Cynthia Sung

Robot-aided exploration of planetary surfaces is essential for understanding geologic processes, yet many scientifically valuable regions, such as Martian dunes and lunar craters, remain hazardous due to loose, deformable regolith. We present a scout-rover cooperation framework that expands safe access to such terrain using a hybrid team of legged and wheeled robots. In our approach, a high-mobility legged robot serves as a mobile scout, using proprioceptive leg-terrain interactions to estimate regolith strength during locomotion and construct spatially resolved terrain maps. These maps are integrated with rover locomotion models to estimate traversal risk and inform path planning. We validate the framework through analogue missions at the NASA Ames Lunar Simulant Testbed and the White Sands Dune Field. Experiments demonstrate (1) online terrain strength mapping from legged locomotion and (2) rover-specific traversal-risk estimation enabling safe navigation to scientific targets. Results show that scout-generated terrain maps reliably capture spatial variability and predict mobility failure modes, allowing risk-aware path planning that avoids hazardous regions. By combining embodied terrain sensing with heterogeneous rover cooperation, this framework enhances operational robustness and expands the reachable science workspace in deformable planetary environments.

Robot-aided exploration of planetary surfaces is essential for understanding geologic processes, yet many scientifically valuable regions, such as Martian dunes and lunar craters, remain hazardous due to loose, deformable regolith. We present a scout-rover cooperation framework that expands safe access to such terrain using a hybrid team of legged and wheeled robots. In our approach, a high-mobility legged robot serves as a mobile scout, using proprioceptive leg-terrain interactions to estimate regolith strength during locomotion and construct spatially resolved terrain maps. These maps are integrated with rover locomotion models to estimate traversal risk and inform path planning. We validate the framework through analogue missions at the NASA Ames Lunar Simulant Testbed and the White Sands Dune Field. Experiments demonstrate (1) online terrain strength mapping from legged locomotion and (2) rover-specific traversal-risk estimation enabling safe navigation to scientific targets. Results show that scout-generated terrain maps reliably capture spatial variability and predict mobility failure modes, allowing risk-aware path planning that avoids hazardous regions. By combining embodied terrain sensing with heterogeneous rover cooperation, this framework enhances operational robustness and expands the reachable science workspace in deformable planetary environments.

Salps are underwater invertebrates considered to be among the world's most energy-efficient examples of jet propulsion. They can swim as solitary individuals or as physically connected colonies, coordinating their jets to produce collective movement. Inspired by salps, we developed the SALP (Salp-inspired Approach to Low-energy Propulsion) system, where individual SALP robots can be physically connected into a multi-SALP group, and we investigate the coupled effects of physical arrangement and jet coordination on the swimming performance and energy efficiency of a two-SALP system. We conduct free swimming tests to evaluate locomotion performance metrics and find that the two-SALP system, when properly coordinated, is able to swim with 15.7% higher speed and 11.3% lower cost of transport than the single SALP. Supporting flow characterization experiments using particle image velocimetry reveal vortex ring structures emanating from robot SALP nozzles. The data suggest that propulsion performance is affected by the spatial arrangement of the vortex ring structure. In particular, we find that SALP systems that produce a parallel vortex ring arrangement produce less vortex circulation and impulse than an in-series vortex ring arrangement. Overall, the SALP system is a useful platform for exploring salp-inspired multi-jet locomotion strategies, enabling decoupling of physical and control parameters to expose underlying locomotion physics in ways that are difficult with the biological salp. These insights advance our understanding of multi-jet locomotion and support the development of more energy-efficient jet-propelled underwater robots in the future.

Robust locomotion over challenging and unstructured terrain requires the ability to design and change the mechanical response of a robot. I will discuss recent work from my group in which we explore design and control algorithms for soft and compliant legged robots constructed from origami. These algorithms allow us to manipulate the passive response of the robot to enable dynamical running and hopping, passive perching, and other types of deformations and interactions. I will also show how these models extend to traditional rigid robots operating in soft environments, such as sandy ground, to produce more robust navigation.

We present the Magnetic Origami Reprogram-ming and Folding System (MORF), a magnetically repro-grammable system capable of precise shape control, repeated transformations, and adaptive functionality for robotic applications. Unlike current self-folding systems, which often lack re-programmability or lose rigidity after folding, MORF generates stiff structures over multiple folding cycles without degradation in performance. The ability to reconfigure and maintain structural stability is crucial for tasks such as reconfigurable tooling. The system utilizes a thermoplastic layer sandwiched within a thin magnetically responsive laminate sheet, enabling structures to self-fold in response to a combination of external magnetic field and heating. We demonstrate that the resulting folded structures can bear loads over 40 times their own weight and can undergo up to 50 cycles of repeated transformations without losing structural integrity. We showcase these strengths in a reconfigurable tool for unscrewing and screwing bolts and screws of various sizes, allowing the tool to adapt its shape to different bolt sizes while withstanding the mechanical stresses involved. This capability highlights the system's potential for task-varying, load-bearing applications in robotics, where both versatility and durability are essential.

Salps, marine invertebrates known for their collective swimming through coordinated jet propulsion, offer a unique model for efficient underwater movement. Inspired by this biological system, we develop the SALP (Salp-inspired Approach to Low-energy Propulsion) robot, a soft underwater robot that swims via jet propulsion similarly to a biological salp. The SALPs can be physically connected into SALP chains and coordinate their jets to achieve various propulsion modes. In our experiments, we compare the swimming performance of the individual SALP with the two-SALP system, focusing on power, acceleration, velocity, and energy efficiency. Results indicate that two SALPs swimming synchronously exhibit a 9.0% increase in steady-state velocity and a 16.6% improvement in transient acceleration compared to a single SALP. Additionally, our analysis of swimming efficiency implies that asynchronous swimming is potentially more energy efficient than the synchronous mode, as reflected by a decrease in the cost of transport (COT).

Soft robots have many advantages over rigid robots thanks to their compliant and passive nature. However, it is generally challenging to model the dynamics of soft robots due to their high spatial dimensionality, making it difficult to use model-based methods to accurately control soft robots. It often requires direct numerical simulation of partial differential equations to simulate soft robots. This not only requires an accurate numerical model, but also makes soft robot modeling slow and expensive. Deep learning algorithms have shown promises in data-driven modeling of soft robots. However, these algorithms usually require a large amount of data, which are difficult to obtain in either simulation or real-world experiments of soft robots. In this work, we propose KNODE-Cosserat, a framework that combines first-principle physics models and neural ordinary differential equations. We leverage the best from both worlds -- the generalization ability of physics-based models and the fast speed of deep learning methods. We validate our framework in both simulation and real-world experiments. In both cases, we show that the robot model significantly improves over the baseline models under different metrics.

Wind tunnel testing augments the undergraduate fluid dynamics curriculum by providing hands-on application of the course material, and a low-cost version of a force balance is desirable.To maximize the utility of wind tunnel-based lessons and laboratory demonstrations, there is also a need for a setup that is easily adaptable to different tests and loading applications.This paper provides such a force balance design, along with detailed evaluation and benchmarking to characterize the accuracy of the force balance.Our force balance uses readily available materials having a total cost under $125.Static load tests show that the force balance is accurate with a mean absolute percentage error of only 2.5%.We demonstrate the system's usefulness and adaptability with classic examples of measuring drag on a sphere and characterizing a NACA 0012 wing, as well as with measuring lift on a foldable wing.Finally, we pilot the force balance in an undergraduate mechanical engineering lab setting and find that students are able to explore the setup, understand the load cell functionality, and use the system to measure drag on a sphere.The force balance enables students to gain hands-on learning experience related to both fluid mechanics and statics, and our user study shows that the force balance is durable through classroom use.The low cost, robustness, and high adaptability of the system makes it suitable for incorporating in multiple labs or for allowing student project teams to utilize the system in their own experiments.

We introduce a self-contained pneumatic actuator capable of 1.43 times stiffness gain from 1332 N/m to 1913 N/m without needing an external air source or valve. The design incorporates an air chamber bellows and a spring bellows, connected and sealed. Stiffness modulation is achieved by altering the air chamber volume. We present an approach for computing the volume, pressurized force, and stiffness of a single bellows component, as well as methods for composing single bellows models to predict the change in stiffness of the dual bellows actuator as a function of air chamber compression. We detail the fabrication of the actuator and verify the models on the fabricated prototype. This actuator holds promise for future integration in tunable stiffness robots demanding high strength and adaptability in dynamic scenarios.

An origami-inspired bistable gripper, featuring a dual-function custom PET linear solenoid actuator that acts both as an actuator and a sensor, is presented. Movements in the permanent magnet plunger, which is directly mounted to the gripper, create induced electromotive force (emf) in the solenoid, and these induced emf measurements are used to detect snap-through actions and light contacts on the gripper. The fabrication methods for the gripper, actuator, and a gel-free soft wearable EMG electrode are outlined, and the actuator's self-sensing method utilizing the time-integral of the induced emf measurements are explored. Because a self-sensing actuator eliminates the need for extra sensors, it allows for further miniaturization of the robot while maintaining its compactness and lightweight design. The paper also introduces a full human-in-the-loop system, allowing users to open or close the gripper with their biceps via a wearable EMG electrode. This system bridges human intent with robotic action, offering a more intuitive interaction model for robotic control.

Actively tuning mechanical properties in soft robots is now feasible due to advancements in soft actuation technologies. In soft manipulators, these novel actuators can be distributed over the robot body to allow greater control over its large number of degrees of freedom and to stabilize local deformations against a range of disturbances. In this paper, we present a hierarchical policy for stiffness control for such a class of soft manipulators. The stiffness changes induce desired deformations in each segment, thereby influencing the manipulator's end-effector position. The algorithm can be run as an online controller to influence the manipulator's stable states – as we demonstrate in simulation – or offline as a design algorithm to optimize stiffness distributions – as we showcase in a hardware demonstration. Our proposed hierarchical control scheme is agnostic to the stiffness actuation method and can extend to other soft manipulators with nonuniform stiffness distributions.

Abstract Aerial vehicle missions require navigating trade-offs during design, such as the range, speed, maneuverability, and size. Multi-modal aerial vehicles enable this trade-off to be negotiated during flight. This paper presents a Bistable Aerial Transformer (BAT) robot, a novel morphing hybrid aerial vehicle that switches between quadrotor and fixed-wing modes via rapid acceleration and without any additional actuation beyond those required for normal flight. The design features a compliant bistable mechanism made of thermoplastic polyurethane (TPU) that bears a large mass at the center of the robot’s body. When accelerating, inertial forces transition the vehicle between its stable modes, and a four-bar linkage connected to the bistable mechanism folds the vehicle’s wings in and out. The paper includes the full robot design and a comparison of the fabricated system to the elastodynamic simulation. Successful transitions between the two modes in mid-flight, as well as sustained flight in each mode indicate that the vehicle experiences higher agility in the quadrotor mode and higher flight efficiency in the fixed-wing mode, at an energy equivalent cost of only 2 s of flight time per pair of transitions. The vehicle demonstrates how compliant and bistable mechanisms can be integrated into future aerial vehicles for controllable self-reconfiguration for tasks such as surveillance and sampling that require a combination of maneuverability and long-distance flight.

We present CurveQuad, a miniature curved origami quadruped that is able to self-fold and unfold, crawl, and steer, all using a single actuator. CurveQuad is designed for planar manufacturing, with parts that attach and stack sequentially on a flat body. The design uses 4 curved creases pulled by 2 pairs of tendons from opposite ends of a link on a 270°servo. It is 8 cm in the longest direction and weighs 10.9 g. Rotating the horn pulls the tendons inwards to induce folding. Continuing to rotate the horn shears the robot, enabling the robot to shuffle forward while turning in either direction. We experimentally validate the robot's ability to fold, steer, and unfold by changing the magnitude of horn rotation. We also demonstrate basic feedback control by steering towards a light source from a variety of starting positions and orientations, and swarm aggregation by having 4 robots simultaneously steer towards the light. The results demonstrate the potential of using curved crease origami in self-assembling and deployable robots with complex motions such as locomotion.

Abstract While there has been much work in the space of embodied intelligence, we as a field have struggled to define what exactly embodied intelligence is and how it should be used. In this paper, we propose that there are multiple types of embodied intelligence, and that these different types of embodied intelligence are suited to different types of tasks. We introduce a method for classifying tasks according to their objective and occurrence, and we describe how existing work in embodied intelligence fits into this framework. We hope that this proposed framework will initiate a discussion to more formally think about the role that embodied intelligence plays and the value that it brings to engineering and robotics.

Abstract The drag coefficient plays a vital role in the design and optimization of robots that move through fluids. From aircraft to underwater vehicles, their geometries are specially engineered so that the drag coefficients are as low as possible to achieve energy-efficient performances. Origami magic balls are 3-dimensional reconfigurable geometries composed of repeated simple water-bomb units. Their volumes can change as their geometries vary and we have used this concept in a recent underwater robot design. This paper characterizes the drag coefficient of an origami magic ball in a wind tunnel. Through dimensional analysis, the scenario where the robot swims underwater is equivalently transferred to the situation when it is in the wind tunnel. With experiments, we have collected and analyzed the drag force data. It is concluded that the drag coefficient of the magic ball increases from around 0.64 to 1.26 as it transforms from a slim ellipsoidal shape to an oblate spherical shape. Additionally, three different magic balls produce increases in the drag coefficient of between 57% and 86% on average compared to the smooth geometries of the same size and aspect ratio. The results will be useful in future designs of robots using waterbomb origami in fluidic environments.

Abstract Robot design is a challenging problem involving a balance between the robot’s mechanical design, kinematic structure, and actuation and sensing capabilities. Recent work in computational robot design has focused on mechanical design while assuming that the given actuators are sufficient for the task. At the same time, existing electronics design tools ignore the physical requirements of the actuators and sensors in the circuit. In this paper, we present the first system that closes the loop between the two, incorporating a robot’s mechanical requirements into its circuit design process. We show that the problem can be solved using an iterative search consisting of two parts. First, a dynamic simulator converts the mechanical design and the given task into concrete actuation and sensing requirements. Second, a circuit generator executes a branch-and-bound search to convert the design requirements into a feasible electronic design. The system iterates through both of these steps, a process that is sometimes required since the electronics components add mass that may affect the robot’s design requirements. We demonstrate this approach on two examples — a manipulator and a quadruped — showing in both cases that the system is able to generate a valid electronics design.

Abstract Robot design is a complex cognitive activity that requires the designer to iteratively navigate multiple engineering disciplines and the relations between them. In this paper, we explore how people approach robot design and how trends in design strategy vary with the level of expertise of the designer. Using our interactive Build-a-Bot software tool, we recruited 39 participants from the 2022 IEEE International Conference on Robotics and Automation. These participants varied in age from 19 to 56 years, and had between 0 and 17 years of robotics experience. We tracked the participants’ design decisions over the course of a 15 min. task of designing a ground robot to cross an uneven environment. Our results showed that participants engaged in iterative testing and modification of their designs, but unlike previous studies, there was no statistically significant effect of participant’s expertise on the frequency of iterations. We additionally found that, across levels of expertise, participants were vulnerable to design fixation, in which they latched onto an initial design concept and insufficiently adjusted the design, even when confronted with difficulties developing the concept into a satisfactory solution. The results raise interesting questions for how future engineers can avoid fixation and how design tools can assist in both efficient assessment and optimization of design workflow for complex design tasks.

We propose a novel design for a lightweight and compact tunable stiffness actuator capable of stiffness changes up to 20x. The design is based on the concept of a coiled spring, where changes in the number of layers in the spring change the bulk stiffness in a near linear fashion. We present an elastica nested rings model for the deformation of the proposed actuator and empirically verify that the designed stiffness-changing spring abides by this model. Using the resulting model, we design a physical prototype of the tunable-stiffness coiled-spring actuator and discuss the effect of design choices on the resulting achievable stiffness range and resolution. In the future, this actuator design could be useful in a wide variety of soft robotics applications, where fast, controllable, and local stiffness change is required over a large range of stiffnesses.