Multitask-Oriented Collaborative Crowdsensing Based on Reinforcement Learning and Blockchain for Intelligent Transportation System

Mengge, Li; Ma, Miao; Wang, Liang; Yang, Bo; Wang, Tao; Sun, Jinping

doi:10.1109/tii.2022.3228935

Cited by 6 publications

(1 citation statement)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…By making full use of the random mobility of mobile users, MCS allocates tasks to well-suited users, which can enhance the flexibility of ubiquitous sensing and ensure high spatiotemporal coverage. This appealing sensing paradigm, which can effectively achieve urban-scale monitoring, has expanded the scope of the IoT and has been widely used in many IoT applications, such as urban sensing [ 10 , 11 ], intelligent transportation [ 12 , 13 , 14 ], and environmental monitoring [ 15 , 16 ].…”

Section: Introductionmentioning

confidence: 99%

UAV-Assisted Cluster-Based Task Allocation for Mobile Crowdsensing in a Space–Air–Ground–Sea Integrated Network

Liu,

Li,

Cheng

et al. 2023

Sensors

View full text Add to dashboard Cite

Mobile crowdsensing (MCS), which is a grassroots sensing paradigm that utilizes the idea of crowdsourcing, has attracted the attention of academics. More and more researchers have devoted themselves to adopting MCS in space–air–ground–sea integrated networks (SAGSINs). Given the dynamics of the environmental conditions in SAGSINs and the uncertainty of the sensing capabilities of mobile people, the quality and coverage of the sensed data change periodically. To address this issue, we propose a novel UAV-assisted cluster-based task allocation (UCTA) algorithm for MCS in SAGSINs in a two-stage process. We first introduce the edge nodes and establish a three-layer hierarchical system with UAV-assistance, called “Platform–Edge Cluster–Participants”. Moreover, an edge-aided attribute-based cluster algorithm is designed, aiming at organizing tasks into clusters, which significantly diminishes both the communication overhead and computational complexity while enhancing the efficiency of task allocation. Subsequently, a greedy selection algorithm is proposed to select the final combination that performs the sensing task in each cluster. Extensive simulations are conducted comparing the developed algorithm with the other three benchmark algorithms, and the experimental results unequivocally endorse the superiority of our proposed UCTA algorithm.

show abstract