Hi All! Looking for some resources and general feedback for an educational research project.
I am conducting an 8 week research project with the University of NSW in Sydney Australia (UNSW).
This Project revolves around the usefulness of UE5 for Prototyping features/logic of a robotic system that navigates and attaches to desks to move from A to B in a workplace environment with success criteria for both the robot and its attached desk. The point is to add as many real world variables as possible for testing larger systems later down the line (multiple NPCs with overlapping paths and goals, moving /chaotic obstacles like people, ordered or hierarchical operations) .
Starting with a digital twin of an office space (already in UE), I am looking to create an NPC system for a small robot that can do the following:
-
Navigate with perception (is this possible?). As in, using an in game camera, could the NPC navigate with some kind of breadcrumb to follow and its own vision? Or will I need to use the navmesh system. This feature would be used to explore possible real world perception systems like cameras/lidar.
-
Have weight effect movement and rotation speed, and eventually the weight of an attached object effect this movespeed.
-
Attach to object and augment bounding box and movement system, augment success conditions in BT. MORE CONCISE GOALS ATTACHED AS IMAGES
The point of this research is to replicate real world constraints to create insights regarding potential real world features of the robotic system and pseudocode for a real system. The goal is not to create a 1:1 clone of the real world system but as a rapid prototyping tool to test potential avenues of research for the eventual real world robot.
I would be extremely grateful if anyone can point me in the way of some useful tutorials and in-engine terms to research in the documentation or rip apart my current goals as to why they can cant be done/dont make sense.
I understand that early developments can massively shift potential features down the line and would like to reduce doubling development by starting with more future proofed systems that consider goals later in the project.
I am also looking for a remote mentor both paid or unpaid to assist in the project if anyone is interested PM me.