A Hybrid Fuzzy-Genetic Algorithm for Performance Optimization of Cyber Physical Wireless Body Area Networks

The use of fuzzy decision-making in datapath selection extends the sensor network lifetime with a uniform distribution of routing load among network nodes. Fuzzy-logic based routing protocols are mostly designed for general wireless sensor networks (WSN). However, such protocols are not compatible with a Wireless Body Area Network (WBAN) comprised of biosensor nodes. WBAN nodes carry inferior computational, communication and energy resources as compared to general WSN nodes. A WBAN routing protocol needs to be designed as per IEEE 802.15.6 WBAN standards to meet high-end QoS requirements of medical applications. This paper presents a fuzzy-logic-based clustering protocol for data routing in WBANs. Nodes are grouped into clusters and cluster head nodes are selected through a Fuzzy-Genetic Algorithm termed as EB-f(g)-MADM. EB-f(g)-MADM makes an assessment of dual attributes of each cluster node in terms of node residual energy and CH selection cost. CH selection cost of a node is the forecasted value of network energy consumption if the node acts as a cluster head. EB-f(g)-MADM utilizes a fuzzy-TOPSIS function which makes a quantitative comparison of cluster nodes and selects the cluster head node possessing the aforementioned attributes closest to their ideally desired values. A Genetic Algorithm-based optimization process adapts the attribute weights for cluster head selection. EB-f(g)-MADM provides enhanced network lifetime with a uniform distribution of routing load. Protocol performance is obtained in terms of network lifetime, throughput and latency. Results are compared with existing WBAN routing protocols and are found to be better.