Attitude Alignment Based on Hole Attitude Estimation in Robot Assembly

Abstract

This article proposes an assembly method based on hole attitude estimation. When the visual accuracy is not high enough and the force perception is easily affected by external interference, the existing assembly plannings are not perfect enough in hole attitude searches, such as low assembly efficiency and high contact force/torque. To improve assembly performance and reliability with limited measurement capability, the article proposes an assembly framework (AF) based on hole attitude estimation. Combined with the variable speed search strategy, the attitude of the hole can be accurately estimated through multiple swings. Under the same measurement conditions, this method reduces the number of searches and reduces attitude search time by up to 40.3% and average contact force/torque by up to 69.5% in various shapes of peg-in-hole assemblies. The experimental results verified the effectiveness of the method.