Robot Control via Local Gaussian Process Regression (Machine Learning)

Course Project, Mentored by Dr. Jiayu Zhou, Dept. of CSE, Michigan State University

• In this project I’ll implement an algorithm approximating the robot dynamics using measured data to reduce computational complexity generated by non-linearity in the dynamics of a robot
• Implemented algorithm approximating the robot dynamics using measured data to reduce computational complexity generated by non-linearities in the dynamics of a 7 DOF robot
• Prediction and correction of non-linearities in robot model by developing an error model that predicts and accounts for errors during a certain movement of the robot
• I evaluated these procedures in terms of Model-based control, e.g. torque control, locally weighted projection regression (LWPR)

https://github.com/yashbagla321/CSE847