Patrick Stalph
Papers
9
Total Citations
86
H-Index
6
About
Patrick Stalph’s research lies at the intersection of machine learning and robotics, with a core focus on enabling robots to learn their own kinematic models through online, adaptive algorithms. His major contributions center on the application and extension of the XCSF learning classifier system—a genetics-based, nonlinear regression tool—for flexible robot arm control. Stalph pioneered methods for learning local linear Jacobians, allowing robots to autonomously map joint configurations to task-space coordinates without relying on pre-defined, rigid models. His work demonstrates that XCSF can efficiently learn sensorimotor control structures, improving robustness by filtering noisy sensory information and handling multiple output dimensions through modularization. With key papers accumulating over 80 citations, his comparative studies of XCSF against algorithms like Locally Weighted Projection Regression (LWPR) have provided critical insights for the field. Stalph’s 2014 monograph, *Analysis and Design of Machine Learning Techniques*, synthesizes these evolutionary solutions for regression, prediction, and control, solidifying his role in advancing autonomous, adaptive robotics.
Research Focus
Key Achievements
Top Papers
- 1Learning local linear Jacobians for flexible and adaptive robot arm control24 citations · 2011
- 2Learning sensorimotor control structures with XCSF16 citations · 2009
- 3
- 4Function approximation with LWPR and XCSF: a comparative study8 citations · 2012
- 5Analysis and Design of Machine Learning Techniques8 citations · 2014
- 6Modularization of xcsf for multiple output dimensions6 citations · 2011
- 7
- 8
- 9Filtering sensory information with XCSF3 citations · 2012