BIO

I am currently pursuing a degree in Mechanical Engineering at FSU with the hope to work in robotics. Robotics and it's versatility is fascinating to me and my interest has only been expanded by this study. It has even caused me to look into working on robotic prosthetics as a potential passion and goal for my career. I just hope to continue hands on work as research has been incredibly fulfilling to me.

The Potential Use of Kimera for Spatial Mapping to Predict Human Intent

Abstract

Many existing robotic lower-limb assistive devices available today are reactionary in the sense that they must first be prompted manually by the user to begin assisting with a particular action. Our research is focused on building a system that will allow the assistive device to predict the user's intent and begin helping with the action before a forced prompt, allowing for a more seamless and intuitive user experience. For example, we would like for the device to be able to recognize that a user wishes to sit down based on visual information revealing that they are approaching a chair and moving in a way consistent with sitting. To do this we first need to find the chairs using existing computer vision toolboxes such as the Simultaneous Localization and Mapping toolbox ‘Kimera’ developed and made open-source by MIT. We hypothesize that Kimera will help us towards our intent-recognition goal by spatially mapping the surrounding environment while also providing semantic labels for the objects in the scene. So far, the project has required an extensive knowledge of the Linux operating system, terminal programming, and extensive debugging. Despite the dearth of instructions for implementing Kimera, we are excited to have Kimera successfully installed and are ready to begin testing it’s capabilities this semester.

Keywords: Robotics, SLAM, Kimera, Human Intent, Mechanical Engineering