This interesting little robot is trying to overcome one of the biggest challenges of bipedal robots, keeping the system stable. The team at the University of Notre Dame is expoloring this challenge by using their robot KURMET for investigating non-steady-state dynamic maneuvers such as jumping. Legged robots offer advantages in terms of mobility over uneven terrain and, particularly in the case of bipeds, mobility in human-centered environments. (This robot isn’t supported by the rod off to the side)
The “fuzzy” control of the robot refers to the training process researchers use to control the robot. There aren’t a set of pre defined actions, instead, KURMET is acting based on a set of inputs that it can replicate. Maybe these evolutionary learning strategies can be used more widely in the future to simplify programming for robotics.
+30
