The State of the Art of Search Strategies in Robotic Assembly

Abstract Assembly robots have been widely used in the manufacturing industry. However, performing precise assembly tasks still poses a great challenge for robots due to numerous sources of the uncertainties such as fixtures, end effectors, or actuators, and the most critical technical bottleneck is to find the best search strategy to improve positioning accuracy in assembling. Search strategies are evaluated in terms of executing time, precision, stability, and applicable geometries and features of parts. This review is highly motivated to gain the state of the art of search strategies and identify some technical means which can further improve existing algorithms. Without losing the generality of this review discussion, robotic assemblies for peg-in-hole (PiH) are focused. Search strategies are classified and discussed based on different criteria respectively such as the number and types of sensors, and the dimensions of search spaces. To tackle with the assembly operations in an ill-structured environment, a search strategy has to fuse and utilize the sensing data from multiple sources such as the integration of force sensing and vision, and the search strategies with the sensing cooperation are explored. As the conclusion, the future research directions on search strategies for automated assembling have been discussed.