Recomendación de Actividades de la Vida Diaria para un Exoesqueleto Bimanual mediante un VLM basado en Edge Computing
TL;DR
This paper presents an ADL recommendation system based on a Vision Language Model (VLM), which runs locally, describes the objects in the scene, and recommends different ADLs based on the objects in the scene and the capabilities of the robotic system.
Abstract
Assistive robotics has established itself as a vital solution for supporting users with motor impairments in performing Activities of Daily Living (ADLs). In this context, the use of bimanual upper-limb exoskeletons presents a promising option for helping people with motor impairments perform ADLs. However, using these devices can be complex for the user, so intuitive interaction methods must be designed to facilitate easy use. This paper presents an ADL recommendation system based on a Vision Language Model (VLM). The model runs locally, describes the objects in the scene, and recommends different ADLs based on the objects in the scene and the capabilities of the robotic system.
