status 5 Dec 2011

VU is generating logs to analyse for preference-based learning.
Ran tests with QI as internal reward (based on gps first, now enhanced with accelerometer and joint force-feedback - not separately tuned, though); leads to comparable results as using distance as a reward.