Reduce Fingerprint Construction for Positioning IoT Devices Based on Generative Adversarial Nets

Fingerprint-based positioning is popular and applicable for Internet of Things (IoT) applications to offer seamless, intelligent and adaptive location-aware services for IoT devices. However, it takes time and cost to build the radio-map. This paper proposed deep convolutional Generative adversarial nets (DCGANs) to minimize the site survey time and cost, and to mitigate signal fluctuations. The radio-map was designed for receiving radio signals from detectable wireless local area network (WLAN) and cellular networks in scalable environments. The proposed fingerprinting-based positioning is a sequential combination of the hybrid support vector machine and long short-term memory algorithms. The experimental results indicate that the proposed method achieves a promising and reasonable positioning performance for IoT devices in scalable wireless environments.