Abstract
This paper presents an innovative control design for a knee replacement exoskeleton, focusing on a dynamic synthesis of human-exoskeleton interactions. The design incorporates an optimally configured nonsingular terminal sliding mode controller (NTSMC), selected for its robust performance in uncertain and variable conditions. A key attribute of the NTSMC is its ability to achieve desired motion states swiftly and accurately, while maintaining nonsingularity, thereby boosting both safety and reliability. Central to our methodology is the application of the Non-dominated Sorting Genetic Algorithm (NSGA-II) in a multi-objective optimization framework. This approach effectively meets the intricate control objectives of the exoskeleton. The NSGA-II algorithm optimizes NTSMC’s setup parameters, balancing the minimization of settling time and integral absolute error. This balance is essential for ensuring the knee exoskeleton’s rapid, precise responses to dynamic movements, and for augmenting the smoothness of the NTSMC’s control signal, which in turn, improves user comfort and minimizes physical strain. Our findings confirm the control system’s adaptability to various patient parameters and its robustness against external factors like human torque. Additionally, the system proficiently monitors and follows squat movements and standard gait patterns.
