Integral control with error modulation in a FAM-based agent for a furuta inverted pendulum

The task oriented approach in the FAM-based agent architecture creates a problem for control laws with integral action. This article proposes a solution to the integral control problem of this architecture. It consists in the modulation of the error signals in the control laws to limit the integral action when the associated condition is not operative. The error modulation makes use of the membership value of the associated fuzzy condition in the agent. Three simulations prepared with the dynamic model of a Furuta inverted pendulum illustrate the integral control problem and the solution proposed. The agent is evaluated in terms of its temporal response.