|
CMPEN 497: Special Topics - Humanoid Robot Simulation
This course investigates state of the art, programmable humanoid robot kinematics and dynamics using a combination of principles of physics, computer vision, and AI. Students learn to model and utilize 17-degrees of freedom humanoid robots with enhanced stability, sensory feedback, and actions.
Robot used: UBTECH Yanshee
Software/languages used:
-
Blender
-
Webots
-
Python
-
OpenCV
The course includes four projects:
-
Humanoid Robot Modeling in Blender:
Students first create a simple physics simulation in Blender to better familiarize themselves with the software.
Once familiar with Blender, students then construct a model of UBTECH's Yanshee robot and create a variety of stable and unstable poses.
-
Humanoid Robot Stability in Webots:
Students use Webots to analyze the stability of a UBTECH Yanshee robot as it moves through three different dances.
This inolves generating data within Webots and processing it later on with Python.
-
Vision and Webots:
Students develop a program to control a simulated UBTECH Yanshee robot in Webots using motion detected by their webcams.
Serving as an introduction to computer vision, students must determine where motion is detected,
send the appropriate command from their vision program to Webots using socket programming, and execute the appropriate command in Webots.
-
Original Dance:
Students design original dances in Webots to keep the robot stable until the very end of the dance, at which point the robot should fall over.
The dances are later tested for stability on a real Yanshee robot.
|
