Hands-on teaching manipulator system and method with imitation learning mechanism

Abstract

The invention provides a hand-by-hand demonstration mechanical arm system with an imitative learning mechanism and a method. The system comprises an action execution module, a sensing module and a learning module, wherein the action execution module comprises a mechanical arm and steering engines; the sensing module comprises a plurality of sensors; and the learning module comprises a controller. According to the system, on the basis of an independent module power supply mode, hand-by-hand off-line demonstration is employed, an action detection device, which comprises gyroscope sensors and an accelerometer, is carried as the sensing module on the mechanical arm to collect the status information of links of the mechanical arm in motion, and then an imitative learning algorithm is applied to the information to guide the mechanical arm system to imitate learning demonstration activities. According to the system and the method, hierarchical control is adopted, the pertinence and information transmission efficiency of a control system are improved, the purpose of demonstration activities can be learned by learning, an imitation task can be still finished when the initial attitude of the mechanical arm or the position of a target object is changed, and a high intelligence degree is achieved.

Description

technical field [0001] The invention relates to a system and method for a hand-held teaching mechanical arm with an imitation learning mechanism, which belongs to the category of intelligent robots. Background technique [0002] The development of robots has considerably affected many aspects of modern human life. People's intelligence requirements for robots are constantly increasing so that they can serve humans better. Robot learning is an effective way to improve the intelligence of robots. In nature, imitation is the most direct and effective learning method for people or animals to master motor skills. Humans and animals are the bionic prototypes of robots, so it is feasible to apply the mechanism of imitation to robots. Many skills or behaviors of humans or animals are gradually formed and developed during the cognitive process of their nervous system. By imitating humans or animals, the cognition and behavior of robots are closer to humans or animals, thereby produc...

AI Technical Summary

Problems solved by technology

At the same time, it reduces the labor intensity of the traditional hand-held teaching method, reduces the cost of the sensor, overcomes the cumbersome image processing problem caused by the use of camera technology to collect teaching behavior, and improves the learning efficiency of the manipulator system.

Images