Towards a Hierarchical Neuromuscular Control Model with Reflex-based Spinal Control - Study with a Simple Running Model

Humans can generate diverse locomotion behaviors through the hierarchical control of the supraspinal and the spinal systems, where the spinal control is often assumed to consist of central pattern generators and reflexes. It has been demonstrated through simulation that human walking and running can be generated with control models where the central pattern generator is controlled in a hierarchical structure. However, no reflex control model in such framework has been proposed. Here we propose that locomotion behaviors can be controlled by a hierarchical control structure where the supraspinal system modulates reflexes with simple signals. We demonstrate in physics simulation that a simple neuromuscular model with a hierarchical reflex control can regulate running speed and height on unknown terrains with height differences of ±10cm.