Assembly Planning from Observations under Physical Constraints
This addresses the challenge of automated assembly planning for robotics, enabling copying of unknown structures from visual input, though it is incremental as it builds on existing methods for detection and planning.
The paper tackles the problem of replicating an unknown assembly of objects from a single photo using object detection and pose estimation, by developing an algorithm that combines physical stability constraints, convex optimization, and Monte Carlo tree search to plan pick-and-place sequences, achieving efficiency and robustness to errors in real robotic systems as demonstrated on a UR5 manipulator.
This paper addresses the problem of copying an unknown assembly of primitives with known shape and appearance using information extracted from a single photograph by an off-the-shelf procedure for object detection and pose estimation. The proposed algorithm uses a simple combination of physical stability constraints, convex optimization and Monte Carlo tree search to plan assemblies as sequences of pick-and-place operations represented by STRIPS operators. It is efficient and, most importantly, robust to the errors in object detection and pose estimation unavoidable in any real robotic system. The proposed approach is demonstrated with thorough experiments on a UR5 manipulator.