Scientific Symposium 2024

Abstract

Enhancing the capabilities of soft transducers while ensuring their compliance and reliability presents notable challenges. This talk will discuss bridging the gap between traditional hard and soft robotics by integrating shape memory polymers with various soft actuation mechanisms, including pneumatics, dielectric elastomers, and soft electromagnetic actuators. The application of this integration extends to microfluidic valve arrays and reconfigurable grippers. Additionally, the integration of soft sensors introduces further complexities, particularly in mitigating electromagnetic interferences. The current research directions involve exploring scalable hybrid actuation systems with self-sensing capabilities and harnessing electroadhesion for force amplification and multiplexing for dexterous robotics and haptics.

Biography

Bekir Aksoy is currently a Research Associate at Northwestern University's Center for Robotics and Biosystems. He completed his PhD at EPFL, Switzerland, specializing in the field of shape programmable multi-stable polymer actuators. Bekir earned his MSc and BSc degrees from Koc University in 2017 and 2014, respectively. During his time at Koc University, he focused on developing MEMS traction sensors for studying cell mechanics. His research interests encompass different areas, including multi-stable polymer actuators, microfluidics, MEMS, and soft force sensors. He has actively contributed to projects involving collaboration with both industrial and academic partners. His current research aims to develop high-performance actuators for dexterous robotics and haptics.

Abstract

Robotic systems need to have adaptable mechanical behavior in order to achieve complex tasks in unstructured, constricted, and delicate environments. For example, in the human body, surgical devices need to be compliant during navigation and deployment to ensure safety, but need to stiffen in particular directions during task execution to ensure precision. This type of adaptability can be achieved on a structural level with jamming, which is a mechanical phenomenon with which the bulk mechanical behavior of a cluster of constituents can be altered when the coupling between the individual constituents is modified. This talk will demonstrate how jamming can be leveraged to achieve multidimensional tunable mechanical behavior (e.g. tunable stiffness, damping, elasticity/plasticity, shape memory) for effective and versatile robot-environment and robot-human interactions. It will outline strategies that enable us to predict, program, and optimize the behavior of these structures for target applications, and discuss how key parameters such as constituent geometry or actuation method impact resulting performance. Several jamming-based composites will be presented for applications ranging from robotic manipulators, to medical robots, to interactive art. Finally, the talk will discuss the future of such robotic composites which can be manufactured or can self-assemble from individual constituents into a desired morphology, can have tunable task-specific mechanical properties, and then disassemble for future re-assembly and reuse.

Biography

Dr. Buse Aktas is an engineer and artist working at the intersection of robotics, mechanics, materials, and design with a focus on developing composite material systems with actively tunable properties. After her Bachelor’s Degree in Mechanical and Aerospace Engineering with a certificate in Visual Arts at Princeton University, she received a Master of Arts in Design at Kadir Has University, focusing on cultural heritage and conducting fieldwork in the form of apprenticeships with traditional craftspeople (locksmiths, broommakers, etc.). She then received her Ph.D at Harvard University in Materials Science and Mechanical Engineering with a secondary field in Critical Media Practice. Her doctoral research in the Harvard Biorobotics Lab focused on developing jamming-based structures for multi-degree-of-freedom robotic interactions. She is currently a recipient of the ETH Postdoctoral Fellowship, and is developing magnetic composites which enable a novel untethered actuation method for the assembly, tuning, and disassembly of robotic composites, as a part of the Multi-Scale Robotics Lab at ETH Zurich.

Abstract

Robots are traditionally designed with immutable physical hardware and control policies that make them specialized for repetitive tasks and structured environments. This talk presents work toward robots that actively change shape to accomplish a variety of tasks in diverse environments. Shape-changing robots are pursued at two levels. First, I will discuss the design of shape-morphing components in the form of variable-trajectory soft actuators. Component-level analysis leads to insight into how stiffness differentials can yield myriad deformations. Inverse models that recapitulate shape-morphing components’ highly nonlinear geometric and material behavior allow for systematic mechanical programming of shape-morphing robotic function. These foundational studies inform the second part of the talk, in which I will discuss how shape-morphing components can be applied to improve robot performance in locomotion and combined locomotion-manipulation tasks.

Biography

Robert Baines is a postdoctoral fellow in the Robotic Systems Lab at ETH Zurich. He received his PhD in Mechanical Engineering and Materials Science at Yale University. With the goal of improving robots’ versatility across different tasks and environments, his research seeks to endow robots with shape-changing capabilities, touching upon topics in soft-bodied actuation and sensing, locomotion, and inverse design. He is the recipient of several awards, including the NSF Graduate Research Fellowship, ETH Postdoctoral Fellowship, and Branco Weiss Society in Science Fellowship. His work on robots that “evolve on demand” was featured on the cover of Nature and disseminated by dozens of media outlets worldwide.

Abstract

We walk over complex terrain, like trails or stairs, seemingly effortlessly, as our neuromuscular system synergistically coordinates two key mechanisms (muscle mechanics and sensory feedback) to maintain agility. Yet, we do not fully understand how these mechanisms contribute, adapt, and compensate to accomplish agility. Orchestrating the primary motor in movement – muscle – is more challenging in unpredictable (i.e., more natural) compared to solid, level terrain. Hence, at the frontiers of biology, muscle physiology, and movement sciences, it is still unclear how animals effectively integrate muscle mechanics – shaped by feedforward control – and sensory feedback to achieve agile locomotion. In my research, I have used unique experimental approaches to explore how guinea fowl, a model for bipedal locomotion, integrate muscle mechanics and sensory feedback to maintain robust locomotion. My work elucidated a gait-specific control mechanism in walking and running over obstacles as well as that it revealed that guinea fowl muscle operating ranges provide a safety factor to potential unexpected perturbations. Lastly, this work showed a modular task-level control of leg length and leg angular trajectory during navigating speed perturbations while walking, with different neuromechanical control and perturbation sensitivity in each actuation mode. Currently, I work on experimental data informed musculoskeletal modeling to further probe at underlying mechanisms to agile locomotion. Elucidating frameworks of functions, adaptability, and individual variation across neuromuscular systems will provide a stepping-stone for understanding fundamental muscle mechanics, sensory feedback, and neuromuscular health. This research has the potential to reveal the functional significance of individual morphological, physiological, and neuromuscular variation in relation to locomotion and agility. Thereby, it provides foundational knowledge for the development of dynamic assistive devices and robots that can navigate through complex terrains.

Biography

Janneke is currently a postdoctoral scholar at KU Leuven, working with Prof Friedl de Groote. She combines state-of-the-art computational modeling approaches, informed and validated by experimental data, to better understand underlying mechanisms to agile locomotion. Before coming to Belgium, Janneke was a postdoctoral scholar at the University of California, Irvine, working with Prof Monica Daley. Here, Janneke focused on how the neuromusculoskeletal system enables navigating complex terrain and how it is impacted by nerve injuries (i.e., impacting neural feedback loops), using guinea fowl as a model species. During this time, Janneke was also awarded the 2021/2022 Graduate Women in Science fellowship, for which she worked with Prof Kiisa Nishikawa at Northern Arizona University, to further understand the decoupling of muscle activation and forces using the novel Avatar approach. Janneke completed her Ph.D. from the University of Idaho (Moscow, USA) working with Prof Craig McGowan. Her Ph.D. projects included studying the jumping and tail-assisted aerial reorientation abilities of desert kangaroo rats to elucidate the musculoskeletal biomechanics and control of these extreme organismal behaviors. Janneke is a divisional assistant editor of Integrative and Comparative Biology (the Society for Integrative and Comparative Biology journal), an Outside JEB (Journal of Experimental Biology) author, and first-author on a paper on how professional societies can be more welcoming to parent members.

Abstract

Lower limb amputations and neuromuscular impairments severely restrict mobility, necessitating advancements beyond conventional prosthetics. Motorized bionic limbs offer promise, but their utility depends on mimicking the evolving synergy of human movement in various settings. While finite-state machines are adequate for rudimentary scenarios their construction becomes infeasible for finer gait phase resolution or multiple locomotion tasks.

This talk explores learning-based approaches that draw upon human demonstrations to emulate the complex synergy of human gait. By conceptualizing human ambulation as a continuum of gait states, we aim to develop a digital twin capable of seamlessly controlling a prosthetic limb. I will outline how models tailored for individuals with lower-limb impairments are constructed and adapted over time to meet evolving gait patterns.

The discussion will extend to the models' adaptability to varied locomotion tasks, such as ascending slopes and climbing stairs, through the examination of model architectures that balance shared and task-specific learning. Furthermore, I will introduce strategies aimed at enhancing these models’ adaptability through continual learning and experience replay. Despite these models' proficiency in preserving previous task knowledge, they often fall short in forward planning.

To bridge this gap, I will present a multitask model that merges anticipatory planning with experience replay. This approach not only maintains past learning but also improves the model's ability to forecast future movements by learning from previous prediction errors. To validate their practicality and resilience, these models were tested against adversarial perturbations, noise, and distribution shifts. Finally, I will share insights from real-time and pilot user experiments, demonstrating the real-world potential of these advanced prosthetic control strategies, and provide a brief outline of future research directions.

Biography

Sharmita Dey completed her Master's degree at the Institute of Artificial Intelligence, TU Dresden, Germany. She conducted her master's thesis, titled "Generalized Decoding of Control Signals from Surface Electromyography Signals," at the DLR, Institute of Robotics and Mechatronics, Oberpfaffenhofen, Germany. Following this, she earned her PhD degree from the Department of Computer Science, University of Goettingen, Germany, focusing on "Learning-Based Biomimetic Strategies for Developing Control Schemes from Lower Extremity Rehabilitation Robotic Devices," for which she was awarded summa cum laude. Parts of her doctoral work has been patented and published as scientific papers. During her doctoral studies, she was selected for a visiting researcher's position with the NASA Jet Propulsion Laboratory, where she contributed to projects aimed at enhancing the traversability of ground robots in perceptually degraded environments. In this role, she became a part of team CoSTAR taking part in the DARPA Subterranean Robotics Challenge 2021. Currently, she holds a Postdoctoral position at the Institute of Computer Science, University of Goettingen, Germany where her research centers on label-efficient learning and foundational models for action understanding as well as learning of interactions and dynamics of moving objects. Her research interests lie in developing robust and adaptive models for intelligent agents using strategies inspired by self-supervised, continual, active, and model-based reinforcement learning. She is also a part of voluntary groups such as Women in Machine Learning (WiML) and Machine Learning for Health (ML4H), where she served as a mentor.

Abstract

With over half a billion years behind them, fish have evolved to swim with remarkable efficiency, agility, and stealth in their three-dimensional aquatic world. Given this, it's natural that engineers often look to fish for inspiration when developing efficient underwater propulsion systems. Over the years, roboticists have been inspired by these biological marvels to design fish-like robots that mimic real fish in terms of morphology, locomotion, movement, and hydrodynamics. Interestingly, the trend has recently shifted from merely drawing inspiration from biology to using robotics as a tool for better understanding biological processes. In this talk, I will first discuss our approach to designing and controlling these robotic fish, rooted in the concept of bio-inspiration. I will then provide examples of how we use both real and virtual robots to explore the mechanisms of collective behavior in schooling fish, specifically for large and small Reynolds numbers, respectively. To conclude, I'll offer a glimpse into my current and future endeavors in the realms of robotics, hydrodynamics, and biology.

Biography

Liang Li is a Project Leader (PI) at the University of Konstanz and the Max Planck Institute for Animal Behavior. He earned his Bachelor's degree in Automation from Chongqing University in 2011 and his PhD in General Mechanics and the Foundations of Mechanics from Peking University in 2017. Between February 2017 and June 2021, he served as a Postdoctoral Research Fellow in the Department of Collective Behaviour at the Max Planck Institute for Animal Behavior in Konstanz, Germany. His research focuses on bio-inspired robotics, swarm robotics, collective behavior in hybrid animal-robot systems, and bio-fluid dynamics in fish schools. Liang Li has contributed to over 37 peer-reviewed journals, including prestigious publications such as Nature Communications and PNAS. He is a senior membership with IEEE and has received science awards from the Messmer Foundation, in addition to being recognized as an outstanding reviewer for the Institute of Physics (IOP). Furthermore, he is an active member of the education committee of the International Society for Bionic Engineering.

Abstract

Imitating nature in robotics, particularly its intricate biology, remains a significant challenge. Despite the progress in bioinspired soft robotics expanding our capabilities, accurately replicating the functionalities of living systems is still beyond our current manufacturing and material science. A promising direction is the direct integration of living systems into engineering, leveraging the inherent capabilities and form factors of biological entities to surpass existing material and manufacturing limitations. My talk explores the integration of living biological components with robotic systems, exploring 'beyond biomimicry' to achieve a profound synergy between biology and engineering. I will concentrate on employing living organisms in biohybrid robotic systems to utilize biological advantages, such as bioelectricity and multifunctional sensing. My work emphasizes the incorporation of fungal mycelium into robotic frameworks, enabling systems with unparalleled sensing, processing, and response capabilities. Furthermore, I will address the challenges associated with using mycelium with robotics, the maintenance and interfacing of biological components, and the broader implications of crafting biohybrid machines. My research aims to bridge the divide between biological science and robotic engineering with the ultimate vision of developing the next generation of biohybrid robots.

Biography

Anand is a multidisciplinary researcher in the fields of biomimetics and biohybrid soft robotics, with a passion for bridging the gap between biology and robotics. He earned his PhD from Scuola Superiore Sant’Anna & the Italian Institute of Technology, Italy. In 2018, he joined Cornell University as a postdoctoral researcher in Dr. Robert Shepherd’s group and is currently serving as a Research Associate. Anand has been awarded a Postdoctoral Fellowship from the Engineering Living Material Institute, as well as a Research Trainee Grant from the Center for Research on Programmable Plant Systems at Cornell University. Anand’s research encompasses 3D printing & Soft Robotics, Biomimetics & Biohybrid Robotics, and Digital Biology & Agriculture, focusing on plants and fungi as model organisms for new robotic designs and materials. He has contributed to writing several prestigious successful grants, including the NSF SITs Program, NIFA NRI initiatives, AFOSR’s SURI Program, and Cornell’s CIDA Initiatives, and he is a Co-PI on the recently awarded ARPA-E GOPHURRS Program. His work has been featured in NSF’s Discovery episodes, CNN, BBC, The Guardian, Scientific American, and Forbes. Driven by his vision for a future where technology and nature coexist in harmony, Anand continues to research the untapped potential of the natural world, aiming to develop the next generation of robotics that can improve the human condition and perform tasks in life-threatening environments.

Abstract

Plants exhibit intelligent behaviors as they explore and colonize their surroundings, overtaking obstacles and moving in complex and mutable environments. While animals employ mobility, plants rely on apical growth, expanding through cell division and elongation from the extremities of their organs. I will show how to mimic in a robot the growing-from-the-tip by embedding a miniaturized 3D printing process in its head. This allows the robot to grow and build its body through layer-by-layer addition of thermoplastic material. The robot autonomously directs itself thanks to the integration of a plant-inspired growth-driven behavioral control. I will show this strategy to be viable for robot navigation and to confer an intrinsic adaptation and resilience to the growing robot in unstructured environments. Additionally, I will explore various plant movements and adaptation mechanisms, such as circumnutations and radial expansion, which have been hypothesized to improve robot movements, particularly in soil. By delving into the plant intelligence paradigm, this talk aims to showcase how plant growth-driven adaptive behaviors can suggest novel design, modeling, and control of robots to leverage the exploration in unstructured scenarios. This approach not only provides novel insights into robot design but also fosters innovative perspectives for preservation and management plans, introducing new applications for engineering-driven plant science.

Biography

Dr. Emanuela Del Dottore received her bachelor’s and master’s degrees in Computer Science and Information Technology from the University of Pisa. In 2017, she received her Ph.D. in Biorobotics from Scuola Superiore S. Anna with a thesis focused on the elaboration and implementation of algorithms and control strategies for plant-inspired robotic systems. She has industry and research experience in embedded systems, programming, systems kinematics, image processing, and data analysis. In her current position as a Researcher in the Bioinspired Soft Robotics laboratory of the Istituto Italiano di Tecnologia (IIT), Dr. Del Dottore is the co-advisor of several Ph.D. students, fellows, and research interns and supports the work of other postdocs in the field of plant-inspired robotics, growing robots, and soft robotics. She has contributed to the writing and management of successful national and international projects. Her current research focuses on formulating and implementing control and decision-making strategies for bioinspired robots based on plant-inspired growth-driven behaviors. In this context, she proposes, investigates, and validates behavioral hypotheses concerning the movement, exploration strategies, and the emergence of high-level behaviors in plants and extracts novel rules for the movements of biomimetic robotic systems. She is passionate about robotics, behavior analysis, swarm and collective behavior, and intelligent systems.

Abstract

Ingestible robots emerge as a transformative solution in combating chronic health conditions such as diabetes, IBD, obesity, and cancer. Despite their promise, these devices face significant spatiotemporal challenges within the gastrointestinal (GI) tract, including issues with retention, locomotion, and energy. This talk will discuss fundamental advances in robotic materials and structures, providing these medical robots with de novo, bioinspired capabilities for actuation, anchoring, and powering. I will highlight two preclinical innovations from my research: (i) e-GLUE, an electroadhesive hydrogel interface designed for enhanced mucosal retention and theranostics, and (ii) IngRI, an ingestible, battery-free, tissue-adhering robotic interface for non-invasive and prolonged electrostimulation of the gut. These two platforms demonstrate the potential for effectively managing various GI conditions through the development of bioinspired robots and robotic biomaterials.

Biography

Dr. Binbin Ying is a Banting Fellow and a postdoctoral researcher in the labs of Prof. Giovanni Traverso and Prof. Robert Langer at MIT and Brigham and Women's Hospital. Dr. Ying earned his PhD degree from McGill University under the supervision of Profs. Xinyu Liu and Jianyu Li in 2020 and worked as a visiting scientist at the University of Toronto from 2018 to 2020. His research has focused on the development of soft robotics and biomaterials for healthcare. Dr. Ying has published over 20 scientific papers. His research has culminated in numerous honors and awards, including Banting Postdoctoral Fellowship, NSERC Postdoctoral Fellowship, Chinese Government Award for Outstanding Self-Financed Students Abroad, and 2020 Materials Horizons Outstanding Article Award. Besides research, Dr. Ying is also passionate about mentorship and serving the research community. He co-founded The Martlets Society, a webinar series for young scholars to connect and exchange expertise.

Abstract

In the last fifteen years, medical robotics has surged, with thousands of clinical systems installed world-wide, and millions of procedures performed. The emergence of new actuation paradigms, and the increasing available computational power and AI-based tools for the design and control of robotic systems have created unprecedented opportunities for their application in medicine. Exciting challenges must be addressed to harness this potential to the benefit of health care.

In this talk, I will share with you my ongoing journey in medical robotics research. We notably demonstrated the design and control of soft robotic devices, for which fundamental advances in their simulation and characterization are needed to fully leverage them in minimally invasive procedures. Our research also focused on developing electromagnetic navigation systems to steer magnetic continuum robots with various levels of autonomy remotely. This journey led me to collaborate with interdisciplinary teams across the world, gathering engineers, clinicians and researchers to translate these development to clinical procedures. Promising applications include neurovascular procedures, cardiac ablations, and fetal surgeries. These efforts started making their way to the clinic, with two spin-off companies from our group currently working toward the commercialization of robotics systems based on our research.

Biography

Quentin Boehler is a senior researcher at the Multi-Scale Robotics Lab, ETH Zurich, where he has been leading the medical robotics activity of the group since 2020. He was born in Strasbourg in 1990. In 2013, he received an engineer’s degree in mechatronics from INSA Strasbourg, and an M.Sc. degree in robotics. He received a Ph.D. degree in robotics in 2016 from the University of Strasbourg, with a doctoral work focused on tensegrity mechanisms and variable stiffness devices for MR-compatible robotics. He was awarded the Best thesis Award from the research commission of the University of Strasbourg, and the First Prize at the 2016 Ph.D. thesis awards organized by the French Robotics Research Group. In 2017, he joined the Multi-Scale Robotics Lab at ETH Zurich as a postdoctoral associate. His current research is on magnetic actuation for medical robotics, including the development and analysis of electromagnetic navigation systems, and of soft magnetic robots for minimally invasive medical interventions.

Abstract

In an era where integration of Artificial Intelligence into society parallels essential utilities like water and electricity; its influence extends across education, healthcare, hospitality, and entertainment. It is a reality that is inevitable and with that, it is also becoming the go-to technology for powering social machines whether for smartly navigating among humans, or interacting with humans in social contexts. I believe that the imperative lies not in questioning its integration but in optimizing its assimilation into our societal fabric. To this end, this talk will delve into some of my doctoral and post-doctoral research that focuses on both aspects of perception of human behaviors by machines, and generation of machine behaviors to design human centric solutions for human-machine interaction. Particularly I will talk about how we can enhance human-robot interaction by (i) modeling multimodal human behavior in a data driven manner to develop goal-oriented social robots, in an iterative manner in both educational and therapy settings, (ii) engaging end-users and stake-holders in the process such as students, teachers, special needs individuals, therapists, etc, and (iii) generating machine behaviors that leverage machines own abilities without the aim of replicating humans. I will present several past and ongoing studies, and the corresponding findings.

Biography

Jauwairia Nasir is a postdoctoral fellow at the Chair of Human-Centered Artificial Intelligence (HCAI) at University of Augsburg since 2023. She earned her PhD, in 2022, at the Computer Human Interaction for Learning and Instruction (CHILI) lab at EPFL, Lausanne, Switzerland, in which the Productive Engagement framework was introduced for social robots, which led to several accolades including the RSJ Pioneering Research Award in robot and human interactive communication at RO-MAN 2023. During her PhD, she was an EU ITN Horizon 2020 Marie Curie fellow at ANIMATAS under which she has been a visiting scholar at the Sorbonne University, Paris, France. She received her masters degree, under the Korean Government Scholarship Program (KGSP), from KAIST, Republic of Korea. Previously, she completed her bachelor's degree at NUST School of Electrical Engineering and Computer Sciences in Pakistan. Her research interests broadly include human-machine interaction, social robotics, multimodal behavioral analysis, and machine learning applied in the fields of typical and atypical education and health care. She also serves as the head of initiatives at the WAILabs, part of the global non-profit Women in AI.

Abstract

The space environment is unique, remote, and unpredictable, and the ability to manufacture in situ offers unique opportunities to address these needs as they arise. However, these challenges are often mirrored on Earth. In hospitals, disaster zones, low resource environments and laboratories, the ability to manufacture customized artefacts at points of need can significantly enhance our ability to respond to unforeseen events. In this talk, I will introduce digital fabrication platforms with codeveloped software and hardware that draw on tools from robotics and human-computer interaction to automate manufacturing of customized artefacts at the point of need. Highlighting three research themes across fabrication machines, programmable materials, and modular assembly, the talk will cover a digital fabrication platform for producing functional robots, a method for programming magnetic material to selectively assemble, and a modular robotic platform for in-space assembly previously deployed in microgravity.

Biography

Martin Nisser is a PhD Candidate in the MIT Computer Science and Artificial Intelligence Laboratory, working with Professor Stefanie Mueller. He holds degrees from MIT, ETH Zurich, and The University of Edinburgh, and has interned and held staff appointments at The European Space Agency, Tesla Motors, Harvard University, and the Boston Dynamics AI Institute. He is a Sweden-America fellow, a Bernard Gold fellow, and his work has appeared in media including BBC News, The NBC Daily Show, The Washington Post, NASA TV, and Popular Science.