NXT Robots and their Applications in Machine Learning Group 2: Roanne Manzano, Eric Tsai, Jacob Robison Mentor: Anjum Gupta Faculty Advisors: Professor Pankaj K. Das and Professor Charles Tu Objective Created algorithms for the robot to move and sense and process their surroundings. Sent and received information through Bluetooth communication between the NXT and a computer interface. Programmed the NXT robots to move through a path and map the robots movement on a computer interface in real time. Motivation Compared the capabilities of various programming environments on the latest Lego NXT Mindstorm robots by running complex programs of our own design. Create the first real environment simulator for NXT that collects precise measurements and accurately map the robots performance. Method/Approach Chose a programming language that allowed Bluetooth communication and complex coding on the robots. Created simple codes on the Lego NXTs in RobotC, NXC and Python to build an understanding of the hardware specifications and programming environments. Tested the robot sensors and Bluetooth capabilities. Implemented algorithms to determine robot’s capabilities to sense and process their environment (e.g. find the brightest spot in a room or opening in an enclosure). Used Python and Python plug-ins to map robots movement in real time. Conclusion Created algorithms to control the Lego robots movement based on the sensor readings. Developed the first virtual simulator for the NXTs in Python. Modeling out the robots movements in a simulator will greatly decrease development time for future machine learning application using the Lego NXTs and is helpful in collecting accurate measurements of the robots’ behaviors.