Sensing Quality-Aware Task Allocation for Multidimensional Vehicular Urban Sensing

Abstract

Vehicular sensing has become attracting an increasing research interest for cost-effective monitoring in urban areas. Even though multiple types of sensing data are required to form a multi-dimensional sensing map in urban sensing applications, most of previous works have only considered the sensing quality of single sensor type. In this paper, we formulate an optimization problem of task allocation to improve the overall sensing quality in multi-dimensional vehicular urban sensing. To mitigate high complexity of the formulated problem, we prove the submodularity of the objective function and present a low-complexity heuristic algorithm called sensing quality-aware task allocation (SQTA) leveraging the property of submodular optimization. Extensive experiments have been conducted by using two real-world datasets, which demonstrate that SQTA can improve the average sensing quality of multiple sensor types and also guarantee sufficient levels of the sensing quality of all sensor types.