Several location-based services require accurate location information in indoor environments. Recently, it has been shown that deep neural network (DNN) based received signal strength indicator (RSSI) fingerprints achieve high localization performance with low online complexity. However, such methods require a very large amount of training data, in order to properly design and optimize the DNN model, which makes the data collection very costly. In this paper, we propose generative adversarial networks for RSSI data augmentation which generate fake RSSI data based on a small set of real collected labeled data. The developed model utilizes semi-supervised learning in order to predict the pseudo-labels of the generated RSSIs. A proper selection of the generated data is proposed in order to cover the entire considered indoor environment, and to reduce the data generation error by only selecting the most realistic fake RSSIs. Extensive numerical experiments show that the proposed data augmentation and selection scheme leads to a localization accuracy improvement of 21.69% for simulated data and 15.36% for experimental data.INDEX TERMS Indoor localization, received signal strength indicator (RSSI), deep neural network (DNN), generative adversarial network (GAN), semi-supervised learning.

Currently, indoor localization is among the most challenging issues related to the Internet of Things (IoT). Most of the state-of-the-art indoor localization solutions require a high computational complexity to achieve a satisfying localization accuracy and do not meet the memory limitations of IoT devices. In this paper, we develop a localization framework that shifts the online prediction complexity to an offline preprocessing step, based on Convolutional Neural Networks (CNN). Motivated by the outstanding performance of such networks in the image classification field, the indoor localization problem is formulated as 3D radio image-based region recognition. It aims to localize a sensor node accurately by determining its location region. 3D radio images are constructed based on Received Signal Strength Indicator (RSSI) fingerprints. The simulation results justify the choice of the different parameters, optimization algorithms, and model architectures used. Considering the trade-off between localization accuracy and computational complexity, our proposed method outperforms other popular approaches.