Learning Motor Control of Redundant Systems

RESEARCH THESIS

Amir Karniel

7.3. Final Remark

The main outputs of the brain are motor commands to the muscles.  The human brain is first of all a motor controller.  Biological motor control is a great challenge for scientists, engineers and physicians.  The classical engineering and mathematical modeling tools are appropriate for linear time-invariant injective systems.  The biological system does not comply with these qualifiers and therefore there is a place and a need for new modeling tools in order to describe and analyze the biological system.  In this thesis, some basic problems of motor control were illustrated.  The possible role of the mechanical nonlinear properties of the muscles in simplifying the control strategy was demonstrated.  A new architecture, learning algorithms and mathematical tools were suggested in order to exploit the virtue of redundancy.  These results constitute another step in the ongoing strive for a better understanding of the biological motor control.  An understanding of the biological motor control system will undoubtedly contribute significantly to the welfare of paralyzed and crippled patients, to a new generation of dexterous robots, and to a better understanding of the mysteries of the human mind and how it operates.