A motion imitation and augmentation system comprising an imitation system, a self-collision avoidance system, and a motion augmented system was proposed in this study. The imitation system captures the human skeleton by using an RGB-D camera and a method involving forward kinematics and the space vector method was developed and integrated to map the human joint angles to the motor angles of the robot's arm. To improve safety of the system, a self-collision avoidance system was incorporated so that the various parts of the robot do not collide with each other while imitating human motions. The imitated motion was recorded and stored in a motion dataset as a reference motion. To address different situations and objects, the reference motion was adjusted using the motion augmented system in which both the adjusted trajectory and its corresponding motor angles were generated using the integrated adaptive constricted particle swarm optimization algorithm. So that, the adjusted motion can adapt to the real situation while maintaining similarity with the reference motion. Experiments conducted in practical situations demonstrated the efficiency of the motion imitation and augmentation system. The robot can mimic a reference motion by using its vision and can generate a suitable motion based on the imitated motion for coping with different objects and situations.
All Science Journal Classification (ASJC) codes
- Computer Science(all)
- Materials Science(all)