Route-Oriented Participants Recruitment in Collaborative Crowdsensing

Abstract

Route-oriented participants recruitment is a critical problem in collaborative crowdsensing, where task publisher uses monetary reward to motivate private cars collecting data along their routes. For map producers, route-oriented crowdsensing scheme helps them achieve maximum roads coverage with a limited budget, by selecting appropriate participants from a group of candidates. Focused on route-oriented participants recruitment problem, this paper first formalizes the road network and vehicle route model. Each vehicle’s route is mapped to a coverage rate on the road set. The recruitment problem therefore transforms to a combinatorial optimization problem, which has proved to be NP-hard. To find a solution, we proposed an approximation algorithm, which leverages submodularity to reduce computation complexity and has a worst performance guarantee. Finally we evaluate the performance of proposed algorithm on real road and trajectory data in Beijing, China.