Gaussian Processes for Signal Strength-Based Location Estimation

By Brian Ferris, Dirk Hähnel, Dieter Fox. 2006

447 citations recorded by the source corpus.

Estimating the location of a mobile device or a robot from wireless signal strength has become an area of highly active research. The key problem in this context stems from the complexity of how signals propagate through space, especially in the presence of obstacles such as buildings, walls or people. In this paper we show how Gaussian processes can be used to generate a likelihood model for signal strength measurements. We also show how parameters of the model, such as signal noise and spatial correlation between measurements, can be learned from data via hyperparameter estimation. Experiments using WiFi indoor data and GSM cellphone connectivity demonstrate the superior performance of our approach.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Electrical and Electronic Engineering

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

In Electrical and Electronic Engineering

Citations in the corpus

Filtering with State-Observation Examples via Kernel Monte Carlo Filter (2013, methodology): “oblem by estimating the conditional density function of the state given observation, and used it as an alternative for the observation model. This heuristic may introduce bias in estimation, however. Ferris et al. (2006) proposed to use Gaussian Process regression for leaning the observation model. ”
Bayesian optimisation for Intelligent Environmental Monitoring (2012, methodology): “For example, GP regression has been used successfully for modelling the signal strength of a wireless signal [6], or gas concentration in indoor and outdoor environments [7].”
Indoor localization in multi-floor environments with reduced effort (2010, methodology): “We may expect that, using more sophisticated methods such as [3], [4], [8], we can further improve the localization accuracy.”
WiFi localization and navigation for autonomous indoor mobile robots (2010, methodology): “Wifi-based localization has been shown to be successful in different scenarios [6], [7], [8], [9] in terms of its ability to support humans or robots to identify their locations.”
GP-BayesFilters: Bayesian filtering using Gaussian process prediction and observation models (2008, background): “2009), the dynamics of robotics arms (Nguyen and Peters 2008), or the spatial distribution of wireless signal strength (Ferris et al. 2006).”
Transferring Localization Models across Space (2008, unknown): “Learning-based methods apply machine learning techniques, such as Gaussian processes (Ferris, Hähnel, and Fox 2006) and kernel methods (Nguyen, Jordan, and Sinopoli 2005), to handle the uncertainty in localization problems.”
RSSI-Based Indoor Localization and Tracking Using Sigma-Point Kalman Smoothers (2009, background): “More recently, particle filters have been used to demonstrate encouraging performance, although at a high computational cost for real time people tracking [4], [26]–[28].”
Sub-meter indoor localization in unmodified environments with inexpensive sensors (2010, background): “Localization with signal-strength mechanisms such as WiFi have been studied in the literature as well [17], [18], [19], [20], [21], including systems that bootstrap automatically without an explicit mapmaking step [22], [23].”
Learning GP-BayesFilters via Gaussian process latent variable models (2009, background): “Over the last years, Gaussian processes (GPs) have been applied with great success to robotics tasks such as reinforcement learning [3] and learning of prediction and observation models [5, 16, 8].”
Monte Carlo Filtering Using Kernel Embedding of Distributions (2014, methodology): “Comparisons are done with (1) the particle filter with k-nearest approach (PF-NN) (Vlassis, Terwijn, and Kröse 2002), (2) the particle filter with Gaussian process regression (PF-GP) (Ferris, Hähnel, and Fox 2006), and (3) Kernel Bayes’ rule filter (KBRF) (Fukumizu, Song, and Gretton 2011).”
Transferring Localization Models over Time (2008, unknown): “Learningbased techniques try to handle the uncertainty in wireless environments and use statistical learning such as KNN (Bahl and Padmanabhan 2000), kernel learning (Nguyen, Jordan, and Sinopoli 2005) and Gaussian process ( Ferris, Haehnel, and Fox 2006; Ferris, Fox, and Lawrence 2007).”
Adaptive Localization in a Dynamic WiFi Environment through Multi-view Learning (2007, methodology): “Localizing users or mobile nodes in wireless networks using received-signal-strength (RSS) values has attracted much attention in several research communities, especially in activity recognition in AI. In recent years, different statistical machine learning methods have been applied to the localizat”
WiFi-SLAM Using Gaussian Process Latent Variable Models (2007, methodology): “Observed values Y may impact the parameters of the covariance function via hyperparameter estimation [Ferris et al., 2006].”