Onboard Learning of Adaptive Behavior: Biologically Inspired and Formal Methods

Abstract

Long lifetime autonomous ground vehicles traveling in rough terrain will need a higher level of autonomy with behaviors that must also adapt to declining vehicle health and unknown environmental conditions. For example, the MER (Mars Exploration Rovers) Spirit and Opportunity have both passed 1900 days of life on the Martian surface, with possible extensions to 2200 days and beyond depending on rover health. Changes in navigational planning due to degradation of the drive motors as they pass their lifetime are currently done on Earth for the Spirit rover. This paper reviews biologically inspired and formal methods of adaptive control and presents an algorithm for onboard learning of navigation strategies using adaptive weight learning within a free flow hierarchy (FFH) behavior framework. We also present the results of some field studies.