Bill Yerazunis

Papers

Total Citations

H-Index

About

Bill Yerazunis is a robotics researcher whose work sits at the intersection of machine learning, control theory, and autonomous manipulation. His primary research areas include model-based reinforcement learning, Gaussian process regression for dynamical systems, and robotic assembly. Yerazunis has made significant contributions to developing learning frameworks that enable robots to operate effectively under real-world constraints—particularly when velocity and acceleration measurements are unavailable. His 2020 paper on derivative-free model learning for reinforcement learning (16 citations) introduced a novel approach using Gaussian Process Regression that allows robots to learn from position data alone, a critical advancement for systems with limited sensing capabilities. His most cited work (2019, 26 citations) applies semiparametric Gaussian processes to navigate a ball in a circular maze, demonstrating how robots can handle complex nonlinear effects like dry friction and contact dynamics. Most recently, his 2024 work on autonomous robotic assembly tackles the grand challenge of enabling robots to assemble functional products from arbitrarily arranged parts, bypassing traditional system constraints. With a growing citation record and a focus on practical, sensor-limited robotics, Yerazunis is pushing the boundaries of what autonomous systems can achieve in unstructured environments.