Execution Times Reconstruction in a LTI System Real-time Simulation

Abstract

A Real-time System (RTS) interacts with physical world by sensors, actuators, Analogical/Digital and Digital/Analogical convertors processing its requirements through Real-time Tasks (RTT). Each RTT is formed by an instances set with at least three time constraints: arrival time, execution time and deadline time. In this sense, it is important to propose a model to describe the behavior of this time constraints; specifically the execution time. The parameter model estimator considered is a digital filter based on Instrumental Variable to use on Autoregressive Moving Average Model (ARMA) (1, 1). The execution times were measured as experimental techniques using QNX 6.5 Neutrino real-time operating system in an AMD Phenom™ 9950 processor with speed of 2.6 GHz and 4GB in RAM memory with a different execution time measurement method. This paper presents a state of the art of Real-time task model, the execution times measurements, execution time model, experimental validation of the minimal probabilistic characteristics to use the Instrumental Variable estimator, experimentation and results. The algorithm was efficient because the execution time reconstructed converges to the execution time measured.