Modeling human target reaching with an adaptive observer implemented with dynamic neural fields. (December 2015)

• Record Type: Journal Article
• Authors: S. Fard, Farzaneh; Hollensen, Paul; Heinke, Dietmar; Trappenberg, Thomas P.
• Is Part Of: Neural networks. Volume 72(2015:Dec.)
• Page Start: 13
• Page End: 30
• Publication Date: 2015-12
• Subjects: Target reaching -- Dynamic neural fields -- Path integration -- Internal model -- Observer -- Adaptive controller; Neural computers -- Periodicals; Neural networks (Computer science) -- Periodicals; Neural networks (Neurobiology) -- Periodicals; Nervous System -- Periodicals; Ordinateurs neuronaux -- Périodiques; Réseaux neuronaux (Informatique) -- Périodiques; Réseaux neuronaux (Neurobiologie) -- Périodiques; Neural computers; Neural networks (Computer science); Neural networks (Neurobiology); Periodicals; 006.32
• DOI: 10.1016/j.neunet.2015.10.003
• Languages: English
• ISSNs: 0893-6080

Abstract

Abstract: Humans can point fairly accurately to memorized states when closing their eyes despite slow or even missing sensory feedback. It is also common that the arm dynamics changes during development or from injuries. We propose a biologically motivated implementation of an arm controller that includes an adaptive observer. Our implementation is based on the neural field framework, and we show how a path integration mechanism can be trained from few examples. Our results illustrate successful generalization of path integration with a dynamic neural field by which the robotic arm can move in arbitrary directions and velocities. Also, by adapting the strength of the motor effect the observer implicitly learns to compensate an image acquisition delay in the sensory system. Our dynamic implementation of an observer successfully guides the arm toward the target in the dark, and the model produces movements with a bell-shaped velocity profile, consistent with human behavior data. Highlights: We developed an internal model for arm control in the neural field frame-work. After learning the model can point to a target without sensory feedback. The robot learns to move in arbitrary directions from few examples. The model adapts to changing forces on the motors and delays in the system. All movements have a bell shaped velocity profile, consistent with human behavior.