Meet the AcroMonk: an underactuated brachiating robot that uses a single actuator and unactuated grippers to snap and release from monkey bars. As the researchers explain, the brachiation action can be achieved by:
using trajectory optimization via direct collocation and stabilization by a model-based time varying linear quadratic regulator (TVLQR) or model-free proportional derivative (PD) control, as well as by a reinforcement learning (RL) based control policy.
[HT] [Credit: Mahdi Javadi, Daniel Harnack, Paula Stocco, Shivesh Kumar, Shubham Vyas, Daniel Pizzutilo, and Frank Kirchner]