Location: Upson 106 Conference Room Next to the Lounge
Abstract: Modular self-reconfigurable robots are typically composed of homogeneous units executing a set of programmed interactions with their neighbors based on a deterministic rule set. Some stochastic modular robotic systems take advantage of their physical design to arrive at a desired state and offer great potential in scalability. We propose combining an innovative hardware design with a control algorithm that enables intermodule interactions with some inherent randomness to ensure successful attachment through permanent magnets. We present the FOAMbots, a scalable, planar, modular robot composed of inflatable units, capable of onboard processing, actuation, sensing, and communication. Each module contains a poro-elastic foam that enables structural integrity while allowing fluid to flow through its volume. Pairs of permanent magnets along the modules’ perimeters enable attachment with adjacent modules, and low-cost, stretchable strain sensors allow modules to communicate and sense their surroundings. This presentation will introduce the hardware, various characterizations of the modular robot’s mechanical and locomotion properties, and a discussion of the algorithms that are currently being implemented to achieve local-to-global changes in the collective’s mechanical properties.