Design, Modeling, and Experimental Validation of a Novel Flexible-Joint Robotic Manipulator using Adaptive Sliding Mode Control

This paper presents the design, fabrication, nonlinear dynamic formulation, simulation, and experimental comparison of a robotic manipulator incorporating a four-spring flexible coupling. The platform is described as a four-degree-of-freedom (4-DOF) arm, and the supplied model distinguishes a motor-side coordinate from three load-side coordinates. Conventional sliding-mode control (SMC) and adaptive sliding-mode control (ASMC) are implemented through a LabVIEW real-time environment and a National Instruments myRIO controller. Reported tests consider no added payload, a 1000 g payload, and an external disturbance at reference angles of 45∘ and 135∘. The supplied simulation and experimental summaries are compared using position-error percentages, settling time, and maximum load torque. Across six matched experimental cases, ASMC reduces the reported position-error indicator by 40.4–50.0%, maximum load torque by 11.8–35.4%, and settling time by 4.6–6.3% relative to SMC. These reductions are calculated from the supplied table values and vary with operating condition and reference angle. The author-reported responses indicate recovery following the applied disturbances and less oscillatory behavior with ASMC. However, the available summaries do not quantify complete chattering suppression, statistical variability, or total actuator energy consumption. The adaptive implementation also introduces a smooth switching function, so the respective contributions of adaptation and smoothing are not isolated. The model structure and Lyapunov candidates are presented with explicit limits: the elastic torque law, numerical implementation details, actuator allocation, and complete closed-loop stability derivation require further specification. The comparison therefore supports the reported performance improvements for the investigated cases while identifying the information needed for full reproduction and broader validation.