Exploring Reinforcement Learning Environment for User-centric Applications in VANET
V Padmapriya* and D N Sujatha
Published: November 19, 2022
DOI: 10.55162/MCET.03.092
Abstract
The past decade has identified Vehicular Ad Hoc Networks (VANET) as one of the promising technology for intelligent communication among vehicles. VANET supports the dissemination of safety, warning, and infotainment messages. Today, there is a growing demand for location-specific and infotainment messages among urban travelers. The class of applications that disseminates infotainment messages are called user centric applications. The task of dissemination is expected to be remunerative to promote cooperation among the vehicles. A vehicle that disseminates these user-centric messages earns a reward in the form of an incentive. Generally, the incentive bring in greed with a threat of malicious behavior in the network. Previously, several incentive-based approaches have been proposed that handle malicious behavior and maintain the equilibrium of rewards with the perspective of the Vehicular Network (VN). However, the Reinforcement Learning (RL) paradigm with its intelligent algorithms combined with vehicular networks is capacitated to handle several challenges in the incentive-based approaches. In this paper, we explore how RL environments can be adopted for the rewarding techniques in VANET. The paper concludes with open research challenges in this area.
Keywords: Applications; Incentive-based; Reinforced Learning; Rewards; User-centric; Vehicular Networks