Improving QoS in Wireless Sensor Network routing using Machine Learning Techniques

Wireless sensor network (WSN) research is now extremely stimulated due to its potential applications in a range of disciplines, including area monitoring, healthcare, environmental observation, and industrial monitoring. The Quality of Service has become one of the main problems in WSN applications due to the increasing demand for WSN. Due to several limitations imposed by the applications using this network, guaranteed QoS in WSN is challenging to establish. Traditional QoS metrics concentrate on network-level metrics including packet reception ratio (PRR), jitter, end-to-end delay, and throughput. A high QoS environment is characterized by low packet delivery latency, high packet reception ratios, and maximum network throughput. The QoS can be assessed at the network or application level. In order to improve QoS in the network, this study focuses on creating and implementing a better path selection approach for WSN routing based on PRR predictions. Regression algorithms are used to forecast the PRR of a specific path, and the path with the best PRR value is selected to improve network quality of service. The strength of the received signal denoted as RSS, link quality indicator, noise floor over the specific multi-hop path, transmission and reception rate in the MAC layer, and routing path length are used to make the forecast. The results of the predictions and the estimated PRR are compared with the actual packet reception ratio collected from various WSN at an industrial environment.