Stability Prediction Of Quadruped Robot Movement Using Classification Methods And Principal Component Analysis

Abstract

This paper introduces a novel technique for predicting the stability of quadruped robot locomo- tion using a central pattern generator (CPG). The proposed method utilizes classification methods and principal component analysis (PCA) to predict sta- bility. The objective of this study is to anticipate the stability or instability of robot movement by mod- ifying controlling parameters, referred to as features. The simulations of robot locomotion are conducted in MATLAB/SIMULINK R©, generating a dataset of 82 observations with different parameters. Machine learn- ing (ML) techniques are then applied, using classi- fication methods and PCA, to determine the stabil- ity condition. Six classification methods, including K-nearest neighbors (KNN), support vector classifier (SVC), Gaussian Naïve Bayes (GaussianNB), logistic regression (LR), decision tree (DT), and random for- est (RF) are implemented using Scikit-learn, an open- source ML library in Python. The performance of these classifiers is evaluated using four metrics: precision, recall, accuracy, and F1-score. The results indicate that KNN and SVC exhibit higher metric values com- pared to the other classifiers, making them more effec- tive for stability prediction.