Intelligent robotic manipulation of cluttered environments

2014-12-08Robotic household assistants of the future will need to understand their environment in real-time with high accuracy. There are two problems that make this challenging for robots. First, human environments are typically cluttered, containing a lot of objects of all kinds, shapes and sizes, in close proximity. This introduces errors in the robot’s perception and manipulation. Second, human environments are highly varied. Improving a robot’s perceptual abilities can tackle these challenge only partially. A robot’s ability to manipulate its environment can help in enabling and overcoming the limits of perception. ❧ We test this idea of manipulation-aided perception in the context of sorting and searching in cluttered, bounded, and partially observable environments. The inherent uncertainty in the world state forces the robot to adopt an observe-plan-act strategy where perception, planning, and execution are interleaved. Since execution of an action may result in revealing information about the world that was unknown hitherto, a new plan needs to be generated as a consequence of the robots actions. Since manipulation is typically expensive on a robot, our goal is to reduce the number of object manipulations required to complete the desired task. ❧ This thesis presents planning algorithms for a robot’s intelligent physical interaction with its cluttered environment. The focus is on using simple manipulation primitives to declutter the world and making task completion easier and faster using the environment’s local structure or context. For object sorting, we present a robust pipeline that combines manipulation-aided perception and grasping to achieve more reliable and accurate sorting results. In the context of environment exploration, we present an adaptive look-ahead algorithm for exploration by prehensile and non-prehensile manipulation of the objects in the environment. This algorithm is then applied to manipulation-based object search in real world. Finally, we add contextual structure to the world in the form of object-object co-occurrence relations and present an algorithm that uses context to guide the object search. ❧ We evaluate the performance and applicability of our planners through extensive simulations and real-world experiments on the PR2 robot. Our results show that purposeful manipulation of clutter to aid perception becomes increasingly useful (and essential) as the clutter in the environment increases, and that intelligent manipulation of a cluttered environment improves the efficiency of robotic tasks