CAREER: Motion Capabilities for Autonomous Mobile Manipulation

University Of Massachusetts Amherst, Amherst MA

Investigators

Abstract

CAREER: Motion Capabilities for Autonomous Mobile Manipulation Abstract Concerted efforts in industry and academia have lead to the design and construction of highly advanced humanoid robots and mobile manipulators. These mechanisms combine dexterous manipulation capabilities with mobility and are equipped with multi-modal sensor suites. Autonomous motion of these complex mechanisms is subject to a multitude of constraints, each requiring the consideration of feedback at a specific frequency. The investigator proposes to devise a novel motion generation generation approach that addresses motion constraints and their feedback constraints. At the low end of the frequency spectrum, the investigator proposes the application of techniques from active learning and utility theory to global motion planning. These techniques permit the exploitation of structure inherent in any real-world motion-planning problem to increase the efficiency of motion planning. The resulting planning methods will satisfy the feedback requirements of autonomous mobile manipulation. To obtain adequate feedback about the state of the environment, perceptual information is provided by novel sensing and modeling techniques, which will also be developed as part of this effort. To address motion constraints that require high-frequency feedback, the investigator proposes to combine multi-objective control methods with planning approaches for constraint satisfaction. These planning methods overcome the susceptibility of control-based motion generation to local minima. The resulting overall approach to motion generation addresses the motion requirements of autonomous mobile manipulation, offering a significant advance towards the autonomous operation of humanoid robots and mobile manipulators in unstructured environments.