Autonomous Adaptation to Simultaneous Unexpected Changes in Modular Robots

To accomplish tasks in the real world, a robot (especially a modular and self-reconfigurable one) must be able to autonomously react and adapt to unexpected and possibly simultaneous changes to its self (such as its sensors, gaits, goals and modules) and dynamic situations/configurations in the environment. Failures, faults and reconfiguration that commonly occur in modular robots result in simultaneous changes that are at times unexpected as well as unpredictable. This paper presents an integrated technique called Surprise-Based Learning (SBL), which aims to enable a reconfigurable robot to autonomously adapt to changes. The main idea is to have a robot perform both learning and representation in parallel by constructing and maintaining a predictive model about itself and the environment. Modular reconfigurable robots, developmental robotics, learning, unexpected changes, fault-tolerance, failure-tolerance, surprises, adaptation.