A multi-objective indoor localization service for smartphones

The bulk of indoor localization applications currently rely on either server-side, cloud-based services that raise critical data-disclosure concerns (e.g., reveal user's location to a central entity), or client-side services that introduce serious performance concerns (e.g., consuming precious smartphone battery and network bandwidth during content uploads). In this paper, we present a novel Multi-objective Indoor Localization Service (MILoS) that provides a fine-grained, energy-efficient indoor localization using only a subset of WiFi-based localization data on the client-side, maintaining user's privacy at the same time. MILoS follows a fingerprinting-based indoor localization model that concurrently optimizes several conflicting objectives (i.e., minimizes the smartphone's energy consumption and maximizes the area coverage induced by WiFi fingerprints importance), using a Multi-Objective Evolutionary Algorithm based on Decomposition (MOEA/D). To the best of our knowledge this is the first time that the WiFi fingerprinting approach is used in a Multi-Objective Optimization setting for indoor localization. We assess our proposed model using real datasets and realistic mobility scenarios.