Intelligence is often best expressed through creative problem solving – a skill that humans frequently depend on, especially in high-stress or time constrained scenarios. In the Apollo 13 incident of 1970, a carbon dioxide filter creatively constructed out of a sock, a plastic bag, book covers, and duct tape helped save the lives of the three astronauts on board. Solving problems by constructing new tools from available objects is colloquially referred to as “Macgyvering”, a term that originated from the popular TV series Macgyver. Unlike humans, today’s robots are limited to using available predefined tools. The construction of tools to solve problems is a particularly interesting challenge for intelligent robotic systems because it demonstrates a more sophisticated level of intelligence than simply tool use. It increases the resourcefulness of robots, allowing them to adapt to a wide range of unpredictable scenarios.
This research focuses on tool construction, contributing a computational framework that enables a robot to construct, or Macgyver, tools out of parts available in the environment. To the best of our knowledge, this is the first work to demonstrate tool construction on a physical robot.