I-OPC: An intelligent optimal path computation system using critical path prediction and deep learning for a time-sensitive network

Abstract

Latency and energy are critical issues when working with time and power-constrained wireless sensor networks. To avoid wasting both time and energy, such systems require selecting optimal communication routes with minimum latency and energy. The energy and latency costs between sender and destination nodes are greatly affected by the occurrence of transmission holes (black hole and grey hole). Therefore, selecting the optimal path must consider the probability of transmission holes and investigate their impact on energy and latency costs. Based on these problems of silent failures, the current work, proposes i-OPC as an intelligent and effective system to address these problems by forecasting sources of such silent failures and resolving them before they occur. The proposed method uses a customized routing schedule and a multi-objective mathematical optimization approach to rank all candidate paths between the source and destination nodes. In addition, i-OPC implements a machine learning technique using deep learning to predict the energy and latency costs of the future location for mobile nodes (). Also, it determines if the future locations of mobile nodes can result in black holes. Experiments produces promising results in terms of delay, energy consumption, packet delivery rate and hole detection rate against existing methods.