with great efforts, I was able to be admitted to my favorite university and major. More fortunately, I met academician Yangsheng Xu, a renowned professor in the field of robotics, who now serves as President of CUHK-Shenzhen. From then on, I began my career in robot research and is realizing my dream of inventing robots.

Prof. Tin Lun Lam

Q：What do you find most interesting in robot research over the years?

Prof. Tin Lun Lam: Research is like an adventure: you know what you want, but you never know what you will get in the end. There are all kinds of uncertainties in research, which is where the fun comes from. The research process can be an emotional roller coaster for you. You may feel like a genius when you have an “aha moment” after racking your brain and coming up with a solution. And when you find out that the idea doesn't work, you may instantly hit rock bottom and feel like an idiot. Then, you either admit that you're a fool and give up or you convince yourself that you've ruled out an unfeasible method and are one step closer to success. And then you move on to find new methods, back and forth, until you succeed. 

The process of creating TreeBot was the longest and most memorable emotional roller coaster I have ever experienced in my research career. The original design idea was very different from the version today. During the development process, there were over ten failed versions with completely different structures and methods. Looking back, the process is a mixture of laughter and tears. You never know what will eventually come out until the last moment.

Prof. Tin Lun Lam is introducing TreeBOT

Q：Under your leadership, Research Center on Intelligent Robots at AIRS has conducted fruitful work in the past year. For example, one of the projects has received research fund from the National Natural Science Foundation of China, and quite a number of papers have been published in top journals and conferences. What’s your take on the development of the Center, as the director?

Prof. Tin Lun Lam: The goal of the Center is to develop cutting-edge robotics technologies in collaboration with leading international experts, to promote the application of new technologies in new fields, to popularize robotics and artificial intelligence knowledge in the community, and to build a world-class hub for the development and application of robot technology with Shenzhen as its core. Since its establishment, the Center has been continuously nurturing top-notch talents and working with top international experts to develop cutting-edge robotics technologies, and has achieved some results. We hope that as time goes by, our work will continue to benefit the society and carry forward the "for Society" spirit of AIRS.