A Neural Network Based Testbed for Modelling Sensorimotor Integration in Robotic Applications1

To model realistic biological behaviors such as perception and motor interactions, one requires a flexible simulation environment in which goal-oriented action and sensory perception can be integrated to produce intelligent behavior. Many researchers have focused their work primarily on one specific aspect of this loop of action and perception. Our challenge is to demonstrate a feasible environment in which both sensory processing as well as motor actions are integrated simultaneously and attached to the outside world. In this report, we describe the preliminary work implemented in our lab towards the integration of sensorimotor interactions using Neural Simulation Language (NSL) and the Rapid Robotics Application Development environment (R2AD) for computing sensory information and creating motor behavior respectively for the purpose of visuomotor coordination in the real world.