Overview
The number of ways in mobile devices can be wirelessly connected to networks is increasing. Handheld devices already exist that support multiple wireless technologies and this is set to become the norm. However, this increased choice in connectivity presents users with the problem. They must select the optimal network technology for the task they are performing. This is a complex choice. It requires users to consider multiple technological and financial constraints. Different applications require different qualities of service (QoS) from the underlying wireless networking technologies and incurring also different costs.
Approach
The aim of this research is to develop an intelligent agent to the perform the network selection task on behalf of the user. This agent will then be embedded in a protocol stack dynamically selecting and maintaining the optimal match between an applications network requirements and the capabilities of the available network technologies and the users preference.
A software agent is embedded in a wireless device stack. Software to manage link-layer handovers requires up-to-date statistic information from the network stack or applications, and be responsive to the dynamic events from the environment. Using cross-layer design, the agent to interface with protocol module.
The agent is modelled on AI concepts from behaviour-based design [2] and code from robotic design toolkits [3]. The agent uses sources of data to make handover and link selection decisions.
