Abstract
This paper presents a lightweight visual servoing framework based on GaussianProcess (GP) regression for real-time autonomous navigation of electric wheelchairs in corridorenvironments. The proposed system leverages Direct Visual Servoing (DVS), where globalimage features are mapped directly to velocity control signals using a GP model. Histogramof Oriented Gradients (HOG) features are extracted from camera images and used as inputs tothe GP, which predicts angular velocity commands without requiring explicit geometric featureextraction or Jacobian computation. The GP model is trained using a dataset collected in variedindoor corridors, with control labels derived from a geometric-based control law. The systemwas deployed on a Raspberry Pi embedded platform and tested on a commercial wheelchair.Experimental results demonstrate that the proposed controller operates at 7 Hz, ensures robusttrajectory correction under visual noise, and achieves over 90% accuracy in corridor-followingscenarios, even under challenging lighting and occlusion conditions. This approach combinesthe adaptability of learning-based control with the simplicity and interpretability of classicalvisual servoing, offering a scalable and computationally efficient solution for assistive mobilityapplications.
DOI: 10.61416/ceai.v28i1.9599