This course envisions a time when robots eventually contribute to society in a meaningful way by solving problems that humans alone cannot. For this transformation to occur, the public needs more than just access to controls literature and theory. It needs access to systems whose cost and performance can be scaled to solve massively parallel problems such as picking crops, and access to tools that make designing complex dynamical systems for providing walking assistance to the elderly as easy as picking specifications from a menu. And finally it needs new capabilities for prototyping and interacting with robotic systems from an early age, both in the classroom and at home. By getting robots into the hands of people through a focused shift towards affordability, customization, and accessible fabrication, we have the potential to organically transform the nature of work by the people who perform it.
The goal of this course and website, therefore, is to enunciate that path forward from where we are now. The slow speed and high cost of developing a new robot often leads engineers to design conservatively and to specify general purpose performance requirements that can satisfy a wide range of user needs across a variety of applications. Generality comes at a cost, however; the need for precision and payload implies more rigid designs, stiffer, heavier components, and bigger actuators. All of these design decisions results in robots ill-suited for what they ideally do best:
- working in dangerous environments where they might easily be damaged,
- performing low-value, repetitive tasks that no one is willing to do, and
- quickly scaling in swarms and teams to solve challenges in parallel.
In order to break the spiraling cost of developing “generalist” robots, this course approaches the design process in a different way: selecting affordable material systems, focusing on the design of customizable, “specialist” platforms, and selecting accessible fabrication strategies that encourage rapid prototyping and iteration. As current approaches in robotic modeling and control generally assume rigid systems with thoughtfully-inserted compliance, this approach also requires new methods for modeling and tuning lower-cost, compliant materials for use in high-performance dynamical systems.
This course thus introduces a holistic workflow for designing, manufacturing, understanding, and tuining dynamical robotic systems that are both affordable and easily prototyped. The goal is to make it possible for non-experts to design and prototype a robot in an hour.