TY - JOUR
T1 - Smart-3DM
T2 - Data-driven decision making using smart edge computing in hetero-crowdsensing environment
AU - Lamaazi, Hanane
AU - Mizouni, Rabeb
AU - Otrok, Hadi
AU - Singh, Shakti
AU - Damiani, Ernesto
N1 - Funding Information:
This work was supported by the Center for Cyber–Physical System, Khalifa University , through the Internal Research Fund, under Grant 8474000137 .
Publisher Copyright:
© 2022 Elsevier B.V.
PY - 2022/6
Y1 - 2022/6
N2 - Mobile Edge Computing (MEC) has recently emerged as a promising paradigm for Mobile Crowdsensing (MCS) environments. In a given Area of Interest (AoI), the sensing process is performed based on task requirements, which usually ask for a specific quality of the sensing outcome. In this work, a two-stage Data-Driven Decision-making Mechanism using smart edge computing (Smart-3DM) is proposed. It advocates the use of smart edge to better fulfill the data-related task requirements. Depending on the type of data to be collected, the minimum quality of the data required, and the heuristics to apply for each type of crowdsensing service, the smart edge orchestrates the selection of workers in MEC. Our approach relies on (a) smart-edge deployment: where a cluster-based distributed architecture using smart edge nodes is considered. Here, two entities are defined: the main edge node (MEN) and the local edge nodes (LENs); and (b) data management offloading where a two-layer re-selection strategy that considers data type and context-awareness is adopted, to reduce data computation complexity and to increase data quality while meeting the task target. The proposed Smart-3DM is evaluated using a real-life dataset and is compared to one-stage local and global approaches. The overall results show that by using two-stage re-selection strategies, better performance with lower processing power (CPU), less Storage(RAM), and improved execution time is achieved, when compared to the benchmarks.
AB - Mobile Edge Computing (MEC) has recently emerged as a promising paradigm for Mobile Crowdsensing (MCS) environments. In a given Area of Interest (AoI), the sensing process is performed based on task requirements, which usually ask for a specific quality of the sensing outcome. In this work, a two-stage Data-Driven Decision-making Mechanism using smart edge computing (Smart-3DM) is proposed. It advocates the use of smart edge to better fulfill the data-related task requirements. Depending on the type of data to be collected, the minimum quality of the data required, and the heuristics to apply for each type of crowdsensing service, the smart edge orchestrates the selection of workers in MEC. Our approach relies on (a) smart-edge deployment: where a cluster-based distributed architecture using smart edge nodes is considered. Here, two entities are defined: the main edge node (MEN) and the local edge nodes (LENs); and (b) data management offloading where a two-layer re-selection strategy that considers data type and context-awareness is adopted, to reduce data computation complexity and to increase data quality while meeting the task target. The proposed Smart-3DM is evaluated using a real-life dataset and is compared to one-stage local and global approaches. The overall results show that by using two-stage re-selection strategies, better performance with lower processing power (CPU), less Storage(RAM), and improved execution time is achieved, when compared to the benchmarks.
KW - Crowdsensing
KW - Data assessment
KW - Data quality
KW - Distributed architecture
KW - Smart edge computing
UR - https://www.scopus.com/pages/publications/85124157626
U2 - 10.1016/j.future.2022.01.014
DO - 10.1016/j.future.2022.01.014
M3 - Article
AN - SCOPUS:85124157626
SN - 0167-739X
VL - 131
SP - 151
EP - 165
JO - Future Generation Computer Systems
JF - Future Generation Computer Systems
ER -