RattleSnake : stochastic folding for chain programmable matter

Abstract

The vision of programmable matter is to create a blob of material that can transform itself into an arbitrary form. One promising approach for achieving programmable matter is to construct a chain of identical nodes that can fold into arbitrary threedimensional shapes. Previous active electromechanical systems have demonstrated this concept but are currently costly, complex, and not robust enough to scale to smaller sizes or larger numbers of nodes. The goal of this thesis is to explore methods of simplifying chain programmable matter by removing the actuator from each node and, instead, putting energy into the system externally through stochastic vibrations. Each node takes this random energy input and rectifies it to produce motion towards the target position. We propose two variants of this system: 1) smart clutches that can be reprogrammed in situ and fold through arbitrary paths in configuration space and 2) ratchets that are programmed ahead of time and are entirely passive. We developed a chain using the ratchet concept and also constructed a new active, electromechanical chain with reduced cost and improved speed and torque compared to previous electromechanical systems. Through experimental and computer simulated studies, we determined that stochastic actuation can simplify and reduce the cost of these systems. We have also identified how the size of the increments of the ratchet, length of the chain, and the amplitude and frequency of agitation affect the folding time and success rate. In addition, we show that passive folding systems should improve in performance as the hardware scales down.