A new perceptron based neural-network algorithm to enhance the scheduling performance of safety-critical WSNs of increased dependability

Wireless Sensor Networks (WSNs) are embedded systems consisting of multiple distributed Sensor Nodes and usually one or more Base Stations (BSs), placed within an area of interest, to monitor and detect given behaviours and changes. Nowadays, WSNs are widely used in Safety-Critical systems where their dependability requirements are determined by the correct operations of the three primary properties: Connectivity, Coverage, and Lifetime of the network. These properties have been mostly addressed independently of each other due to the complexity of addressing them simultaneously within a networked environment. This paper proposes a Perceptron-based Artificial Neural Network (ANN) analyser to analyse the performance of the Scheduling algorithms (e.g., where nodes alternate between awake (ON) and sleep (OFF) states) in WSNs using a MATLAB simulation environment. This approach uses a neural network to learn, train, and test the performance of such algorithms, to better the overall dependability of the network. The simulation results show possible ways to improve the lifetime of nodes using a more dynamic connectivity /coverage strategy. In particular, nodes that switched ON more than four times were identified and classified. This has the benefit of improving network lifetime and hence its service availability and reliability attributes (dependability) by balancing the workload or the sleep-schedule of those nodes, the network’s lifetime increases by avoiding unnecessarily depleting nodes’ energy.

This version of the article has been accepted for publication, and is subject to Springer Nature’s AM terms of use (https://www.springernature.com/gp/open-research/policies/accepted-ma...), but is not the Version of Record and does not reflect post-acceptance improvements, or any corrections. The Version of Record is available online at: https://link.springer.com/book/9789819986118