Spatial Statistics-based Participant Selection for Optimised Mobile Crowd-Sensing

Mobile Crowd-Sensing (MCS) is emerging as an important technology for future wireless networks and associated services like Mapping, Localization, Weather prediction, Pollution monitoring, etc. It is also being investigated for its potential in enabling the creation of Digital Twins of networks and cities as well. In MCS, few networked devices sense the parameter of interest and report it to a central entity, which utilizes this data to generate usable intelligence. For the MCS-based system to be reliable, it is imperative that the number and placement of the sensors are appropriate. If the number is less than adequate, the derived intelligence is not reliable. If the number is more, the additional data results in higher reporting and processing overheads. So, in order to realize the full potential of MCS without increasing the computational burden, in this work, a spatial statistics-based participant sensor selection scheme has been proposed. Radio Environment Mapping has been presented as an application of MCS. The number of participants is arrived at on the basis of the spatial correlation function. Then the participants are selected on the basis of the distance from the regular lattice centers. The results indicate a significant increase in accuracy in REM creation by the proposed method.