PBCA - Prediction Based Channel Allocation

Cellular infrastructures as they exist today are statically configured, and do not provide scalable, flexible and adaptive radio resource reconfiguration. In GSM networks, many dynamic channel allocation (DCA) schemes have been proposed to improve the network performance and user experience. DCA schemes can be categorized into cell based centralized, cell based distributed, and mobile handset based solutions. However, most of these existing schemes are event-driven and are reactively triggered by the arrival of active calls, which has a delayed response to network traffic condition. In order to adjust network configuration to the changing traffic conditions, a new class of information driven adaptation schemes are needed. The emergence of next generation smart phones with considerable computation capability and the localization technologies adopted in cellular networks enable a new class of proactive and prediction-based information driven network reconfiguration schemes. This paper makes a case for Prediction Based Channel Allocation algorithm as a means to dynamically reconfigure cellular network channel resource based on location and traffic prediction. Simulation results show that the proposed scheme outperforms an existing DCA scheme in term of channel allocation efficiency under various network conditions, and meanwhile shows a good tolerance to prediction errors.