Anshuman Radhakrishnan, Westview High School, Nikhil Murthy, and Jennie Wang
Advances in the field of Artificial Intelligence have made it feasible to rapidly develop and implement Machine Learning models. This growth has an extensive impact on the robotics world by improving decision making and software quality, ultimately allowing for more intelligent robots than ever before. The organization, For Inspiration and Recognition of Science and Technology (FIRST) runs worldwide robotics competitions for K-12 students – the subprogram FIRST Tech Challenge (FTC) allows students to build and program Android Robots.
Challenges involved in designing these robots are numerous, including unreliability of robot hardware, changing environments due to outside influences, and general imprecision of sensors and control systems. Intelligent software and rigorous testing is necessary in order to achieve optimal performance in the FTC game. We determined that it was necessary to develop and test decision-making software in order to improve the quality of our programs.
This paper presents an attempt to test machine learning models and their applications in the context of the FIRST Tech Challenge and design decision-making networks that enable Android robots to intelligently adapt to their environment, allowing for a streamlined testing process and improved software quality. This paper’s contribution to the software quality field is a process of testing decision-making software in robots performing in uncertain environments.