Localization Reliability Improvement Using Deep Gaussian Process Regression Model

Abstract

With the widespread use of the Global Positioning System, indoor positioning technology has attracted increasing attention. Many systems with distinct deployment costs and positioning accuracies have been developed over the past decade for indoor positioning. The method that is based on received signal strength (RSS) is the most widely used. However, manually measuring RSS signal values to build a fingerprint database is costly and time-consuming, and it is impractical in a dynamic environment with a large positioning area. In this study, we propose an indoor positioning system that is based on the deep Gaussian process regression (DGPR) model. This model is a nonparametric model and it only needs to measure part of the reference points, thus reducing the time and cost required for data collection. The model converts the RSS values into four types of characterizing values as input data and then predicts the position coordinates using DGPR. Finally, after reinforcement learning, the position coordinates are optimized. The authors conducted several experiments on a simulated environment by MATLAB and physical environments at Tianjin University. The experiments examined different environments, different kernels, and positioning accuracy. The results showed that the proposed method could not only retain the positioning accuracy, but also save the computation time that is required for location estimation.

Recent advances in data-driven wireless communication using Gaussian processes: A comprehensive survey

“…Such interpretation strengths promote a large number of GP models to empower diversified wireless communication applications. There are five popular GP models using different kernels to support various wireless communication tasks, such as the GP models with stationary spectral mixture (SM) [24], [48] and compositional kernels [49] (see section IV-A), non-stationary (NS) kernels [19], [50], [51], [52], [53] (see section IV-B), deep kernels [54], [55] (see section IV-C), and multi-task kernels [19], [30] (see section IV-D). Furthermore, GPs have scalability variations with distributed inference to scale large data on a big number of edge devices (see section IV-E).…”

Distributed Machine Learning for Wireless Communication Networks: Techniques, Architectures, and Applications

“…The most popular DL models are neural networks (NNs), and only those NNs with enough hidden layers (usually, at least two layers) are considered "deep". Other multi-layer architectures, for instance, deep Gaussian process [34], neural process [35], and deep random forest [36], are regarded as DL structures. An advantage of DL compared to conventional ML is spontaneous feature extraction, through which hand-crafted efforts can be avoided [37].…”

Signal Fuse Learning Method With Dual Bands WiFi Signal Measurements in Indoor Positioning

“…In this experiment, we want to test the performance of location estimation in the physical field environment with several different methods, including our proposed method, KNN [36], WKNN [21], Cluster KNN [22], and DGPR proposed by Fei Teng [26]. In the traditional indoor position system, KNN, WKNN and Cluster KNN were proposed to categorize the fingerprint database into a small group, then a user can derive his position in this small group easily.…”

“…To solve the above problem, Fei Teng et al used a deep Gaussian regression model for indoor positioning. This model is a nonparametric model, and it only needs to measure part of the reference points, thus reducing the time and cost required for data collection [26]. However, according to our investigation, their researches all are based on the 2.4G signal band.…”