Joint Offloading and Resource Allocation for Satellite Assisted Vehicle-to-Vehicle Communication

• Published in: IEEE systems journal Vol. 15; no. 3; pp. 3958 - 3969
• Main Authors: Cui, Gaofeng, Long, Yating, Xu, Lexi, Wang, Weidong
• Format: Journal Article
• Language: English
• Published: New York IEEE 01.09.2021; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Communication; Communication satellites; Computation offloading; Deep reinforcement learning (DRL); Edge computing; Lagrange multiplier; Markov processes; Mixed integer; mobile edge computing; Nonlinear programming; Processor scheduling; Resource allocation; Resource management; Satellite broadcasting; satellite communication; Satellites; Task analysis; vehicle-to-vehicle (V2V); Vehicles; Voltage control
• ISSN: 1932-8184; 1937-9234
• DOI: 10.1109/JSYST.2020.3017710

Satellite assisted vehicle-to-vehicle (V2V) communication can provide services for vehicles in depopulated areas, and it can be employed as an effective complementary component for terrestrial vehicular networks. Since the available communication and computing resource for satellites are scarce, task offloading, computing and communication resource allocation, which are coupled with each other, are critical issues for satellite assisted V2V communication. To tackle these problems, we formulate the joint offloading decision, computing and communication resource allocation problem for satellite assisted V2V communication as a mixed-integer nonlinear programming problem with minimum weighted-sum end-to-end latency, and we decouple it into two subproblems. First, the Lagrange multiplier method is adopted to obtain the optimal computing and communication resource allocation with fixed offloading decision. Then, the results of the resource allocation subproblem are fed into the offloading decision problem, which is formulated as a Markov decision process. To maximize the long-term reward of offloading decision, a deep reinforcement learning based method is adopted to learn the optimal offloading decision. Finally, the simulation results show that the proposed joint task offloading and resource allocation approach has superior performance compared with other schemes.