Mathis Richter
Papers
10
Total Citations
134
H-Index
6
About
Mathis Richter is a leading researcher in neurally inspired cognitive robotics, specializing in Dynamic Field Theory (DFT) and embodied cognition. His work centers on developing autonomous robotic architectures that bridge high-level behavioral organization with low-level sensory-motor control, enabling robots to perform complex, human-like tasks in real time. Richter’s most influential contribution is the creation of the **cedar software framework** (over 30 combined citations), a powerful tool for building and testing embodied cognitive systems that integrate perception, action, and learning. His seminal 2012 paper on a human-cognition-inspired robotic architecture (65 citations) established foundational principles for autonomous action selection and behavioral sequencing. Richter has also advanced reinforcement learning in neural dynamics, introducing DN-SARSA(λ) for learning behavioral sequences from delayed rewards, and developed methods for parsing action sequences and understanding spatial relations in visual scenes. His recent work on diagonal structured state space models for efficient streaming processing on neuromorphic hardware (Loihi 2) demonstrates ongoing innovation at the intersection of cognitive science and energy-efficient AI. With over 130 total citations, Richter’s research provides essential tools and theoretical insights for building truly autonomous, embodied artificial cognitive systems.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3
- 4Parsing of action sequences: A neural dynamics approach8 citations · 2015
- 5Autonomous reinforcement of behavioral sequences in neural dynamics7 citations · 2013
- 6
- 7
- 8Autonomous Robot Hitting Task Using Dynamical System Approach5 citations · 2013
- 9
- 10Autonomous Reinforcement of Behavioral Sequences in Neural Dynamics2 citations · 2012