As noted one type of control that can be applied to an intelligent actuator is neural network control. These have been applied to various types of problems in all areas of science and technology. There are two main steps in the application of neural network control to intelligent actuators.
Feed forward network and nonlinear control have been shown to be most appropriate for continuous and differential activation function as associated with intelligent actuators. This is due to the universal approximation capability as it applies to intelligent actuators. Recurrent networks are a type of network used with intelligent actuators and are commonly used for applications such as system identification. If we have a set of input-output data, system identification is able to map the data pairs using a type of mapping logic. Using this type of network to control the intelligent actuator is able to capture the dynamics of a system.
The next type of control we are discussing for the intelligent actuator is the Bayesian control probability. A large number of intelligent actuator control algorithms have been produced using Bayesian probability. The controls typically serve as state and space estimators of the variables used in the control of the intelligent actuator.
As noted these are types of artificial intelligence of machines and intelligent actuators. An intelligent actuator using artificial intelligence perceive information about their environment and take action to maximize the positive outcome of the application and control of the intelligent actuator.
Thomson has partners in design and development that can assist you in your intelligent actuator application Call us today to discuss your needs.