Towards accurate and privacy-preserving localization using anchor quality assessment in Internet of Things

Abstract

Crowdsourced localization plays a significant role for the applications in Internet of Things. Even though existing studies have proposed privacy-preserving localization algorithms to protect location information from being disclosed to untrusted parties, they still need to consider improving the accuracy of the algorithms so as to enable practical localization services. In this paper, by extending the research scope of previous literature, we investigate a novel localization problem of achieving both accuracy improvements and privacy preservation from the perspective of anchor quality assessments. Specifically, the quality of anchor (QoA) is proposed to quantitatively describe the capability of anchor users to provide accurate data by taking into account factors such as residual energy, localization distance, and localization angle. Based on efficient and privacy-preserving maximum and minimum computation techniques, a privacy-preserving QoA assessment method is proposed. Besides, an accurate and privacy-preserving localization (APPL) algorithm is developed by using the privacy-preserving assessment method. The APPL algorithm selects anchor users who can provide accurate data to participate in localization based on the results of the QoA assessment. It is proved that the APPL algorithm is correct, accurate, privacy-preserving and efficient by theoretical analyses. Both experiments and simulations are conducted to evaluate the performance of the APPL algorithm and validate the analytical results.