Comparative Analysis of Levenberg-Marquardt and Neural Network Algorithm for 6-DOF Robot Manipulators in Precision Agriculture

Robotic arms are increasingly being utilized in agriculture, where agility and precise movement are essential for their effective implementation. To enhance the performance of these manipulators, many researchers are turning to Artificial Intelligence (AI) to address challenges in accurately predicting the joint angles needed for inverse kinematics. Traditionally, this has been done using computational methods like the Levenberg-Marquardt (LM) Algorithm, but there is a growing interest in using AI techniques, such as neural networks, to improve efficiency. Despite this, few studies have directly compared these methods in real-world applications. This paper presents a comparative analysis of two approaches: the Bi-LSTM (Bi-Directional Long Short-Term Memory), a type of Recurrent Neural Network (RNN), and the LM algorithm, focusing on the time required to compute inverse kinematics. The study was conducted on a 6-DOF robotic manipulator with five joint angles. The results show that Bi-LSTM outperforms the LM algorithm, achieving an average absolute joint angle error of just 0.3065 degrees, which is impressively low. Moreover, Bi-LSTM proved to be significantly faster, completing the calculations 12.75 times quicker than the LM algorithm. However, it is worth noting that the AI-based approach consumed 1.2 times more power compared to the traditional method. This study highlights the potential of AI to revolutionize robotic arm manipulation in agriculture, offering both speed and precision.