Statistical Method For Robotics Design Planning

Voiceover

Presenter Information(s)

Gage Koehler

Abstract or Description

This experiment will test various methods of assorting data concerning First Robotics Competitions to create better ways for teams to create goals concerning weight, size, and drivebase type that are beneficial for specific strategies and playstyles. Different forms of data collection and statistical analysis  will be tested to compare the results and determine the methods that contribute to the most effective design methods.

Historically, robotics teams have primarily relied on empirical methods and heuristics to determine robot specifications. Although these approaches have yielded successful results in the past, there still exists a gap in potential over long periods of time, which a consistent and versatile method of generating specifications would solve. There is an absence in large-scale data-driven methods for defining optimal robot parameters. 

This research holds immense potential to benefit robotics teams around the world, especially newer teams that especially struggle to define specific parameters. By creating an algorithmic approach to this, teams will have access to evidence-based insights that exceed the abilities of insight and experience. The specifications derived will be tailored to maximize the robots potential, to increase competitive success.