Improving the quality of service in behavioural-based mobile robotic applications

In complex control systems such as autonomous guided vehicles, where the environment is unpredictable, mobile robotic architectures should incorporate internal monitoring systems, in order to dynamically measure the vehicle quality of service, and adjusting the control robot parameters for the best vehicle fit. In this paper an adaptive quality of service management system (AQoSMS), integrated in an intelligent planner of a real-time mobile robotic architecture is presented. The performance parameter requirements requested from the application level, configured depending on the robot state and on the environment, the real values of the performance parameters and the available slack time in the system provided by the real-time operating system are traded off in run-time by the AQoSMS. In order to satisfy and improve the quality indexes required for the current mission, the AQoSMS performs the adaptation of robot control parameters (i.e. speed) or temporal characteristics of the heterogeneous load executed by the system (i.e. task period or computational time).