Legged locomotion systems often suffer from the curse of dimensionality. Unlike traditional wheeled robots, where the actuator coordinates intuitively correspond to task space attributes such as speed and direction, legged platforms usually have large number of actuators and joints, which are much farther away from the specification of the locomotion task.

Part of my research investigates the use of simple dynamical systems, such as a pogostick, as an abstract control interface for complex locomotors such as hexapods. The challenge is first to understand and control the dynamical properties of these abstractions, or templates, and second to find controllers that will anchor these abstractions in the original complex system. The result is a hierarchical decomposition of the control problem, resulting in a very simpl and intuitive high level control interface for a complex robot.

As part of my dissertation, I have presented an example of this approach, using the Spring-Loaded Inverted Pendulum(SLIP) model to control a simulated hexapod robot.