M. Jansen

Arizona State University, University of Bonn

Papers

Total Citations

H-Index

About

M. Jansen has pioneered the integration of neural network control theory with physical robotic systems, focusing on the critical gap between simulation and real-world deployment. Their foundational work on globally stable neural robot control, including the 1993 study on I/O-stability with a global neural net inverse model in the feedback loop, established rigorous stability guarantees for learning-based controllers—a rare achievement that bridges control theory and machine learning. Jansen’s 2005 paper on payload adaptation demonstrated how four separate three-layer perceptrons could learn complex dynamics (mass-coupling, Coriolis, viscous, and static friction forces) directly from point-to-point trajectories, enabling robots to autonomously adjust to varying loads without manual retuning. More recently, their 2017 work on fast prototyping for robot locomotion in real-world environments (8 citations) argued compellingly that morphology and controller design must be co-optimized with real environmental conditions, moving beyond simulation-only approaches. Though Jansen’s citation counts are modest, their contributions represent a principled, mathematically rigorous approach to neural robot control that has influenced subsequent work on stable learning-based control and embodied intelligence.