No abstract

This paper addresses the challenge of terrainadaptive dynamic locomotion in humanoid robots, a problem traditionally tackled by optimization-based methods or reinforcement learning (RL). Optimization-based methods, such as model-predictive control, excel in finding optimal reaction forces and achieving agile locomotion, especially in quadruped, but struggle with the nonlinear hybrid dynamics of legged systems and the real-time computation of step location, timing, and reaction forces. Conversely, RL-based methods show promise in navigating dynamic and rough terrains but are limited by their extensive data requirements. We introduce a novel locomotion architecture that integrates a neural network policy, trained through RL in simplified environments, with a state-of-the-art motion controller combining model-predictive control (MPC) and whole-body impulse control (WBIC). The policy efficiently learns high-level locomotion strategies, such as gait selection and step positioning, without the need for full dynamics simulations. This control architecture enables humanoid robots to dynamically navigate discrete terrains, making strategic locomotion decisions (e.g., walking, jumping, and leaping) based on ground height maps. Our results demonstrate that this integrated control architecture achieves dynamic locomotion with significantly fewer training samples than conventional RLbased methods and can be transferred to different humanoid platforms without additional training. The control architecture has been extensively tested in dynamic simulations, accomplishing terrain height-based dynamic locomotion for three different robots.

No abstract

This paper presents our findings in exploring various approaches for turning on a novel prototype biped which takes inspiration from humanoids, but features fundamental differences that increase its stability while reducing its cost and complexity. This non-anthropomorphic robotic system modifies the traditional humanoid form by aligning the legs in the sagittal plane and adding a compliant element to the feet. As this approach to locomotion is relatively new, turning methods have yet to be explored. Turning on this unique platform is a nontrivial problem that we examined by adding additional DoF in the forms of arms or hip actuators. The turning strategies tested include using the hand or foot as a pivot point, utilizing the arms like a tail or reaction wheel, and adding another DoF to each leg. The methods were tested quantitatively to assess their rotational accuracy and qualitatively to evaluate their viability in certain situations.

No abstract