Search

We report on the methodology of developing a hybrid model and utilizing it as a template to initiate running behavior in a legged robot. The hybrid model is comprised of a physics-based, rolling, spring-loaded, inverted pendulum (R-SLIP) model, and a data-driven model that compensates for unmodeled dynamics using Gaussian process (GP) regression. The hybrid R-SLIP-GP model retains the R-SLIP's intrinsic running dynamics, as well as improves the template's accuracy without intensive data training and iteration efforts. The proposed hybrid model was evaluated via simulation and in empirical robot running experiments. The results confirm that the added GP-based model greatly improved the model's accuracy, especially discrepancies resulting from the originally difficult-to-model complex leg-ground interactions. In addition, the R-SLIP-GP model was utilized as a complete motion template to design and control the robot's running motion. The experimental results demonstrate that the hybrid template can initiate more stable and power-efficient running motion in the robot than using the R-SLIP template. [ABSTRACT FROM AUTHOR]

In this paper, we report on the development of the clock-torqued rolling spring loaded inverted pendulum (CTR-SLIP) model. The new model, which adds clock-based torque control on the leg orientation of the previously developed R-SLIP model, has two advantages: first, regulating the model to follow its passive dynamic running (i.e., at a fixed point) significantly increases the model's basin of attraction; and second, formulating the model closer to the empirical robot enables the model to serve as the transient and steady-state running template of the robot as the anchor. These features enable the model/robot to perform speed transition from one fixed point profile to another, and the experimental validation confirms that the robot can successfully transition between two running speeds bidirectionally. The achievement of variable-speed running by the proposed method has a unique merit—it is purely model-based, and there is no need for further tuning, optimization, or learning processes. Regarding the robot, the proposed strategy only requires it to have simple position control to regulate its leg orientations, and there is no need for other sensory modules to provide information for feedback. [ABSTRACT FROM AUTHOR]