When Crowdsourcing Meets Unmanned Vehicles: Toward Cost-Effective Collaborative Urban Sensing via Deep Reinforcement Learning

Ding, Lige; Zhang, Dong; Cao, Mingzhe; Ma, Huadóng

doi:10.1109/jiot.2021.3062569

Cited by 10 publications

(1 citation statement)

References 41 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…its locations, called task assignment [3][4][5][6]. Some studies on spatial crowdsourcing usually concentrate on the problem of task assignment [7][8][9][10][11], which is to allocate tasks to unmanned vehicles, and they presume all tasks are simple and easy.…”

mentioning

confidence: 99%

Diversity-aware unmanned vehicle team arrangement in mobile crowdsourcing

Liu

Feng

Peng

et al. 2022

J Wireless Com Network

View full text Add to dashboard Cite

With the continuous development of mobile edge computing and the improvement of unmanned vehicle technology, unmanned vehicle could handle ever-increasing demands. As a significant application of unmanned vehicle, spatial crowdsourcing will provide an important application scenario, which is about to organize a lot of unmanned vehicle to conduct the spatial tasks by physically moving to its locations, called task assignment. Previous works usually focus on assigning a spatial task to one single vehicle or a group of vehicles. Few of them consider that vehicle team diversity is essential to collaborative work. Collaborative work is benefits from organizing teams with various backgrounds vehicles. In this paper, we consider a spatial crowdsourcing scenario. Each vehicle has a set of skills and a property. The property denotes vehicle’s special attribute (e.g., size, speed or weight). We introduce a concept of entropy to measure vehicle team diversity. Each spatial task (e.g., delivering the take-out, and carrying freight) is under the time and budget constraint, and required a set of skills. We need to assure that the assigned vehicle team is diverse. To address this issue, we first propose a practical problem, called team diversity spatial crowdsourcing (TD-SC) problem which finds an optimal team-and-task assignment strategy. Moreover, we design a framework which includes a greedy with diversity (GD) algorithm and a divide-and-conquer (D&C) algorithm to get team-and-task assignments. Finally, we demonstrate efficiency and effectiveness of the proposed methods through extensive experiments.

show abstract