Utilizing Artificial Intelligence to Design Delay and Energy-Aware Wireless Sensor Networks
Ranjana Thalore, Vandita Vyas, Jeetu Sharma, Vikas Raina · 2021
There is an arduous pursuit for a protocol design that is energy-efficient for Wireless Sensor Networks (WSNs) due to the limited battery capacity of nodes. The development of new protocols is taken as a step to achieve energy efficiency in the communication stack. Architectural repressions, energy exhaustion, error lenience, channel, and network constitution are the design constraints of WSNs. Artificial Intelligence mechanisms can be implemented in WSNs by two methods: The designers have the global objective as well as design in mind. The designer considers and creates a set of self-interested agents who evolve and interact in a stable manner, in their structure, through evolutionary techniques for learning. This chapter presents various approaches to design delay and energy aware WSNs. The QualNet version 6.1 network simulator is used for simulations of various models. These are done to evaluate performance constraints like network lifetime, end-to-end delay, and throughput. The assessment precedent for LR-PAN realized in 3D and 2D terrains are the valuation of various QoS parameters. ML-MAC protocol is designed to analyze a randomly deployed network in terms of design constraints like the conservation of energy and enhancement of lifetime. A practical deployment method of sensor nodes incorporating 3D is presented in the chapter and two different types of networks—homogeneous and heterogeneous WSNs—are analyzed through extensive simulations. Long-distance transmission is expensive in WSNs since power consumption is proportional to dα for transmitting over distance d, where α is the path loss exponent that varies between 2-6. There is a requirement for relay nodes in the network that can perform data sensing, data aggregation, and data routing. A Layered Relay-Medium Access Control (LR-MAC) Protocol introduces the concept of multi-layering for the relay (routing) nodes. The QoS parameters are analyzed for three different deployment strategies viz. random, grid, and circular.