Massimo Regoli acquired a Bachelor’s degree in Computer Engineering (cum laude) and a Master’s degree in Artificial Intelligence (cum laude), both from Sapienza University of Rome. During his master thesis he worked for about a year with the Nao robots, implementing smart soccer-related behaviours based on Bayesian classifiers for the RoboCup competition.
After a brief period of work in a software house, he joined the Istituto Italiano di Tecnologia, where he is a third year Ph.D. student in Bioengineering and Robotics. He is studying how to exploit the tactile sensors of the iCub robot in order to endow it with skills such as grasping and manipulation. During his Ph.D. he gained practical and theoretical knowledge on machine learning and control theory. As part of his project he has implemented a controller that regulates the pressure applied by the fingertips on an object to maintain a stable grip. He used machine learning techniques (mainly reinforcement learning, neural networks and Gaussian mixture regression in the context of learning by demonstration) to improve this controller so that the robot can manipulate the object to achieve a better grasp.
He also investigated the problem of in-hand object recognition using tactile sensors, in which he used a machine learning approach (using algorithms based on Regularized Least Squares) that allowed the iCub robot to discriminate objects through tactile interaction.
Regoli M., Pattacini U., Metta G., and Natale L., Hierarchical Grasp Controller Using Tactile Feedback, in IEEE-RAS International Conference on Humanoid Robots, Cancun, Mexico, 2016. Selected as best paper finalist (interactive session). [pdf]
Regoli M., Jamali N., Metta G., and Natale L., Controlled Tactile Exploration and Haptic Object Recognition, in Proc. IEEE International Conference on Advanced Robotics, Hong Kong, China, 2017. Selected as best paper finalist. [pdf]