Abstract
In the past, a robotic arm needed to be taught to carry out certain tasks, such as selecting a single object type from a fixed location and orientation. Neural networks have autonomous abilities that are being deployed to aid the development of robots and also improve their navigation accuracy. Maximizing the potentials of neural network as shown in this study enhances the positioning and movement targets of industrial robots. The study adopted an architecture called XBNet (Extremely Boosted Neural Network) trained using a unique optimization approach (Boosted Gradient Descent for Tabular Data (BGDTD) that improves both its interpretability and performance. Based on the analysis of the simulations, the result demonstrates accuracy and precision. The study would contribute significantly to the advancement of robotics and its efficiency.