Human body posture mapping method applied to action imitation of anthropomorphic robot

Abstract

The invention discloses a human body posture mapping method applied to action imitation of an anthropomorphic robot. The method comprises the following steps that 1) the three-dimensional position information of notes of human skeleton series is acquired through a depth camera; 2) vectors of the human skeleton series are constructed according to the notes of the human skeleton series, virtual human joints are established according to the vectors of the human skeleton series and the robot joint structure so as to form a human skeleton model; 3) a connecting rod reference coordinate system basedon connecting rods of the human skeleton model is established; 4) the human body mapping joint angles are calculated by utilizing the human body connecting rod skeleton vectors according to the structure characteristics of robot joints; and 5) the human body mapping joint angles are applied to robot joints according to the limit of each joint angle of the robot. The method has the advantages thatthe human body mapping joint angles can be accurately calculated by adopting a geometric analysis method based on the connecting rod skeleton vectors and the virtual human joints, the human body posture can be mapped to the robot end through the joint angles, and therefore the calculation loss is low, and the accuracy is high.

Description

technical field [0001] The invention relates to the technical field of human-computer interaction, in particular to a human body posture mapping method applied to motion imitation of a humanoid robot. Background technique [0002] In recent years, robot technology has developed rapidly and has been more and more widely used in fields such as industry, medical treatment, scientific research, education and training, and family daily life. At the same time, the increasingly complex and diverse application environments put forward higher requirements for the adaptability and intelligence of robots. Robot imitation learning technology can improve robot learning efficiency, improve robot intelligence, and free developers from heavy programming work. It is an important part of imitation learning to properly represent the obtained human teaching information and apply it to the robot. The characterization link needs to establish a way to represent the taught actions and map the obs...

AI Technical Summary

Problems solved by technology

At present, most of the researches using the analytical method of geometric analysis have not combined the characteristics of the human body and the characteristics of the joint structure of the robot to accurately analyze the calculation of the joint angle, so to a certain extent, it limits the similarity of the humanoid robot's imitation of human gestures and actions.

Images