Designing a Self-Optimization System for Cognitive Wireless Home Networks

Abstract

In this paper, we describe the design and implementation of an extensible and flexible self-optimization framework for cognitive home networks that employs cognitive wireless networking and agent-based design principles. We provide a “first-principles” derivation of its architecture based on a careful analysis of user requirements on wireless home networks (WHNs), state the related design objectives and constraints, and address those in the context of present-day and emerging radio platforms. We utilize the cognitive resource manager as an architecture of the individual agents. This architecture serves as a “constraint that de-constrains,” i.e., it allows achieving high system flexibility, while providing structural constraints to ensure robustness. We show that the designed system is capable of solving complex utility maximization problems constrained with user, operator, and regulatory policies in the crowded ISM bands. The system successfully operates on an extensible parameter configuration space across multiple protocol layers. For example, the prototype employs diverse optimization algorithms, and can also benefit from radio environment map information on primary transmitters propagated through the integrated policy mechanism. The proposed system delivers both efficient and robust radio resource management, and enables comfortable WHN experimentation by providing extensibility of sensory inputs, actuation parameters, and optimization algorithms.