Data buffering in mobile robot control systems

Abstract

The article deals with the question of data buffering, which comes from sensors to the control system of a mobile robot. The mobile robot control algorithms are developed on the assumption that the sampling frequency of sensor signals is constant. However, in real systems the execution time of the program between consequent sensor requests is random. To reduce the error of signal processing it is necessary to use data buffering. The dependencies for calculating the time between two consequent requests to the sensors in a polling procedure implemented by a digital von Neumann type controller are found. Using the queuing theory the probability distribution of the number of occupied buffer cells is obtained. The minimal buffer size to ensure the required reliability of data transmission is determined. The performed computer modeling determined that the estimated buffer size should be corrected upwards.