Behavior Coordination of Mobile Robotics Using Supervisory Control of Fuzzy Discrete Event Systems

Abstract

In order to incorporate the uncertainty and impreciseness present in real-world event-driven asynchronous systems, fuzzy discrete event systems (DESs) (FDESs) have been proposed as an extension to crisp DESs. In this paper, first, we propose an extension to the supervisory control theory of FDES by redefining fuzzy controllable and uncontrollable events. The proposed supervisor is capable of enabling feasible uncontrollable and controllable events with different possibilities. Then, the extended supervisory control framework of FDES is employed to model and control several navigational tasks of a mobile robot using the behavior-based approach. The robot has limited sensory capabilities, and the navigations have been performed in several unmodeled environments. The reactive and deliberative behaviors of the mobile robotic system are weighted through fuzzy uncontrollable and controllable events, respectively. By employing the proposed supervisory controller, a command-fusion-type behavior coordination is achieved. The observability of fuzzy events is incorporated to represent the sensory imprecision. As a systematic analysis of the system, a fuzzy-state-based controllability measure is introduced. The approach is implemented in both simulation and real time. A performance evaluation is performed to quantitatively estimate the validity of the proposed approach over its counterparts.

A Hierarchical Hybrid Architecture for Mission-Oriented Robot Control

“…Based on the information contained on those vectors an arbitrator must determine the actions to perform. This paradigm is later extended on [12,13].…”

Novel Yinger Learning Variable Universe Fuzzy Controller

“…Where h is radius of local space (  k ), and data-window is changed by adjusting it. (,) DAB is the distance function which is defined by (8) (,) ,, ( ,)(,) (,)…”

Logic inference based construction of a supervisor for a discrete event system

“…Indeed, SCT is actively utilized in robotics nowadays. Recent publications in this area concern single robot control [11], [12], robot groups control [13], [14], [15], robots formation control [16], and swarm robotics [17], [18]. On the other hand, the inference search machines, implemented at a robot control system, are applied to represent and process the knowledge obtained during the robot functioning.…”