An in-depth analysis of the impact of battery usage patterns on performance of task allocation algorithms in sparse mobile crowdsensing

Abstract

Mobile Crowdsensing leverages the sensing capabilities of multiple mobile devices to execute large-scale sensing tasks by breaking them into smaller tasks for execution on individual mobile devices. Task allocation algorithms are used to efficiently distribute these smaller sensing tasks to a subset of participants while optimizing system-level goals (such as location accuracy or data quality) for participant selection. The sensing tasks, e.g., collecting GPS tagged data, are often energy-intensive and battery consumption during sensing task execution remains a major concern for participants. So far no in-depth study exists that evaluates the impact of battery consumption on allocation algorithms. In this work, we conducted an in-depth study on the effects of battery consumption patterns of smartphone users. We studied the impact of battery consumption patterns extracted from a real-world data-set on standard as well as state-of-the-art algorithms to show how different battery usage patterns affect the performance of allocation algorithms. Our work provides an important insight into factors affecting the performance of allocation algorithms and advocates incorporating battery usage patterns for the future development of these algorithms.