In vitro gait simulators recreate in vivo kinematics and kinetics in foot and ankle specimens. The choice of underlying control system inherently influences the realism of resultant simulations. A physiologically accurate phantom model was developed, consisting of six segments and nine tendons. Two control systems were tested: a three-stage cascade controller and a force controller, which took the muscle forces calculated in a cascade learning trial as inputs. The controllers were tested on two target motions, a range of motion trial where the tibiotalar, subtalar and metatarsophalangeal joint of the hallux were controlled to perform a sinusoidal motion and a physiological motion of the stance phase of gait.

The cascade controller was more accurate and repeatable than the force controller, with average and maximum errors for the cascade controller of 0.27° ± 0.01° and 1.12° ± 0.04°, compared to 0.66° ± 0.11° and 2.57° ± 0.95° for the force controller. However, the force controller was approximately two thirds faster, with a feedback loop iteration time of 19.0 ms ± 4.1 ms, compared to 57.4 ms ± 12.6 ms. Both simulators showed a higher linear correlation than previous simulators, with all trials achieving r > 0.99. Therefore, both control systems showed high accuracy and repeatability, but varied in speed and should be chosen on a case-by-case basis, depending on the targets of the study.

Clinical Relevance— Creating an in vitro gait simulator which accurately and repeatably recreates in vivo motions is important in building a tool that will help quantify kinetic and kinematic effects of pathological and injurious conditions. In addition, by replicating the muscle forces using a physiological optimization, the control systems will enable the design and validation of improved surgical and orthotic interventions.