Algorithm to Automatically Generate Non-Collision Trajectory to Perform Pick-and-Place Operation with a Mobile Manipulator

In recent years, the use of mobile manipulators for performing pick-and-place operations has increased in multiple workspaces. However, mobile robots often struggle with precise positioning owing to large position estimation errors. Additionally, working with a robot arm can be inconvenient because operators must teach the robot arm each individual workspace. In this study, to address these challenges, we propose a coordinate system matching method based on marker recognition. Additionally, we present an algorithm that creates a trajectory on the basis of the environmental information obtained through sensors and the marker frame of the recognized workspace. This trajectory is then corrected to avoid collisions between the robot and the environment, thereby allowing the mobile manipulator to perform pick-and-place operations in multiple workspaces without operator instructions.