Wavelet Transform DC-GAN for Diversity Promoted Fingerprint Construction in Indoor Localization

Abstract

Wi-Fi positioning is currently the mainstream indoor localization method, and the construction of fingerprint database is crucial to the Wi-Fi based localization system. However, the accuracy requirement needs enough data sampled at many reference points, which consumes significant manpower and time. In this paper, we convert the acquired Channel State Information (CSI) data to feature maps using complex wavelet transform and then extend the fingerprint database by the proposed Wavelet Transform-Feature Deep Convolutional Generative Adversarial Network model. With this model, the convergence process in training phase can be accelerated and the diversity of generated feature maps can be increased significantly. Based on the extended fingerprint database, the accuracy of indoor localization system can be improved with reduced human effort.