SCHEDULE

Soft-bodied robots usually share the same body for sensing, actuation, and interaction. We have clarified that variations in morphology of the soft body would lead to change in the function of sensing. However, control of soft robot still faces a challenging issue when the soft body is partially damaged, since the embedded sensing elements may generate unexpected feedback. Our aim is to tackle this scientific challenge with a three-fold research purpose: (1) Clarify the correlation between output of embedded sensors and critical change (such as being cut, worn, or trimmed) of the soft body based on dynamic simulation. (2) Propose a compensation strategy based on the study in (1), called morphological compensation, for minimization of any difference in sensing feedback. (3) Implement such compensation on actual soft robotic mechanisms, considering its benefit in sensing, and interaction. Obtained results are expected to contribute to the science of sensorized soft bodies, and control of soft robots. 

TBA

TBA

TBA

TBA

TBA

In this talk, the speaker will present a novel approach to pneumatic soft sensing that allows for online and offline tunability, a step towards active sensing. Pneumatic soft sensors are a type of soft sensor that use air pressure as a sensing modality. They can be integrated into various soft robotic actuators and provide feedback on deformation, contact force, and curvature. Apart from the ease of integration to soft robots that are also driven by pneumatic actuation, air shows great versatility in changing the sensor morphology and stiffness by being embedded within a soft elastomer cover. This talk will demonstrate how to shape the “Air” for active soft sensing by manipulating the geometry, material properties, and inflation pressure of pneumatic soft sensors. The controllability of the "Air" created a tunable mechanical filter to favor the sensing signal that adapts to the environment. The talk will highlight the modeling, fabrication, and applications of this technique for in-hand manipulation and large-area sensing. 

Walking and crawling animals can quickly form their gaits within minutes of being born, thanks to their neural locomotion control circuits, which are genetically encoded. They can adapt their movement quickly to traverse various substrates and even take proactive steps to avoid obstacles. Biological studies reveal that complex locomotion behavior is primarily achieved through several components, including neural control networks (central pattern generators, premotor networks) with plasticity and memory, sensory feedback, and dynamic body-brain-environment interactions. In addition to neural computation, morphological computation (biomechanics) also plays a critical role in robust behavior. In this talk, I will discuss how these components can be translated into neural control and sensing technologies to achieve bio-inspired locomotion and adaptation in rigid and soft robots.