Extension of the ALVl -Architecture for Robust Visual Gui Miniature Robot*
Markus Krabbes, Stephan Völker, Fachgebiet Neuroinformatik · 1997
Extensions of the ALVlNRCArchitecture are introduced for a KHEPERA-miniature robot to navigate visually robust in a labyrinth. The reimplemantation of the ALVINNapproach demonstrates, that also in indoor-environments a complex visual robot navigation is achievable using a direct input-output-mapping with u multilayerperceptron network, which is trained by expert-cloning. With the extensions it succeeds to overcome the restrictions of the small visualjield of the camera by completing the input vector with history-components, intrduction of the velocity dimension and evaluation of the network’s output by a dynamic neural jield. This creates the prerequisites to take turns which are no longer visible in the actual image and so make use of several alternatives of actions ($e. at crossings).