RSS Remeasurement Estimation for Indoor Positioning System with Generative Adversarial Network Model

With the rapid development of computer technology in recent years, the demand for services to obtain location information is increasing day by day. The expansion of mobile computing has prompted wireless fidelity(WiFi)-based indoor localization to be one of the most attractive and promising techniques for ubiquitous application. The main problem with this technology is that received signal strength (RSS) values need to be manually sampled to create a WiFi fingerprint database in the offline database creation stage. As any change in the environment may affect signal propagation, the fingerprint database must be repeatedly updated and maintained, and this is a time-consuming task. In this study, we propose the Adaptive Context Generative Adversarial Networks (ACOGAN) model by adopting random sampling RSS values as the feedbacks and predicting the residual RSS values as the missing fingerprints to solve the remeasurement problem. By employing feedbacks, the missing fingerprints can be rebuilt efficiently and accurately using the ACOGAN model. The experimental results not only illustrate the superiority of the proposed ACOGAN model on field and simulated environments but also demonstrate its robustness to satisfy the time consumption on the learning progress.