20221124 Us Patent Reinforcement Learning

Publication of US Patent on Reinforcement Learning for Control Systems – My US patent, detailing a novel method for training control arrangements in systems such as robotics, was officially published. I developed an architecture where an actuator, trained via reinforcement learning, corrects the feedback information supplied to a regulation device rather than modifying the output control signals directly. This methodology prevents the regulation device from treating actuator corrections as disturbances, optimizing the overall control strategy. Patent