The increasing complexity of the future wireless networks leads to the requirement for self-organization. This is true especially in home networking where users are typically not networking professionals and cannot be expected to perform complex optimization and management tasks. In this context, cognitive radio concept combining cross-layer optimization and learning mechanisms is a promising solution. We demonstrate a cognitive home networking prototype, which addresses practical problems users face with the present-day wireless networks at home. The prototype shows how nodes using IEEE 802.11 radios and WARP boards operate under the Cognitive Resource Manager (CRM). The nodes achieve the desired performance by handling network dynamics and controlling parameters taking independent or cooperative decisions and operating in different layers of the protocol stack. This is done using multiple control loops which are supported by the CRM architecture. We demonstrate the use of machine learning for online estimation of network activity patterns to enable more efficient Dynamic Spectrum Access (DSA) using Hidden Semi-Markov Models (HSMM). The demonstration showcases dynamic spectrum allocation and policy-based behavioral changes in a home environment, where several multimedia streams and data communication flows are competing against each other and against external, also primary, interferers.
Self-organizing home networking based on cognitive radio
ZANELLA, ANDREA
2011
Abstract
The increasing complexity of the future wireless networks leads to the requirement for self-organization. This is true especially in home networking where users are typically not networking professionals and cannot be expected to perform complex optimization and management tasks. In this context, cognitive radio concept combining cross-layer optimization and learning mechanisms is a promising solution. We demonstrate a cognitive home networking prototype, which addresses practical problems users face with the present-day wireless networks at home. The prototype shows how nodes using IEEE 802.11 radios and WARP boards operate under the Cognitive Resource Manager (CRM). The nodes achieve the desired performance by handling network dynamics and controlling parameters taking independent or cooperative decisions and operating in different layers of the protocol stack. This is done using multiple control loops which are supported by the CRM architecture. We demonstrate the use of machine learning for online estimation of network activity patterns to enable more efficient Dynamic Spectrum Access (DSA) using Hidden Semi-Markov Models (HSMM). The demonstration showcases dynamic spectrum allocation and policy-based behavioral changes in a home environment, where several multimedia streams and data communication flows are competing against each other and against external, also primary, interferers.Pubblicazioni consigliate
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.