Motor Design and Control for Scalable Distributed Actuation

Abstract

Machine design is often constrained to a limited number of controllable degrees of freedom due to the cost and complexity associated with integrating large numbers of actuators. This thesis explores the hardware and control development for a motor architecture designed for distributed actuation, where many controllable degrees of freedom are required across macro scale structures. Availability of a low cost and easily integrated actuator at these scales would open new regimes for fields such as robotics, manufacturing, human computer interaction and wireless communication.

A survey of prior distributed actuation research is conducted, including shape memory alloy, piezoelectric, hydraulic, and electric motor topologies. A new approach using a multiplexed two-phase axial flux PCB motor is designed and iterated on through empirical testing and simulation. These motors are integrated into a modular 64 actuator array, and a proof of concept is built capable of distributed linear motion for interpolation of a surface or as independent degrees of freedom. The prototype achieves 21𝜇m linear resolution over 45mm of stroke, with a 1.9N stall force, and a density of 104 actuators per square foot. Motor commutation is achieved through multiplexing of individual motor windings, allowing for sub-linear cost and component count scaling. Actuator performance over a number of performance parameters is addressed, including output torque, speed, mass, resolution, range of motion, as well as parameters critical to scalability including motor footprint, cost and power consumption. Finally two applications of serially distributed actuation are discussed, including the design of modular continuum robots from a discrete toolkit of structural elements, as well as a serpentine actuator with many controlled degrees of freedom.