Mobile Crowdsensing

Load-Balanced Data Collection in Mobile Crowdsensing Systems

The efficient operation of mobile crowdsensing systems involves the non-trivial issue of distributing work loads among mobile and heterogeneous devices to solve crowdsensing tasks. These loads should be balanced to optimize short and long-term system performance with respect to costs, quality of results, user satisfaction and further metrics specific to a certain application.

Consider the following example:

We seek to create an up-to-date tempo-spatial noise level map of an area. We are given a tessellation of this area and a group of crowd participants equipped with smart phones is scattered over the area. Using the external microphones of their smart phones, the crowd participants may provide sensed noise levels with location and time stamp. The question we would like to answer is: How can we cover the whole area while minimizing the amount of time a single participant is involved?

We adapt and combine knowledge from resource allocation, scheduling, online optimization, big data analytics, parallel computing, and game theory to provide answers to this question.

A major part of our research and experiments is to test the performance of load balancing algorithms for mobile crowdsensing systems.