Latency-Sensitive Task Allocation for Fog-Based Vehicular Crowdsensing

Abstract

The development of the Internet of Vehicles has attracted a large number of innovative vehicular applications. Specifically, vehicular crowdsensing represents an emerging paradigm that utilizes on-board sensing and computing capacity to provide location-based services. The conventional vehicular crowdsensing method aims to improve the quality of service or reduce the cost while guaranteeing coverage and data quality. Inappropriate sensing vehicle selection may cause high execution latency, which, however, has rarely been considered. In this article, we propose a fog-based vehicular crowdsensing scheme, where on-board sensors can act as sensing participants and vehicles can act as fog nodes. It unifies crowdsensing participant selection and computing task allocation to jointly optimize execution latency and cost. Furthermore, we present a heuristic algorithm named as random chemical reaction optimization (RCRO) to solve the proposed problem. Finally, we evaluate the proposed scheme and algorithm by simulating real-world traffic. Numerical results show the performance of the RCRO algorithm and the superiority of the proposed method in comparison with the conventional method.