Mission

Scalable and Resilient Swarm Robotics for Extreme Environments

Welcome to the Nemitz Robotics Group at Tufts University. Our group specializes in swarm engineering, the design and implementation of scalable robotic systems. We push the boundaries of low-cost intelligence and fully automated robot manufacturing, requiring no human intervention. Our systems are resilient to structural deformations and electromagnetic interference that would ordinarily disable conventional control methods. Tailored for extreme environments, our swarm robotics technologies enable critical applications such as cave rescues, subsea operations, explosive ordnance disposal, environmental monitoring, and space exploration. We envision deploying large swarms of robots in humanitarian crises, with designs swiftly customized and mass-produced to meet mission-specific needs.

Machines that Evolve and Learn: We explore automated design strategies using genetic algorithms to access design spaces beyond human intuition. By creating robots with unique body-actuator configurations, we employ reinforcement learning to develop the most effective control schemes. This approach allows each custom robot to learn and optimize its behavior within its environment.

Microfactories: We are pioneering new strategies for fully automated fabrication and testing of soft robots to enhance manufacturing reproducibility. Our microfactories seek to combine advanced techniques like fused granulate fabrication, vision-based 3D printing, pick and place systems, automated print removal, and material recycling. This enables single-click production pipelines that produce high-quality robots.

Fluidic Computation and Control: We research fluidic alternatives to electronic intelligence to increase robot robustness against physical impacts and electromagnetic interference. By developing fluidic computers—including logic gates, memory elements, and state machines—we create control mechanisms that function reliably in extreme conditions where traditional electronics might fail.

Funding Sources and Collaborators