Joint Crowdsensing and Offloading Algorithms for Edge-Assisted Internet of Intelligent Vehicles

Abstract

In this paper, we aim to propose a new joint crowdsensing and offloading scheme that considers the benefits of social welfare. To induce the sensing participation, we adopt the ideas of cooperative multi-agent reinforcement learning (CMARL) to develop a novel crowdsensing algorithm. Due to the limitation of computation and communication resources in the IoIV system, the Lozano, Hinojosa, and Mármol solution (LHMS) is applied to solve the IoIV resource allocation problem. Our proposed crowdsensing and offloading algorithms are tightly coupled and work together to reach a consensus with reciprocal advantages. The main merits possessed by our hybrid approach are its flexibility and adaptability to current IoIV system situations. Performance evaluations on the proposed scheme show the superiority of our joint approach by comparing it with three existing baseline protocols.

Keywords

• Internet of Intelligent Vehicles
• vehicular crowdsensing
• vehicular offloading
• cooperative multi-agent reinforcement learning
• Lozano
• Hinojosa