Dynamic, biped walking is a difficult motor coordination problem for robots. "RunBot," shown here, contains two
joint motors and two hip motors and an upper-body component. These
actuators are dynamically controlled by a neural network that can adapt
via simulated synaptic plasticity to novel situations, such as a change
in the terrain. This way, RunBot's walking pattern looks rather
human-like (see Video S2 5.1 MPG), and the little machine successfully tackles difficult walking situations (see Manoonpong et al., e134).
No comments:
Post a Comment
Note: Only a member of this blog may post a comment.