Gait Prediction Analysis Based on Particle Swarm Optimization Algorithm Support Vector Machine

Abstract

Gait is an emerging biometric trait, and gait prediction is a process that enables the prediction of the motion state of the human lower limb based on image features, joint angles, or human kinematic metrics. This article proposes a support vector machine (SVM) model based on particle swarm optimization (PSO) algorithm. Specifically, the particle swarm optimization (PSO) algorithm is used to optimize the penalty factor and kernel function parameters of the support vector machine (SVM). Meanwhile, by utilizing MediaPipe in conjunction with a motion camera within a neural network application, it is possible to generate lower limb keypoint features. This can be used to obtain datasets for training and testing models. The experimental prediction results show that the model has an accuracy of 96.29% for the lower limb gait prediction results, which has high prediction accuracy and can be used in the control scenarios of lower limb exoskeleton robots.