Resolving Wireless Sensor Networks Issues using Machine Learning Techniques: A Review

Abstract

Wireless Sensor Network (WSN) is the most encouraging advancements for some real-world applications due to its size, low cost, and easily deployable nature. Because of some factors like temperature, humidity, wind, pressure, battery life, etc., the performance of WSN will change dynamically, and therefore it requires depreciating dispensable redesign of the network. The conventional WSN is controlled by external programs, which makes it the networks hard to respond dynamically. In order to provide a quick response for dynamic changes, Machine learning (ML) techniques can be applied on WSN. Machine Learning is a technique that is self-instructed from experience and acts without human intervention or re-program. In this paper, Machine learning techniques for solving various issues in WSN are presented; we discussed machine learning techniques for anomaly, fault, and event detection. Finally, we presented the statistical analysis of Machine learning techniques used to solve the issues of WSN.