WiFi-SLAM Using Gaussian Process Latent Variable Models

By Brian Ferris, Dieter Fox, Neil D. Lawrence. International Joint Conference on Artificial Intelligence, 2007

581 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/a9f3f81a1373dba3a60d86b24f2fb24f34b8abd9

How researchers see this paper

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

1. Zheng Yang
2. Chenshu Wu
3. Qiang Yang
4. Suining He
5. Dieter Fox
6. Jonathan Ko
7. Yunhao Liu
8. Manikanta Kotaru

Citations in the corpus

• GP-BayesFilters: Bayesian filtering using Gaussian process prediction and observation models (2008, background): “In (Ferris et al. 2007, 2006), GPs represent observation models for a person's or robot's position based on wireless signal strength.”
• Indoor Tracking: Theory, Methods, and Technologies (2015, background): “The most popular approach is most likely Wi-Fi-SLAM [77].”
• UJIIndoorLoc: A new multi-building and multi-floor database for WLAN fingerprint-based indoor localization problems (2014, methodology): “In [8], authors categorised approaches according to the technique used for localization into several paradigms, including calibration-free localization [9], WLAN based techniques [10], Dead-reckoning [11], simultaneous localization and mapping (SLAM) [12] and multi-modal sensing [13].”
• Past, Present, and Future of Simultaneous Localization and Mapping: Toward the Robust-Perception Age (2016, unknown): “as WiFi, can be used to improve robot navigation without any prior knowledge of the location of the antennas [79].”
• 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 ).”
• Wi-Fi Fingerprint-Based Indoor Positioning: Recent Advances and Comparisons (2016, methodology): “like signal-strength-based SLAM [109] utilizes the Gaussian process [65] to estimate the position of robots.”
• Locating in fingerprint space: wireless indoor localization with little human intervention (2012, methodology): “WiFi-SLAM [9] uses the Gaussian process latent variable models to relate RSS fingerprints and models human movements (displacement, direction, etc.”
• FedLoc: Federated Learning Framework for Data-Driven Cooperative Localization and Location Data Processing (2020, methodology): “Early variants of GPSSM were learned by finding the maximum a posteriori (MAP) estimates of the latent states, generating various successful positioning applications, among others the RSS-based WiFi localization [123], the human motion capture [124], and the IMU-based slotcar tracking [125], etc.”
• 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”
• Learning GP-BayesFilters via Gaussian process latent variable models (2009, methodology): “…introduction of so-called back-constraints (Lawrence and Quiñonero Candela 2006), which have been applied successfully in the context of WiFi-SLAM, where the goal is to learn an observation model for wireless signal strength data without relying on ground truth location data (Ferris et al. 2007).”
• No need to war-drive: unsupervised indoor localization (2012, methodology): “Recently, WiFi-SLAM [27] proposed the us­age of the WiFi signal strength for SLAM based on a Gaussian Process Latent Variable Model (GP-LVM).”
• Transferring Multi-device Localization Models using Latent Multi-task Learning (2008, methodology): “Since the localization problem is a regression problem, we follow ( Ferris, Fox, & Lawrence 2007 ) to report the average error distance3 and the standard deviation.”
• Transferring Localization Models across Space (2008, unknown): “In order to reduce the calibration effort or the size of labeled data, (Ferris, Fox, and Lawrence 2007) applied Gaussian-Process-Latent-Variable models (GP-LVMs) to exploit the latent-space locations under an unsupervised framework.”
• Hierarchical Gaussian process latent variable models (2007, methodology): “… is exploited in several of the successful applications of the GP-LVM: learning style from motion capture data (Gro­chow et al., 2004) learning a prior model for tracking (Urtasun et al., 2005; Urtasun et al., 2006) and robot simultaneous localisation and mapping (Ferris et al., 2007).”
• Accurate indoor localization with zero start-up cost (2014, background): “depth cameras [20] or odometry/laser-based simultaneous localization and mapping [8], which are unavailable for today’s mobile devices, unlike Ubicarse.”
• Smartphones Based Crowdsourcing for Indoor Localization (2015, methodology): “WiFi-SLAM [9] uses the Gaussian process latent variable models to relate RSS fingerprints and models human movements (displacement, direction, etc.”