Optimized inverse kinematics modeling and joint angle prediction for six-degree-of-freedom anthropomorphic robots...

Area of Science:

• Robotics
• Artificial Intelligence
• Computational Kinematics

Background:

• Inverse kinematics is essential for robot control, determining joint configurations for desired end-effector poses.
• Six-degree-of-freedom (six-DoF) anthropomorphic robots present significant inverse kinematics challenges due to complex mathematics, nonlinearities, and multiple solutions.
• Existing methods often struggle with computational demands and generalizability.

Purpose of the Study:

• To systematically explore and identify optimal Artificial Intelligence (AI) models for solving the inverse kinematics problem in six-DoF anthropomorphic robots.
• To balance predictive accuracy with computational efficiency in AI-driven inverse kinematics solutions.
• To enhance model interpretability using Explainable AI (XAI) techniques.

Main Methods:

• Systematic exploration and rigorous evaluation of various AI models.
• Bayesian optimization for hyperparameter tuning to select the optimal regressor.
• Five-fold cross-validation on a public dataset for performance assessment.
• Explainable AI (XAI) using SHAP (SHapley Additive exPlanations) for feature importance analysis.

Main Results:

• The selected AI model achieved high accuracy in predicting six joint angles, with a mean squared error between 1.934 × 10^-3 and 3.522 × 10^-3.
• Demonstrated significant computational efficiency, with prediction times of approximately 1.25 ms per sample.
• SHAP analysis confirmed model interpretability and highlighted key feature importances.

Conclusions:

• AI models offer a viable and efficient solution for the complex inverse kinematics of six-DoF anthropomorphic robots.
• The developed approach successfully balances predictive accuracy and computational speed, crucial for real-world robotic automation.
• This research advances the state-of-the-art by providing interpretable and efficient kinematic solutions for robotic systems.