Answers are grounded in this paper and cite their sources.
High-dynamic motion control of large-sized quadruped robots is highly challenging due to the model inaccuracies introduced by substantial size and weight. Hurdling is particularly difficult, as it requires not only well-designed gait and trajectory planning but also a careful balance between model accuracy and real-time performance during control. We first measured and analyzed canine hurdling gait characteristics and full-body motion trajectories using motion capture systems. Inspired by these observations, we designed biomimetic hurdling gaits for quadruped robots, formulated parameterized hurdling trajectory equations, and developed a three-phase hurdling strategy consisting of approaching, jumping, and tracking. To resolve the contradiction between model fidelity and real-time control, we proposed a three-level dynamics model together with a hierarchical model predictive control framework. Finally, we implemented the running, jumping, and hurdling motions on a 55 kg large-scale quadruped robot in a dynamic simulation environment, achieving a hurdling speed of 2.0 m/s and a hurdle height of 0.45 m. In addition, vertical and forward jumping experiments were conducted on the physical prototype, achieving a maximum jumping height of 0.78 m and a jumping distance of 1.1 m, thereby verifying the feasibility of the proposed algorithm under real-world conditions and hardware constraints.
No takes yet. Share an insight, caveat, or question.
Huang et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: