CrowdRep: A Blockchain-based Reputation System for Crowdsourced HD Map Update.

Benny Wijaya,Kun Jiang,Mengmeng Yang,Tuopu Wen,Xuewei Tang,Diange Yang,Yanhai Ma,Ricky Albert
DOI: https://doi.org/10.1109/itsc55140.2022.9922458
2022-01-01
Abstract:Crowdsourced vehicles are expected to be the answer to an efficient yet economic solution for maintaining an HD (High-Definition) map for autonomous vehicles. It is the building block that helps keep the autonomous vehicle safe and sound. However, in realizing this idea, there comes problem such as the reliability of the data provided by crowdsourced vehicles and malicious entities that can compromise map database security. In this paper, we propose a novel idea of implementing blockchain technology to solve the reliability problem and eliminate the effect of malicious intent from unknown entities. We explore the idea of using reputation to value the contribution given by crowdsourced vehicles and update it according to the consensus generated by the peer-to-peer members in the area of update. Essentially, these two key elements are combined to solve the problem faced by crowdsourced solutions for HD map updates. We create a numerical simulation to show that our method can group vehicles into categories according to the quality of data it provides. Furthermore, we show that the system can handle a significant amount of attackers (not a majority) with only a slight decrease in the map element localization accuracy.
What problem does this paper attempt to address?