Visual motion control for autonomous wheelchair sidewalk following

The independent mobility of wheelchair users is pivotal for their overall well-being. Negotiating electric wheelchairs through constrained spaces, especially for those with disabilities, presents notable challenges. In response, this research introduces a pioneering vision-based approach for sidewalk navigation, leveraging tactile paving features and advanced Machine Learning (ML) models. Our methodology encompasses the creation of a specialized dataset, the formulation of a custom control law, and the deployment of a lightweight real-time Gaussian Process (GP) model on a Raspberry Pi platform. Thorough experimentation substantiates the efficacy of our approach, demonstrating accurate and dependable autonomous wheelchair navigation on sidewalks. Beyond technical contributions, our solution offers a cost-effective alternative to conventional sensor-dependent systems, profoundly enhancing user mobility and overall quality of life. By empowering wheelchair users with enhanced navigation capabilities, this research strives to foster independence and inclusion in their daily lives.