Optimizing ABB MIG welding robot through polynomial trajectory planning and artificial intelligence integration

Abstract

The optimization of the ABB MIG welding robot involves intricate trajectory planning, typically handled through mathematical approaches. However, conventional methods often face challenges due to complex mathematical equations and derivations. This study presents ANN integration as a way to circumvent these restrictions and enable the customization of trajectory planning features. The effectiveness of this method is illustrated by the utilization of a virtual six-degree-of-freedom (DOF) robot manipulator. Analyzing and selecting hyper-parameters to optimize the ANN’s performance is a part of the research. Subsequently, sample data are employed to assess the robustness of the developed ANN topology. This assessment is conducted by comparing the outcomes of the artificial neural network approach with the results obtained using conventional mathematical methods. When compared to current state-of-the-art methods, the suggested method achieves better accuracy and efficiency in analyzing the dependability of trajectories with both interval and random uncertainty. All things considered, this paper’s findings greatly aid in the evaluation and planning of manipulators dependability and safety.