Research On Privacy-preserving Mobile Services Based On Secure Multi-party Computation

With the rapid development of wireless communication technology,the mobile Internet is gradually becoming the main way for people to surf the Internet.Mobile smart terminals represented by mobile phones and tablet computers are rapidly popularizing,and people’s living habits are also changing accordingly.Taking mobile localization and mobile crowdsourcing as an example,mobile localization can help people find parking spaces and restaurants;mobile crowdsourcing technology can use as many mobile terminals as possible to record various data quickly and in real time,providing powerful industrial system as support.However,if the mobile user’s information is inadvertently leaked,or even maliciously uploaded,malicious adversaries can even infer the participant’s location,health status and other personal information.Therefore,the privacy preserving scheme in the mobile environment should protect the privacy of mobile users.(1)Aiming at the privacy preserving problem of mobile localization,a privacy preserving system for indoor location service based on secure multi-party computation is proposed.To achieve the design goals of high accuracy and efficiency,the system takes advantage of the characteristics of WiFi signals and magnetic field signals,and delegates the calculation to edge servers.To protect WiFi fingerprint privacy,we propose a secure knearest neighbor algorithm.To protect the privacy of magnetic field fingerprints,a secure dynamic time warping algorithm is proposed.The system locates through two steps,first,the secure three-party WiFi fingerprint localization technology is used to obtain an approximate result,and then the three-party magnetic field fingerprint localization scheme is used to obtain a more precise position.At the same time,in order to realize the secure dynamic time warping algorithm,we designed a series of basic algorithms,such as secure branch algorithm,secure finding minimum algorithm and so on.We use edge servers to conduct experiments on public datasets.The results show that the efficiency of the privacy preserving scheme of WiFi fingerprint positioning is more efficient than the existing privacy preserving localization scheme,and the localization precision of the privacy preserving scheme of hybrid positioning is better than that of the existing schemes.(2)Aiming at the privacy preserving problem of task allocation in mobile crowdsourcing scenarios,a privacy-preserving crowdsourcing task allocation scheme based on secure multi-party computation is proposed.In order to protect the privacy of task attributes,we propose a secure logistic regression algorithm,which can assign tasks to participants according to their preferences;In order to protect the privacy of participant information and task spatiotemporal information,we propose a secure greedy algorithm to solve this NP-hard problem,where users can acquire multiple tasks in one round of interaction with a mobile crowdsourcing platform.Finally,we demonstrate through security analysis that the server can ensure that the privacy of participants and tasks is not leaked under an honest and curious threat model.The experimental evaluation on real data sets shows that this scheme is feasible and effective.