Max Robotics — Global Service Robot Directory + US Certification

About

Craig Sherstan is a researcher whose work spans two interconnected domains: intelligent prosthetics control and reinforcement learning. His early contributions focused on improving the usability of myoelectric prostheses for upper-limb amputees, applying real-time machine learning to reduce the cognitive burden of device operation. His most-cited work, "Application of real-time machine learning to myoelectric prosthesis control" (2015, 62 citations), demonstrated how learned predictions about user intent could meaningfully improve prosthetic responsiveness. Alongside collaborative multi-joint control systems and a broader vision of prosthetic arms as wearable intelligent robots, Sherstan helped advance the field toward more intuitive, adaptive assistive devices. His research later evolved toward foundational questions in reinforcement learning, particularly around temporal abstraction and value estimation. His work on Gamma-Nets (2020) explored how agents can generalize value estimates across multiple timescales—an important challenge in long-horizon decision-making. He has also investigated how the successor representation can accelerate learning in predictive knowledge frameworks grounded in general value functions. Taken together, Sherstan's research reflects a commitment to building intelligent systems that are both theoretically grounded and practically deployable in dynamic, real-world environments.