Learning to Control Redundant Musculoskeletal Systems with Neural Networks and SQP: Exploiting Muscle Properties

Keyphrases

• Neural Network 100%
• Musculoskeletal System 100%
• Learning Control 100%
• Muscle Properties 100%
• Muscle Stimulation 100%
• Control Policy 66%
• Controller 66%
• Stiffness 33%
• Machine Learning Approach 33%
• Forward Model 33%
• Musculoskeletal Modeling 33%
• Biomimetic 33%
• Movement Dynamics 33%
• Target Position 33%
• Learning Task 33%
• Optimality Criteria 33%
• High Linearity 33%
• Highly Nonlinear 33%
• Human Arm 33%
• Co-contraction 33%
• Sampling Efficiency 33%
• Sequential Quadratic Programming 33%
• High Redundancy 33%
• Biomechanical System 33%
• Musculoskeletal Apparatus 33%
• Muscle-driven 33%
• Actuator Redundancy 33%

Engineering

• Policy Control 100%
• Optimality 50%
• Robot 50%
• Nonlinearity 50%
• Target Position 50%
• Forward Model 50%
• Learning Task 50%
• Learning Approach 50%
• Classical Technique 50%
• Biomimetic Muscle 50%

Computer Science

• Neural Network 100%
• Robot 50%
• Biomimetics 50%
• Machine Learning Approach 50%
• Target Position 50%
• Sequential Quadratic Programming 50%