Fingerprinting Indoor Positioning Method Based on Kernel Ridge Regression with Feature Reduction

Le, Yanfen; Jin, Shijialuo; Zhang, Hena; Shi, Wei; Yao, Heng

doi:10.1155/2021/6631585

Cited by 8 publications

(10 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To overcome the nonlinear localization problem, a kernel-based PCA was proposed in [14]. To make the RSSI vectors linearly independent, a nonlinear mapping was formed between the fundamental space of RSSI vectors R d and a low-dimensional feature space F , that is, ζ : R d → F .…”

Section: Feature Extraction By Pcamentioning

confidence: 99%

“…To make the RSSI vectors linearly independent, a nonlinear mapping was formed between the fundamental space of RSSI vectors R d and a low-dimensional feature space F , that is, ζ : R d → F . Hence, the RSSI vectors' covariance matrix in the spatial domain is given by [14]:…”

Section: Feature Extraction By Pcamentioning

confidence: 99%

“…Discriminant analysis such as linear discriminant analysis (LDA) or Fisher's discriminant analysis (FDA) [12,13] and principal component analysis (PCA) [14][15][16] are among the well-known feature dimension reduction techniques for RSSI fingerprinting systems. In [14], a new collection of principal components (PCs) was obtained by reconstructing RSSI measurements based on the PCA.The PCA aims to reduce the mean square reconstruction error as much as possible by selecting a portion of PCs while maintaining relevant data from all of the APs for the localization model. In [15], a kernel PCA was developed to achieve a nonlinear mapping between the PCs and the user position, which shows an improvement on [14].…”

Section: Introductionmentioning

confidence: 99%

“…In [14], a new collection of principal components (PCs) was obtained by reconstructing RSSI measurements based on the PCA.The PCA aims to reduce the mean square reconstruction error as much as possible by selecting a portion of PCs while maintaining relevant data from all of the APs for the localization model. In [15], a kernel PCA was developed to achieve a nonlinear mapping between the PCs and the user position, which shows an improvement on [14]. To reduce the impact of environmental interference, AP clusters were constructed in [16], where hierarchical PCA was used to obtain cluster RSSI features by converting the raw data to a linearly independent dimensional matrix.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

WiFi Fingerprinting Indoor Localization Based on Dynamic Mode Decomposition Feature Selection with Hidden Markov Model

Babalola¹,

Balyan²

2021

Sensors

View full text Add to dashboard Cite

Over the years, WiFi received signal strength indicator (RSSI) measurements have been widely implemented for determining the location of a user’s position in an indoor environment, where the GPS signal might not be received. This method utilizes a huge RSSI dataset collected from numerous access points (APs). The WiFi RSSI measurements are nonlinear with distance and are largely influenced by interference in the indoor environment. Therefore, machine learning (ML) techniques such as a hidden Markov model (HMM) are generally utilized to efficiently identify a trend of RSSI values, which corresponds to locations around a region of interest. Similar to other ML tools, the performance and computing cost of the HMM are dependent on the feature dimension since a large quantity of RSSI measurements are required for the learning process. Hence, this article introduces a feature extraction method based on dynamic mode decomposition (DMD) for the HMM to effectively model WiFi fingerprint indoor localization. The DMD is adopted since it decomposes RSSIs to meaningful spatial and temporal forms over a given time. Here, the mode forms are analytically reconstructed to produce low-dimensional feature vectors, which are used with the HMM. The localization performance of the proposed HMM-DMD is compared with other well-known ML algorithms for WiFi fingerprinting localization using simulations. The results show that the HMM-DMD algorithm yields a significant localization performance improvement, accuracy, and reasonable processing time in comparison with the state-of-the-art algorithms.

show abstract

Section: Feature Extraction By Pcamentioning

confidence: 99%

Section: Feature Extraction By Pcamentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

WiFi Fingerprinting Indoor Localization Based on Dynamic Mode Decomposition Feature Selection with Hidden Markov Model

Babalola¹,

Balyan²

2021

Sensors

View full text Add to dashboard Cite

show abstract

“…A combination of built-in smartphone sensors, portable monitors, and geographic data comprise of the Expoapp system in providing a suite of dynamic environmental exposure measurements 27 . Other sensors attempt to accurately classify participant location using indoor positioning methods based on the Received Signal Strength (RSS) fingerprint or a cluster principal component analysis-based indoor positioning algorithm 28,29 . Wireless temperature skin sensors have been used to test the hypothesis whether the temporal association between personal air temperature and blood pressure was mediated via skin temperature, and whether this relationship was seasondependent 30 .…”

mentioning

confidence: 99%

A randomized cross-over trial investigating differences in 24-h personal air and skin temperatures using wearable sensors between two climatologically contrasting settings

Constantinou

Oikonomou

Konstantinou

et al. 2021

Sci Rep

View full text Add to dashboard Cite

The influence of elevated air temperatures recorded in various urban microenvironments in adversely impacting biologically relevant disease end points has not yet been extensively tackled. This study is a post hoc analysis of the TEMP pilot trial, a randomized 2 × 2 cross-over trial that examined changes in metabolic and stress hormonal profiles of healthy adults in two settings (urban vs. rural) with distinctly different climatological characteristics during the Mediterranean summer. This analysis aimed to study the association between the 24-h personal air or skin temperature sensor measurements and the diary-based location type (indoors vs. outdoors) in urban (seaside) vs. rural (higher in altitude) microenvironments. Out of 41 eligible participants, a total of 37 participants were included in this post-hoc TEMP trial analysis. Wearable sensors recorded personal air temperature, skin temperature, and activity (as a surrogate marker of physical activity) in each setting, while a time-stamped personal diary recorded the types of indoor or outdoor activities. Temperature peaks during the 24-h sampling period were detected using a peak finding algorithm. Mixed effect logistic regression models were fitted for the odds of participant location (being indoors vs. outdoors) as a function of setting (urban vs. rural) and sensor-based personal temperature data (either raw temperature values or number of temperature peaks). During the study period (July–end of September), median [interquartile range, IQR] personal air temperature in the rural (higher altitude) settings was 1.5 °C lower than that in the urban settings (27.1 °C [25.4, 29.2] vs. 28.6 °C [27.1, 30.5], p < 0.001), being consistent with the Mediterranean climate. Median [IQR] personal air temperature in indoor (micro)environments was lower than those in outdoors (28.0 °C [26.4, 30.3] vs 28.5 °C [26.8, 30.7], p < 0.001). However, median [IQR] skin temperature was higher in indoor (micro)environments vs. outdoors (34.8 °C [34.0, 35.6] and 33.9 °C [32.9, 34.8], p < 0.001) and the number of both personal air and skin temperature peaks was higher indoors compared to outdoors (median [IQR] 3.0 [2.0,4.0] vs 1.0 [1.0,1.3], p < 0.007, for the skin sensors). A significant association between the number of temperature peaks and indoor location types was observed with either the personal air sensor (OR 3.1; 95% CI 1.2–8.2; p = 0.02) or the skin sensor (OR 3.7; 95% CI 1.4–9.9; p = 0.01), suggesting higher number of indoor air temperature fluctuations. Amidst the global climate crisis, more population health studies or personalized medicine approaches that utilize continuous tracking of individual-level air/skin temperatures in both indoor/outdoor locations would be warranted, if we were to better characterize the disease phenotype in response to climate change manifestations.

show abstract

Artificial intelligence techniques in wireless sensor networks for accurate localization of user in floor, building and indoor area

Chadha

Jain

Kumar

2022

Multimed Tools Appl

View full text Add to dashboard Cite

Fingerprinting Indoor Positioning Method Based on Kernel Ridge Regression with Feature Reduction

Cited by 8 publications

References 31 publications

WiFi Fingerprinting Indoor Localization Based on Dynamic Mode Decomposition Feature Selection with Hidden Markov Model

WiFi Fingerprinting Indoor Localization Based on Dynamic Mode Decomposition Feature Selection with Hidden Markov Model

A randomized cross-over trial investigating differences in 24-h personal air and skin temperatures using wearable sensors between two climatologically contrasting settings

Artificial intelligence techniques in wireless sensor networks for accurate localization of user in floor, building and indoor area

Contact Info

Product

Resources

About