Control system and method for learning variable impedance

Abstract

The invention provides a control system and method for learning variable impedance. The system mainly comprises a variable impedance controller, a system Gaussian process model, a variable impedance control strategy, and a strategy learning algorithm. The system does not need any priori knowledge of an environment, and the system Gaussian process model is constructed according to the interaction data. The long-term inference and planning of the system is carried out in a Bayesian mode. The system can extract more useful information from the limited observation data, and completes a complex force control task through the least interaction time. Through the adding of an energy los item to a cost function, the system achieves the balance between an error and energy, and enables a robot to begood in compliant capability. Finally, the obtained variable impedance control strategy allows the target rigidness and damping parameters to be adjusted at the same time at different stages of the task. The system and method can be widely used for the compliant control tasks: double-mechanical-arm assembly, multi-mechanical-arm cooperation and robot gait control, and guarantee the safety and robustness of the interaction operation.

Description

technical field [0001] The invention relates to a robot compliance control, in particular to an efficient learning variable impedance control system and control method. Background technique [0002] As robots are increasingly used in contact manipulation tasks in unstructured environments, such as compliant assembly and human-computer interaction, it is difficult to establish an accurate dynamic model of the system due to the complexity of the tasks and the variable and unpredictable contact environment. How to make the robot perform new tasks safely, efficiently and quickly, and precisely control the contact force in different environments is a new challenge for the robot. Impedance control is widely used in robot interactive control tasks due to its good adaptability and robustness. Since the force control characteristics are determined by the impedance parameters of the robot, the selection of inertia, stiffness, and damping parameters is highly task-dependent and is usu...

AI Technical Summary

Problems solved by technology

[0003] For manipulation tasks that require force control, the number of learning explorations should be as low as possible, since a large number of physical interaction attempts may cause damage to the robot or workpiece, and large sampled data are time-consuming and expensive, which is impractical

Images