The fundamental problem of the project is how to define and assess the quality of LoT mobility data, in the presence of the unique characteristics of LoT. To this end, we conducted an in-depth analysis of the data quality dimensions of the mobility data and the factors that affect the data quality dimensions in the context of Location of Things (LoT). Moreover, we performed an extensive study of the state-of-the-art techniques for mobility data quality management and low-quality mobility data exploitation, and based on that, we identified new opportunities for quality-aware data management and exploitation in the challenging LoT setting. Finally, we summarized the means to mitigate mobility data quality issues and proposed an integral framework for mobility data quality management in the LoT.
Within the proposed framework, we studied data quality modeling techniques that can be adaptative to the decentralized, dynamic, and heterogeneous computing environment. Specifically, taking the data statistics (quantile computation) task as a case study, we analyzed and modeled the relationship between data errors and processing latencies among a set of decentralized and heterogeneous computing nodes. In another task of positioning data cleaning, we studied and proposed a generalized model for capturing and mitigating the uncertainty of mobility data (positioning data) in LoT.
On top of the proposed data quality modeling techniques, we studied specific data quality enhancement algorithms. On the one hand, we studied missing data imputation for wireless positioning data, considering modeling the spatiotemporal dependencies within a mobility data sequence and between multiple mobility data sequences. On the other hand, we studied decentralized, continuous proximity-based outlier detection in an edge computing fashion.
To measure and improve the efficiency of the planning of data quality tasks in LoT, we proposed an edge-resident task coordination mechanism for optimizing resource usage. Such a mechanism has proven effective in the application of spatiotemporal quantile monitoring. As a facilitator work, we also implemented a testbed for deploying and evaluating data quality management algorithms in the decentralized, dynamic, and heterogeneous LoT environment.
To disseminate the project results in the cyber world, we have created a project website (
http://msca-malot.github.io/(odnośnik otworzy się w nowym oknie)) a GitHub organization (
https://github.com/msca-malot(odnośnik otworzy się w nowym oknie)) and a Twitter social media account (
https://twitter.com/MalotMsca(odnośnik otworzy się w nowym oknie)). These portals will continue to be used to expand the impact of the project's current and subsequent outcomes. Partially project outcomes have also been presented in academic and networking events, including top-tier conferences VLDB 2022 and SIGMOD 2022, invited talks at HUST, SUSTech, BUPT, etc., and seminars at AAU and RUC.
