Assembly task recognition with planar, curved and mechanical contacts

The assembly-plan-from-observation (APO) method for robot programming is described. The APO method aims to build a system that has threefold capabilities. The method observes a human performing an assembly task, it understands the task based on this observation, and it generates a robot program to achieve the same task. Emphasis is placed on the main loop of the APO, which is task recognition. Using object recognition results, the task recognition module determines what kind of assembly task is performed. The method extends the task recognition module to handling curved objects and other mechanical contacts. Task models for these cases are defined. Two concepts are verified, i.e., that such task models are useful in recognizing assembly tasks, and that it is possible to generate robot motion commands for repeating the same assembly task