Using Physics Simulators to Aid in Real-Time Robot Planning

Michael Tong

EECS Department
University of California, Berkeley
Technical Report No. UCB/EECS-2018-68
May 15, 2018

http://www2.eecs.berkeley.edu/Pubs/TechRpts/2018/EECS-2018-68.pdf

In this report we focus on building a framework for a feedback loop for robot planning that incorporates physics simulations in modeling complex interactions between the robots and their corresponding environment. By using a physics simulator to handle the dynamics, one can consider the tasks of building an environment and the dynamics of the robot separately, making it much easier to design and implement tests. Additionally, using a physics simulation allows for the testing of hypothetical environments and scenarios that would be prohibitively expensive to physically build and test. This report will show how to conceptually build such a framework, describe in detail the realization of such a framework, and discuss the results of experiments performed using the framework.

Advisor: Ronald S. Fearing


BibTeX citation:

@mastersthesis{Tong:EECS-2018-68,
    Author = {Tong, Michael},
    Title = {Using Physics Simulators to Aid in Real-Time Robot Planning},
    School = {EECS Department, University of California, Berkeley},
    Year = {2018},
    Month = {May},
    URL = {http://www2.eecs.berkeley.edu/Pubs/TechRpts/2018/EECS-2018-68.html},
    Number = {UCB/EECS-2018-68},
    Abstract = {
In this report we focus on building a framework for a feedback loop for robot planning that incorporates physics simulations in modeling complex interactions between the robots and their corresponding environment. By using a physics simulator to handle the dynamics, one can consider the tasks of building an environment and the dynamics of the robot separately, making it much easier to design and implement tests. Additionally, using a physics simulation allows for the testing of hypothetical environments and scenarios that would be prohibitively expensive to physically build and test. This report will show how to conceptually build such a framework, describe in detail the realization of such a framework, and discuss the results of experiments performed using the framework.}
}

EndNote citation:

%0 Thesis
%A Tong, Michael
%T Using Physics Simulators to Aid in Real-Time Robot Planning
%I EECS Department, University of California, Berkeley
%D 2018
%8 May 15
%@ UCB/EECS-2018-68
%U http://www2.eecs.berkeley.edu/Pubs/TechRpts/2018/EECS-2018-68.html
%F Tong:EECS-2018-68