Fault Classification and Blockchain-Based Incentive Scheme for Smart Wireless Sensor Networks

Abstract

Due to the implementation in uncertain and dangerous environments, wireless sensor networks (WSN) are prone to software, equipment, and system failures. At the sensor level, classifiers are used to classify hardover, spike, drift, data loss, erratic, random, and stuck faults. The data quality from the collected information is determined by the level of participation from all crowd sensor networks (CSN) entities, including service customers, service provider, and data collectors. For CSNs, we propose a blockchain-based incentive methodology. The incentives are being used to entice data collectors to join the network as well as participants. To prevent privacy leakage, advanced encryption standard (AES128) technique is used. Accuracy, precision, F₁ score are used to compare the results of the first situation. In this work, simulations show that the extremely randomized trees (ERT) algorithm achieves a higher rate of fault detection. The performance is evaluated by comparing the execution times of all of the consensus mechanisms, while the encryption process is affirmed by correlating the execution times. The incentive system is examined by computing the input string's gas cost and mining time.