Skip to content

Sprint Goal

Gokul Krishna Chenchani edited this page Sep 20, 2023 · 13 revisions

Research and Development Questions

Research Questions

  • Which encoding deep learning approach can be used for comparative evaluation?
  • Which fusion methods can be implemented for comparative evaluation?
  • Why is this approach necessary in the current @Work scenario?
  • Comparative run-time analysis with current YOLOv8 model used in the YouBot for performance and test on arbitrary surfaces.

Development Questions

  • Deploy the final study and implementation on the real YouBot hardware and test the performance and efficiency compared with YOLOv8 model.

SPRINT GOALS

15th OCTOBER

  • Complete RnD proposal submission and register for RnD - Draft for Deebul by 30th September.
  • Collect all the dataset related to RoboCup@Work and 6d dataset of HOPE (Simulated).
  • Create/Train a model (E.g. ResNet50, ResNet18) on the RoboCup@Work and 6d HOP datasets.
  • Article/Blog on Error correcting encoding and other encoding methods (Jupyter).
  • What fusion methods to be used to get a single output for the used datasets - Article/Blog on the findings.
Clone this wiki locally