Incentivizing Point Cloud-Based Accurate Cooperative Perception for Connected Vehicles

Abstract

In this paper, we propose a Cooperative Perception framework with an Incentive Mechanism (CPIM) to tackle the lack of incentive mechanisms of cooperative perception among connected vehicles. This mechanism comprises three key components: point cloud area value assessment, bargaining mechanism, and greedy algorithm-based selection mechanism. The point cloud area value assessment quantifies the significance of point clouds in specific spatial areas, contributing to informed decision-making during the subsequent bargaining process. The bargaining mechanism uses a bargaining model to determine a price for each area that satisfies all participating vehicles. The selection mechanism enables vehicles to quickly make near-optimal purchasing decisions by greedily selecting those with the highest marginal utility per price in each area. The core innovation of the proposed CPIM lies in the unique design of point cloud area value assessment and vehicle bargaining components, which are specially crafted for rapid response and low bandwidth consumption in cooperative perception. Furthermore, the selection mechanism enables fast and accurate execution of the selection process. Extensive experiments demonstrate that compared to other cooperative perception frameworks, our approach achieves up to 89.03% reduction in bandwidth resource consumption without compromising accuracy. The proposed CPIM framework offers a promising solution to enhance both the effectiveness and efficiency of cooperative perception in vehicular networks.