July 14, 2020
(Be sure to watch Part 1 of this talk.)
In the second half of Using Computer Simulation in Mechanical Engineering, Dan Negrut, professor in the UW-Madison Mechanical Engineering Department, discusses the idea of using computer simulation to design better robots and autonomous vehicles. In order to simulate robots and autonomous vehicles, we rely on a physics-based simulation engine (called “Chrono” in this case)—a topic he discussed in Part 1. In designing robots, there are two matters of concern: Does the robot have the body to do something? and, does the robot have the brains to do what’s expected of it? Issues are covered such as physics simulation, sensor simulation, machine learning for autonomy, and virtual world simulation—the four pillars that support the idea of robotics simulation. Finally, Negrut discusses a test case in which machine learning was used to train an autonomous vehicle to follow a lead vehicle in off-road platooning operations.
To learn more about UW-Madison EPD’s Master of Engineering in Data Analytics (MEDA) online master’s degree program, click here.
If you have comments about this video or suggestions for future topics, please let us know here.
Note: If you are unable to see captions, select “CC” at the bottom of the video and choose “English.”
July 14, 2020