LAUSR

Crowdsensing refers to an approach for collecting of data from a large number of smart devices and sensors carried by many individuals and has been employed for numerous applications, which include pollution monitoring, traffic monitoring and noise sensing. It is an important mechanism for building applications in the smart environments enabled by the internet-of-things. However, often a given problem may dictate that samples are drawn from a defined sub-population of participants, for example based on characteristics of the participant such as location, demographics or other profile attribute, rather than from any possible member of the whole population. In this article we introduce an approach for crowdsensing with a consideration for how to sample from specific sub-populations in a region, delineated in a dimension-based way analogous to the multi-dimensional data model used in data warehousing. Simulation and performance results are provided demonstrating the approach’s ability to maintain active participants, provide coverage of the region of interest, and to be able to scalably sample the variable of interest in relation to the sub-population. This is the first work to our knowledge to address and propose an approach to the specific problem of crowdsourcing from specific attribute-defined sub-populations.