Autonomous sensing method for intelligent processing machine

Abstract

The invention relates to an autonomous sensing method for an intelligent processing machine. The method comprises: normalizing the data of multiple sensing areas of the intelligent processing machinethrough feature vectors to a unified standard interval and extracting as valid data, thereby acquiring effective feature variables; modeling the dynamics of an unknown nonlinear processing system witheffective feature variables according to different processing modes; using a dynamic RBF neural network identifier to locally approximate the unknown dynamics of the processing system modeled, thereby acquiring a learning training result of the dynamic RBF neural network to establish a constant neural network and form a mode library of the constant neural network under different processing modes;constructing a set of dynamic estimators based on the constant neural network; and performing the difference operation between the timing dynamic data currently received by a sensor and the dynamic estimator and judging, thereby achieving the autonomous sensing. The autonomous sensing is performed by using a dynamic mode recognition method through effective physical features, thereby achieving amore concise and accurate autonomous sensing method for the intelligent processing machine.

Description

technical field [0001] The invention relates to the technical field of intelligent control, in particular to an autonomous sensing method of an intelligent processing machine. Background technique [0002] With the scale and complexity of the manufacturing industry, the manufacturing equipment and production environment have become more and more complex, and the dynamic characteristics of the manufacturing process have also been continuously enhanced, resulting in an increasing number of decisions for production planning and control. The degree of flexibility of the currently commonly used flexible manufacturing systems is constrained by the complexity of decision-making. If the complexity of the system cannot be accurately grasped and controlled before the decision-making, the flexible manufacturing system may not be able to adapt to the change of the processing object, resulting in the inflexibility of the flexible system. , unable to complete the more complex manufacturin...

AI Technical Summary

Problems solved by technology

[0004] Compared with other processing machines, the biggest difference between intelligent processing machines is that they have the ability to perceive themselves and the outside world actively or passively. Various learning algorithms are usually used to model the autonomous perception of intelligent processing machines. Since intelligent processing machines All information is random time series information, and its autonomous perception can be regarded as a process of identification and identification of dynamic patterns in essence, and the existing learning algorithms cannot solve the problem of fast identification of dynamic patterns well, making The autonomous perception process of intelligent processing machines has high computational complexity but insufficient accuracy

Images