Using Synthetic Data to Enhance the Accuracy of Fingerprint-Based Localization: A Deep Learning Approach

Abstract

Human-centered data collection is typically costly and implicates issues of privacy. Various solutions have been proposed in the literature to reduce this cost, such as crowdsourced data collection, or the use of semi-supervised algorithms. However, semi-supervised algorithms require a source of unlabeled data, and crowd-sourcing methods require numbers of active participants. An alternative passive data collection modality is fingerprint-based localization. Such methods use received signal strength (RSS) or channel state information (CSI) in wireless sensor networks to localize users in indoor/outdoor environments. In this paper, we introduce a novel approach to reduce training data collection costs in fingerprint-based localization by using synthetic data. Generative adversarial networks (GANs) are used to learn the distribution of a limited sample of collected data and, following this, to produce synthetic data that can be used to augment the real collected data in order to increase overall positioning accuracy. Experimental results on a benchmark dataset show that by applying the proposed method and using a combination of 10% collected data and 90% synthetic data, we can obtain essentially similar positioning accuracy to that which would be obtained by using the full set of collected data. This means that by employing GAN-generated synthetic data, we can use 90% less real data, thereby reduce data-collection costs while achieving acceptable accuracy.

Joint Coordinate Optimization in Fingerprint-Based Indoor Positioning

“…On the other side, in ideal scenarios where the number of RPs and APs are enough (e.g., (150, 9, 15)), all PRAs have similar performance, particularly 2D-GPR and CGPR. This observation does not reduce the strength of the proposed 2D-GPR algorithm, because several studies on PRAs show that they are saturated by increasing the number of training data [1], [12].…”

“…In CGPR the aim is to obtain two functions that can map any RSS vector ∀s ∈ R ×1 to the 2D Cartesian coordinates as follows = (s) + , = (s) + and , ∼ N (0, ), (2) where and are pattern recognition functions that can convert RSS vectors to the and coordinates, respectively. The and should be optimized with the training dataset in (1). The optimization process of and are similar, and in the following, we only describe this procedure for .…”

“…S ATELLITE signals cannot penetrate in indoor environments, and due to the non-line-of-sight error, the accuracy of satellite-based positioning methods is not enough [1], [2]. The fingerprint-based wireless local area networks (WLANs) positioning is used in indoor environments to reach a high accuracy positioning [3].…”

Generative Adversarial Networks (GANs) in networking: A comprehensive survey & evaluation

“…The process of updating and is depicted in Algorithm 1 [6]. Convergence occurs when ( ; ) = 0.5, which means that discriminator is no longer able to distinguish between real and fake data.…”

A Systematic Review of Location Aware Schemes in the Internet of Things

“…Cloud computing with better prediction algorithms (process at cloud) can lead to accurate localization schemes while consuming minimal resources. In the literature, some authors proposed ML-based localization schemes [ 115 , 116 , 117 ] to improve the localization accuracy. By integrating ML with the localization, the progressive likelihood surpassed the posterior likelihood.…”