The Role of Human Relations and Interactions in Designing Memory-Related Models for Sensor Networks

Recently, the use of Wireless Sensor Networks has become substantial in most of our life aspects. These networks have many issues and challenges at the design phase (e.g., memory and power consumption). There exists a huge amount of works and studies that offer and provide solutions for many of these challenges. However, the issues of predicting memory requirements and memory management have not received enough attention in sensor networks literature. Yet, most of the studies in this field focus on issues related to power consumption and connectivity of sensor nodes. This paper has two main purposes: first, we propose a metric for measuring the strength of a relation between two sensors. In the proposed metric, we involve three important characteristics of human relations and interactions: encounter frequencies, duration of encounters, and regularities of encounters. We then exploit this metric in predicting memory requirements in a sensor network. Second, based on the estimated memory size, we propose an approach for memory management in a sensor network. The proposed approach is based on two concepts: social capital in sociology and preferential return in human mobility. The results show that our approach is effective in managing sensor memories comparing to other approaches in the literature.