Compliant actuators and control methods have been known to exhibit similarities in human musculoskeletal systems. In order to better understand and improve the effects of force optimization under closed- loop conditions, a physical joint model was constructed with an agonist and an antagonist muscle operating under linear control. Utilizing LabVIEW software, compliant McKibben air muscles and Merlin stretch sensors, the author was able to incorporate a Bang-Bang controller and propose the inclusion of pulse width modulation (PWM) as well as a Proportional-Integral-Derivative (PD) controller to study the reflex loops under various levels of feedback sensitivity. The feedback mechanism, similar to proprioceptive muscle spindle feedback, will be based on the input given from the stretch sensors to achieve a desired movement.

Motions are controlled by the Central Nervous System (CNS). Demonstration of an open-loop muscle model working in conjunction with an inertial system would be extremely difficult. However, understanding the control process of a muscle can be demonstrated through the closed-loop control method. This was accomplished by constructing a linear bang-bang controlled computer program to simulate a musculoskeletal model that incorporates proprioceptive feedback. Simulink software was also utilized to show how the damping and stiffness coefficients, as in human muscle spindle and Golgi Tendon Organ feedback loops, can be adjusted to optimize stability.