Stanford Robotics

Artificial Intelligence Laboratory - Stanford University

Neuromuscular Control Research

Movement is central to human life. In order to build robots that mimic human capabilities, it is important to first understand
how movement results from the intricate coordinated actions of nerves, muscles, tendons, bones, and other physiological components. Understanding movement requires understanding how the many components work, from the physics of dynamic musculoskeletal motions to the central nervous system’s control strategies for orchestrating motor tasks. While these can be partially observed with experiments, modeling and simulating the neuromusculoskeletal system’s dynamics elucidates many unobserved and ill-understood aspects.

This project aims to understand neuromuscular control of human movement by combining multiple approaches: 
We plan to achieve this by combining multiple approaches: by applying insights
from humanoid control theory to identify which experiments probe the biological motor system best;
by reconstructing recorded motions with detailed biomechanical models and predicting the
accompanying neuromuscular activity; and by using our model's predictions and experimental
recordings to decode the neural activation patterns that control motor task
  1. By applying insights from humanoid control theory to identify which experiments probe the biological motor system best.
  2. By reconstructing recorded motions with detailed biomechanical models and predicting the accompanying neuromuscular activity.
  3. And by using our model's predictions with experimental recordings to decode the neural activation patterns that control motor tasks.

Musculoskeletal Models:



 
 

Controlling a physiologically accurate humanoid biomechanical model.