In Internet of Things (IoT), data are collected from various sources (e.g., sensors and databases) to describe a phenomenon under-observation. Due to the lack of knowledge about the measurement environment and the limited accuracy of data sources, the IoT data inevitably appears uncertain, imperfect, and conflicting. This would lead to high data conflict among different data sources, and the final result of data fusion is degraded. This paper presents an edge-based data conflict resolution approach for IoT to tackle these issues while maintaining an accurate and reliable result. In the proposed approach, data conflict is significantly alleviated using two phases. The first is Faulty Data Detection and Correction (FDDC), which runs based on the confidence interval and estimated data. The second is Data Conflict Measure and Fusion (DCMF), which is developed to compute the degree of conflict among different data sources based on Fuzzy measures and calculate each data source’s credibility degree. Then, data fusion is carried out based on the Minimum Mean Square Error (MMSE) criterion accordingly. The simulation results reveal the remarkable effectiveness and efficiency of the proposed approach for solving IoT conflicting data.

在物联网（IoT）中，从各种来源（例如传感器和数据库）收集数据，以描述观测不足的现象。由于缺乏有关测量环境的知识以及数据源的准确性有限，IoT数据不可避免地显得不确定，不完善和冲突。这将导致不同数据源之间的高数据冲突，并且数据融合的最终结果会降低。本文为物联网提出了一种基于边缘的数据冲突解决方法，以解决这些问题，同时保持准确而可靠的结果。在提出的方法中，使用两个阶段可以显着缓解数据冲突。第一个是故障数据检测和纠正（FDDC），它基于置信区间和估计数据运行。第二个是数据冲突度量和融合（DCMF），开发该工具是基于模糊测度来计算不同数据源之间的冲突程度，并计算每个数据源的可信度。然后，根据最小均方误差（MMSE）准则进行数据融合。仿真结果表明，该方法解决物联网冲突数据具有显着的有效性和效率。