A neuro-fuzzy multi-topology adaptive routing framework for QoS-aware healthcare IoT communications

Healthcare IoT network loss their potential of reliability transmission, continuous monitoring, and intervention system due to critical network conditions along with its traffic characteristics. Thereby, it can directly affect the performance of the remote healthcare applications includes healthcare life quality check, patient safety, and clinical effectiveness in terms of QoS and robustness performance. In this paper presents a novel neuro-fuzzy multi-topology adaptive routing (NF-MTAR) method for enhancing reliable transmission of the healthcare IoT network. It integrates the neuro-fuzzy intelligence with multi-topology virtual partitioning enables dynamic optimization of the network resources based on network condition and its traffic characteristic. NF-MTAR method incorporates two unique innovations such as (i) neuro-fuzzy search engine identifies an optimum path to reach specific root node selected by incorporating five key parameters includes traffic flow intensity, resource utilization, residual energy, link quality, and node connectivity. (ii) Virtual Software Defined Networking (V-SDN) provides multi-topology virtual partitioning (elliptical, linear, and random) within the network, carry data transmission over multiple topologies for different traffic critical simultaneously. COOJA simulator is used to create three-layer 6LoWPAN architecture which capable of allowing dynamic network configuration and improve centralized policy management. Evaluation metrics are confirmed that the reasonable improvement is achieved such as high throughput (94.3%), reduce end to end delay (18.4ms), improve energy efficiency (31.2%), network lifetime (42.8%), and reliability (99.8%) by the proposed NF-MTAR method as compared to other state-of-art-methods. Thus, it provides potential improvement for next-generation medical monitoring and intervention system.