Adaptive control in multi-task mobile robotic applications

Abstract

Mobile robotic systems are often conditioned by limitations in available resources. Normally these systems try to offer a robust behaviour while pursuing several goals simultaneously, which leads to the execution of several tasks competing for CPU time and other shared hardware resources as sensors and actuators. If not properly managed, these limitations may provoke poor performance and, even worse, blocking of the entire system operation. In this work we propose a series of control policies for dynamic adaptation that have been integrated on a modular architecture. We also present two demonstrators to evaluate working implementations on real world applications.