Jan Kneissler

Papers

Total Citations

H-Index

About

Jan Kneissler’s research lies at the intersection of evolutionary computation and robotics, with a focus on how learning classifier systems can enable adaptive, robust control. His most significant contribution is demonstrating that XCSF—a nonlinear regression system rooted in evolutionary algorithms—can efficiently learn to control a robot arm, even when sensory information is noisy or incomplete. In his highly cited 2013 paper (12 citations), Kneissler introduced a filtering mechanism that leverages forward velocity kinematics, allowing the system to predict how motor actions change the arm’s state. This innovation dramatically improved learning robustness and control performance, addressing a critical challenge in real-world robotics: dealing with imperfect sensor data. His earlier 2012 work (3 citations) laid the groundwork by first exploiting predictive knowledge of motor activity. While his citation counts are modest, Kneissler’s work is notable for bridging theoretical evolutionary computation with practical, noisy physical systems—a difficult and often underappreciated task. For students and researchers, his research offers a compelling example of how biologically inspired learning algorithms can be adapted for real-time, sensor-driven control in robotics.