Transferring Insights on Mental Training to Robot Motor Skill Learning

Humans exhibit impressive abilities to plan and improve movements with mental forecasts. Through this, highly complex motor skills can be learned and enhanced in a short amount of time. Transferring these abilities of mental imagery and mental training to the field of robotics could possibly improve motor skill learning of robots as well. The approach introduced in this thesis focuses on the subject “Learning without doing”. To this end, this thesis presents a concept to train robots in "mental environments", for improved motor skill learning. The thesis connects the basic concepts of mental imagery used by humans with existing techniques of robotics. For this connection, neuropsychological inspired concepts are discussed and linked to robotic approaches for motor skill learning such as dynamical movement primitives (DMP). The thesis presents an overview of human motor skill learning and introduces a concept to transfer these structures to a mixture of experts based architecture of movement primitives. As a proof of concept, the single components of this introduced concept are applied to the example of robot ball catching and advantages and current limitations of the approach are discussed.

[1]  Stephen M. Kosslyn,et al.  A Simulation of Visual Imagery , 1977, Cogn. Sci..

[2]  M. Farah.  The neural basis of mental imagery , 1989, Trends in Neurosciences.

[3]  J. Decety,et al.  The cerebellum participates in mental activity: tomographic measurements of regional cerebral blood flow , 1990, Brain Research.

[4]  Jeffrey L. Elman,et al.  Finding Structure in Time , 1990, Cogn. Sci..

[5]  C. Hall,et al.  The Use of Imagery by Athletes in Selected Sports , 1990 .

[6]  Chung-Ming Kuan,et al.  Forecasting exchange rates using feedforward and recurrent neural networks , 1992 .

[7]  Andrew R. Barron,et al.  Universal approximation bounds for superpositions of a sigmoidal function , 1993, IEEE Trans. Inf. Theory.

[8]  Geoffrey E. Hinton,et al.  An Alternative Model for Mixtures of Experts , 1994, NIPS.

[9]  A. Sirigu,et al.  The Mental Representation of Hand Movements After Parietal Cortex Damage , 1996, Science.

[10]  Anil K. Jain,et al.  Artificial Neural Networks: A Tutorial , 1996, Computer.

[11]  D. Wolpert.  Computational approaches to motor control , 1997, Trends in Cognitive Sciences.

[12]  P. Moerland,et al.  Some Methods for Training Mixtures of Experts , 1997 .

[13]  Daniel B. Willingham,et al.  A Neuropsychological Theory of Motor Skill Learning , 2004 .

[14]  Preben Alstrøm,et al.  Learning to Drive a Bicycle Using Reinforcement Learning and Shaping , 1998, ICML.

[15]  Scott T. Grafton,et al.  Role of the posterior parietal cortex in updating reaching movements to a visual target , 1999, Nature Neuroscience.

[16]  Jerry Banks,et al.  Discrete Event Simulation , 2002, Encyclopedia of Information Systems.

[17]  Jun Nakanishi,et al.  Control, Planning, Learning, and Imitation with Dynamic Movement Primitives , 2003 .

[18]  Stefan Geyer,et al.  Imagery of voluntary movement of fingers, toes, and tongue activates corresponding body-part-specific motor representations. , 2003, Journal of neurophysiology.

[19]  S. Blakemore,et al.  Action prediction in the cerebellum and in the parietal lobe , 2003, Experimental Brain Research.

[20]  H. Groenewegen.  The Basal Ganglia and Motor Control , 2003, Neural plasticity.

[21]  D. Papineau.  The Oxford Companion to the Mind , 2004 .

[22]  Rick Grush,et al.  The emulation theory of representation: Motor control, imagery, and perception , 2004, Behavioral and Brain Sciences.

[23]  S. Kosslyn.  Mental images and the Brain , 2005, Cognitive neuropsychology.

[24]  R. Andersen,et al.  The posterior parietal cortex: Sensorimotor interface for the planning and online control of visually guided movements , 2006, Neuropsychologia.

[25]  Julien Doyon,et al.  Functional neuroanatomical networks associated with expertise in motor imagery , 2008, NeuroImage.

[26]  M. Desmurget,et al.  Movement Intention After Parietal Cortex Stimulation in Humans , 2009, Science.

[27]  J. Doyon,et al.  Contributions of the basal ganglia and functionally related brain structures to motor learning , 2009, Behavioural Brain Research.

[28]  Nicolas Schweighofer,et al.  Motor learning without doing: trial-by-trial improvement in motor performance during mental training. , 2010, Journal of neurophysiology.

[29]  Jan Peters,et al.  Learning table tennis with a Mixture of Motor Primitives , 2010, 2010 10th IEEE-RAS International Conference on Humanoid Robots.

[30]  Daniel Erlacher,et al.  Applications of lucid dreams: An online study , 2012 .

[31]  K. Semendeferi,et al.  Human prefrontal cortex: evolution, development, and pathology. , 2012, Progress in brain research.

[32]  Jan Peters,et al.  Probabilistic Movement Primitives , 2013, NIPS.

[33]  Tadas Stumbrys,et al.  The phenomenology of lucid dreaming: an online survey. , 2014, The American journal of psychology.

[34]  Dirk Pflüger,et al.  Modeling and Simulation - An Application-Oriented Introduction , 2016, Springer undergraduate texts in mathematics and technology.

[35]  Kristsana Seepanomwan.  Mental imagery in humanoid robots , 2016 .