Neural Control, Memory, and Learning for Complex Behaviors in Multi Sensori-Motor Robotic Systems

• Applicant: Professor Dr.-Ing. Poramate Manoonpong
• Subject Area: Automation, Mechatronics, Control Systems, Intelligent Technical Systems, Robotics
• Term: from 2010 to 2016
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 175092235

Living creatures like walking animals and humans impress the observer by the elegance of their movements. They can perform diverse behaviors including anticipation and learning. However, the complex achievements of biological systems are still far from being realized or implemented in artificial agents due to complex multi-input multi-output sensori-motor coupling. Thus, the central goal of this proposal is to investigate and develop different neural mechanisms and learning together with proactive (anticipatory) decision making for generating complex biologically-inspired behaviors including natural movements and effective locomotion over difficult terrains, versatile proactive behaviors, versatile memory-guided behaviors, and goal-directed behaviors in physical walking robots by interacting with complex environments. All these complex behaviors have to cooperate or compete under the concept of modular neural control. Specifically this approach will for the first time combine five important neural aspects in the autonomous walking robots: 1) muscle model, 2) adaptive forward models (efferent copy), 3) short-term memory, 4) multi-synaptic plasticity (long-term memory), and 5) predictive motor models (action planning). It will, as a result, provide a better understanding of the general control, memory, plasticity, and predictive principles in embodied neural sensori-motor function. In addition to the development of the neural mechanisms, we will also develop an open-source multi sensori-motor robotic platform for performing here experiments and for studying artificial and biological walking systems in general.

• DFG Programme: Independent Junior Research Groups