In recent years research in autonomous robots has been more and more successful in developing algorithms for generating behavior from a generic task-independent objective. However, without a specific task it is difficult to evaluate the behavior. The same difficulty is also faced in characterizing the behavior of freely moving animals. Together with Eckehard Olbrich, I investigate methods based on information theoretic quantities that are able to deal with high-dimensional systems. An important feature of our method is to provide a length-scale-dependent quantity. This allows to isolate the complexity of behavior on the coarse level, on finer levels and on the noise level. We apply this method to toy examples and data from high-dimensional robotic systems.
Paper: G. Martius and E. Olbrich. Quantifying emergent behavior of autonomous robots. Entropy, 17(10):7266, 2015. [ bib | DOI | http ]
See also the Supplementary page.