Co-operative Extended Kohonen Mapping (EKM) for Wireless Sensor Networks

Abstract

This paper discusses a methodology to manage wireless sensor networks (WSN) with self-organising feature maps, using cooperative Extended Kohonen Maps (EKMs). EKMs have been successfully demonstrated in other machine-learning contexts such as learning sensori-motor control and feedback tasks. Through a quantitative analysis of the algorithmic process, an indirect-mapping EKM can self-organise from a given input space, such as the WSN's external factors, to administer the WSN's routing and clustering functions with a control parameter space. Preliminary results demonstrate indirect mapping with EKMs provide an economical control and feedback mechanism by operating in a continuous sensory control space when compared with direct mapping techniques. By training the control parameter, a faster convergence is made with processes such as the recursive least squares method. The management of a WSN's clustering and routing procedures are enhanced by the cooperation of multiple self-organising EKMs to adapt to actively changing conditions in the environment.

Cooperative agent-based SANET architecture for personalised healthcare monitoring

“…The evaluation of co-operative EKMs on SANETs is executed using the framework conceived by Chaczko, et al [4]. The Jadex framework by Universität Hamburg, shown in Figure 3: Java Simulation Environment [5] allows for convenient monitoring and tracking of SANET events by pre-assigning event trajectories in the network field in an interactive manner.…”

“…Reinforcement learning is driven by continuous control space functions [17], as indirect-mapping methods provide fluid decision choices than direct mapping. The accuracy in sensory stimuli control is important where external factors directly affect the network's robustness and reliability, allows for developing a feature map approach using co-operative EKMs with indirect mapping to improve the responsiveness of the tracking process [4] [17].…”

“…The environmental concern shown in Figure 2: Cooperative EKM Process [4] for a given SANET domain can be summarized in the following statement tasks:…”

“…Reinforcement learning heuristics, based upon nondeterministic biological concepts, provides a best effort optimum routing and clustering organization solution depending on the current situation [4]. …”

Sensor Actor Network Modeling utilizing the Holonic Architectural Framework

“…Determining an optimum routing and clustering organization needs adaptation depending on the situation and context [4]. Therefore, this closes the feedback loop to ensure that knowledge from the environment is retained, so predictive models can determine future clustering and routing paths can better model system state behavior.…”