ANFIS-Based Resource Mapping for Query Processing in Wireless Multimedia Sensor Networks

Abstract

AbstractWireless multimedia sensor networks (WMSNs) are usually resource constrained, and where the sensor nodes have limited bandwidth, energy, processing power, and memory. Hence, resource mapping is required in a WMSN, which is based on user linguistic quality of service (QoS) requirements and available resources to offer better communication services. This paper proposes an adaptive neuro fuzzy inference system (ANFIS)-based resource mapping for video communications in WMSNs. Each sensor node is equipped with ANFIS, which employs three inputs (user QoS request, available node energy, and available node bandwidth) to predict the quality of the video output in terms of varying number of frames/second with either fixed or varying resolution. The sensor nodes periodically measure the available node energy and also the bandwidth. The spatial query processing in the proposed resource mapping works as follows. (i) The sink node receives the user query for some event. (ii) The sink node sends the query through an intermediate sensor node(s) and cluster head(s) in the path to an event node. A cluster head-based tree routing algorithm is used for routing. (iii) The query passes through ANFIS of intermediate sensor nodes and cluster heads, where each node predicts the quality of the video output. (iv) The event node chooses the minimum quality among all cluster heads and intermediate nodes in the path and transmits the video output. The work is simulated in different network scenarios to test the performance in terms of predicted frames/second and frame format. To the best of our knowledge, the proposed resource mapping is the first work in the area of sensor networks. The trained ANFIS predicts the output video quality in terms of number of frames/second (or H.264 video format) accurately for the given input.