HSL Meeting: Brandon Koo; Cody Paige (12-1p, 5/8, Rm 31-270)

HSL Meeting

Monday, May 8, 2023
12-1pm, Room 31-270
Zoom link: https://mit.zoom.us/j/99313053821


Talk #1: 

  • Speaker: Brandon Koo
  • Title:  Neural control of motion prediction
  • Abstract—Modern exoskeleton systems, as well as other tandem human-robot interactive systems such as collaborative agents, lack real-world applicability due to low fluency (a measure of seamlessness between human and agent). One solution approach to the fluency issue is to predict motion, thereby potentially alleviating said incompatibility specifically due to robotic agents lagging behind in motion. By targeting the time lag between electrical stimulation and physical motion present in human muscle groups, as well as advances in machine learning, this study proposes the development of a motion prediction algorithm which can be used to drive exoskeleton or other tandem-use hardware with greater fluency.


Talk #2: 

  • Speaker: Cody Paige
  • Title: Development and preliminary results of a VR user study for geological exploration or Lunar and planetary surfaces
  • Abstract—As part of MIT’s work with the Resources for Exploration and Science of OUR Cosmic Environment (RESOURCE) project with NASA Ames and the Solar System Exploration Research Virtual Institute we are testing both the scientific and operational usefulness of a virtual reality environment for local, small-scale (< 5 m) geological analysis for Lunar and planetary surface exploration missions.  Specifically, we are testing a virtual reality (VR) environment for geological exploration. We incorporate local environmental data such as temperature, luminosity, humidity, and wind. The data was collected in Svalbard, Norway, from three locations near Longyearbyen. The sites were selected based on their distinct geological features including 1) a riverbed in a glacially carved valley (<10 cm-scale features), 2) the base of a recent glacial retreat (last 100 years, 10-50 cm) and 3) a permafrost feature (>1 m).  This data was rendered in VR and is being used to assess users’ abilities to answer questions about the relevant local geology for differing feature scales (<10 cm, 10-50 cm and >1 m). The VR environment is being compared to a traditional desktop application. Here we discuss the data collection techniques used in Svalbard as well as the lessons learned from this field work and preliminary results from the study.