Wireless sensor networks and machine learning centric resource management schemes: A survey

Wireless Sensor Network (WSN) is a heterogeneous, distributed network composed of tiny cognitive, autonomous sensor nodes integrated with processor, sensors, transceivers, and software. WSNs offer much to the sensing world and are deployed in predefined geographical areas that are out of human interventions to perform multiple applications. Sensing, computing, and communication are the main functions of the sensor node. However, WSNs are mainly constrained by limited resources such as power, computational speed, memory, sensing capability, communication range, and bandwidth. WSNs when shared for multiple tasks and applications, resource management becomes a challenging task. Hence, effective utilization of available resources is a critical issue to prolong the life span of sensor network. Current research has explored various methods for resources management in WSNs, but most of these approaches are traditional and often fall short in addressing the resource management issues during real-time applications. Resource management schemes involves in resource identification, resource scheduling, resource allocation, resource utilization and monitoring, etc. This paper aims to fill the gap by reviewing and analysing the latest Computational Intelligence (CI) techniques, particularly Machine Learning (ML) and Artificial Intelligence (AI). AIML has been applied to countless humdrum and complex problems arising in WSN operation and resource management. AIML algorithms increase the efficiency of the network and speed up the computational time with optimized utilization of the available resources. Therefore, this is a timely perspective on the ramifications of machine learning algorithms for autonomous WSN establishment, operation, and resource management.