inverted pendulum

We all know by now that quadcopters can do impressive things, such as be made out of Lego, be controlled with your mind, or be impervious to crashing, but did you know that they can be quite the entertainer, performing juggling stunts in midair? Some of the more impressive feats quadcopters can perform are juggling balls, and balancing an inverted pendulum. Now, Dario Brescianini, a student at ETH Zurich’s Institute for Dynamic Systems and Control, and some of his colleagues have developed an algorithm that allows for the adorable little fliers to juggle an inverted pendulum.

An inverted pendulum, as the name suggests, is a pole reminiscent of a standard pendulum that you might find on a clock, but the center of mass is above the pivot point. This makes the inverted pendulum difficult to balance, and thus, even more difficult for a not-entirely-steady, flying robot to accurately throw toward another not-entirely-steady flying robot. In order to achieve that goal, Brescianini turned to a 2D mathematical model.

Quadcopters juggling

After the model detailed acceptable behavior for the quadcopters, the team tested it out in the real world, allowing two quadcopters to play catch with an inverted pendulum. Considering the quadcopters would catch and balance the pendulum on one end of the stick, rather than on the curved sides, the team needed to protect the quadcopter from potentially being impaled. So, the ends of the pendulum were wrapped in a balloon, which is filled with flour.

Though the 2D model paved the way for the quadcopter balancing act, it wasn’t perfect, as catching the pendulum proved difficult. To correct this issue, the team had to include a learning algorithm so the system could update its catching methods as it found out if they worked or not.

Brescianini and company have taught us an important life lesson today: If two individual tasks are difficult to perform on their own, combining them clearly makes performing them a lot easier.