Dynamic Legged Locomotion in Realistic Terrains

Algorithmic and Physical Performance Advancements and Challenges



Legged robots have the potential capability of versatile mobility in human oriented environments. This can enable in the near future their deployment in realistic domestic and industrial workplaces as well as in complex and unstructured environments like those resulted after disasters. The critical mobility challenge is therefore to traverse vast unstructured terrains and balance under unknown perturbations during operations. In recent years, there are several achievements in the locomotion and balancing of legged machines when traversing irregular terrains, demonstrating capabilities that allow them to cope with moderate terrain roughness.  Their performance is however still very far from the rich locomotion and balancing skills demonstrated by humans or animals. Concerning the latest, there are also few new findings in the biomechanics of human/animal locomotion, which in combination with the progress in legged robots hardware performance, can permit to execute highly dynamic gait modulations and balancing reactions as well to generate these responses through the computation of whole body algorithms that take into account multimodal feedback including proprioceptive and visual cues. The latest is allowed by the simultaneous increase of the on-board computational power as well as the algorithmic advancements and optimization.   

The proposed workshop will discuss the most recent achievements in legged locomotion and balancing from the recent findings of the fundamental principles of human locomotion skills in biomechanics, to the most novel locomotion and balancing algorithms applied to the recently developed high performance legged robot platforms.