Roadside IRS Assisted Task Offloading in Vehicular Edge Computing Network

Vehicular edge computing (VEC) has been recognized as a promising technique to process delay-sensitive vehicular applications. Nevertheless, in order to accommodate the rapid growth in the number of connected vehicles, it’s inevitable that there will be an increasing deployment of conventional infrastructure with limited communication ranges. This could potentially lead to escalating costs and impede the full realization of the VEC system. In this paper, a roadside intelligent reflecting surface (IRS) assisted VEC network is introduced, where the IRS is deployed outside the coverage of roadside units (RSUs) to extend the service range. Furthermore, the maximum total number of successful offloading tasks problem within the scheduling time problem is formulated, encompassing the optimization of offloading decisions, computation resource allocation and phase shift of IRS. To tackle the formulated challenging problem, we first decouple the original problem into two subproblems. Then, a heuristic algorithm is proposed, where a many-to-one matching algorithm is proposed to joint optimize offloading decision and the computation resource, and an iterative algorithm is utilized to optimize the phase shift coefficients of IRS. The simulation results validate the effectiveness of the proposed algorithm in comparison to other schemes, and the IRS can effectively maintain network performance even when there are intervals in RSU coverage areas.