DeepLocBox: Reliable Fingerprinting-Based Indoor Area Localization

Laska, Marius; Blankenbach, Jörg

doi:10.3390/s21062000

Cited by 25 publications

(21 citation statements)

References 56 publications

(86 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Therefore, it can be concluded that the positioning is accurate because the previous RSSI influences the current RSSI. 6.94 2D-CNN-DLB [20] 7.08 RSS clustering (affinity propagation) [21] 8.08 3D clustering (K-means) [21] 14.80…”

Section: Resultsmentioning

confidence: 99%

See 1 more Smart Citation

Movement Path Data Generation from Wi-Fi Fingerprints for Recurrent Neural Networks

Shin¹,

Choi

Yoon

2021

Sensors

View full text Add to dashboard Cite

The recurrent neural network (RNN) model, which is a deep-learning network that can memorize past information, is used in this paper to memorize continuous movements in indoor positioning to reduce positioning error. To use an RNN model in Wi-Fi-fingerprint based indoor positioning, data set must be sequential. However, Wi-Fi fingerprinting only saves the received signal strength indicator for a location, so it cannot be used as RNN data. For this reason, we propose a movement path data generation technique that generates data for an RNN model for sequential positioning from Wi-Fi fingerprint data. Movement path data can be generated by creating an adjacency list for Wi-Fi fingerprint location points. However, creating an adjacency matrix for all location points requires a large amount of computation. This problem is solved by dividing indoor environment by K-means clustering and creating a cluster transition matrix based on the center of each cluster.

show abstract

Section: Resultsmentioning

confidence: 99%

“…The experiment compares the positioning performance of the proposed method and that of the existing methods 3-Layer DNN, P-DNN [4], DNN-DLB [20], 2D-CNN-DLB [20], RSS clustering [21], and 3D clustering [21]. The existing methods determine the position using the Wi-Fi fingerprint inputs.…”

Section: Experimental Scenariomentioning

confidence: 99%

Movement Path Data Generation from Wi-Fi Fingerprints for Recurrent Neural Networks

Shin¹,

Choi

Yoon

2021

Sensors

View full text Add to dashboard Cite

show abstract

“…The utilized network backbone architectures cover a wide range of models. Multi-layer perceptron [19] either directly receives the fingerprint vectors or a low-dimensional embedding via stacked (denoising) autoencoder (S(D)AE) [4]- [7], [20]. Convolutional neural networks (CNN) can be directly applied on the fingerprint vector (1D-convolution) [21], on time-series fingerprints [21], or on 2D-input images, which are constructed from the original fingerprint vector [15], [22].…”

Section: Related Workmentioning

confidence: 99%

“…The overlapping grid cell encoding is adopted from our previous work in [19]. To support within grid cell regression, we encode each fingerprint position within the local coordinate system of the grid cell it is located in.…”

Section: B Data Encodingmentioning

confidence: 99%

Multi-Task Neural Network for Position Estimation in Large-Scale Indoor Environments

Laska

Blankenbach

2022

IEEE Access

Self Cite

View full text Add to dashboard Cite

Pedestrian localization within large-scale multi-building/multi-floor indoor environments remains a challenging task. Fingerprinting-based approaches are particularly suited for such large-scale deployments due to their low requirements of hardware installments. Recently, the fingerprinting problem has been addressed by deep learning. Existing models are mostly task specific by providing floor classification or position estimation within a small area. A strategy to support localization within large-scale environments is to sequentially apply hierarchical models. This has several drawbacks including missing scalability and increased deployment complexity on smartphones. We propose a unifying approach based on training a single neural network that classifies the building/floor and predicts the position in a single forward-pass of the network. Our model classifies a grid cell and performs within grid cell regression, which solves the performance degradation of applying regression within large areas. To reduce the error in case of misclassified grid cells, we propose a novel technique called multi cell encoding learning (multi-CEL), where a model simultaneously learns several redundant position representations within an overlapping grid cell encoding. For three public WLAN fingerprinting datasets, we demonstrate that multi-CEL surpasses existing state-of-the-art multi-task learning neural networks and even outperforms regression neural networks explicitly trained for 2D-positioning by up to 17%.

show abstract

“…In the literature, the accuracy of approximation of a function is often given as a statistical value determined on the basis of data unknown to the model. The data can be, e.g., the mean value, the median, or the relative mean square error (RMSE) between the predicted and the actual values [55]. In this study, for the set of data acquired on the basis of the above measurements made on the physical models and on the basis of calculations allowing the determination of the value of E 1−n (value of evaporation from the initial moment (1st day) to a current moment (n-th day)) permitted the development of two approximating functions.…”

Section: Calculationsmentioning

confidence: 99%

Application of the TDR Sensor and the Parameters of Injection Irrigation for the Estimation of Soil Evaporation Intensity

Walczak

Lipiński

Janik

2021

Sensors

View full text Add to dashboard Cite

The objective of the study was to develop a precise method of determination of the evaporation rate in a soil irrigated with the use of a mobile injection irrigation system. Two methods of constructing functions approximating the value of evaporation have been developed. In the first method, the domain comprises the parameters of injection irrigation, i.e., the dose and the depth of injection, and in the second, the volumetric moisture of soil in the layer immediately below the soil surface, which was measured with time-domain reflectometry (TDR) sensors. For that purpose, a laboratory experiment was carried out, based on 12 physical models. The study was conducted on a natural soil material, with particle size distribution of its mineral parts corresponding to that of a loamy sand soil. It was demonstrated that evaporation intensity increases with irrigation and decreases with increase in the depth of water application. Using TDR sensors, it was also shown that evaporation intensity increases proportionally to the weighted arithmetic mean of the volumetric moisture. Comparison of the two methods indicates that the evaporation intensity of injection-irrigated soil can be estimated with higher accuracy when the domain of the approximating function is the injection depth and dose than when the domain of the function is the weighted mean of volumetric moisture of the surface horizon of the soil. However, the method using TDR sensors for the estimation of evaporation intensity of an injection-irrigated soil has a greater potential for the construction of universal approximating models. In addition, the advantage of the method based on the use of TDR sensors is that it uses arguments for the approximating function, f2θ, in real time.

show abstract

DeepLocBox: Reliable Fingerprinting-Based Indoor Area Localization

Cited by 25 publications

References 56 publications

Movement Path Data Generation from Wi-Fi Fingerprints for Recurrent Neural Networks

Movement Path Data Generation from Wi-Fi Fingerprints for Recurrent Neural Networks

Multi-Task Neural Network for Position Estimation in Large-Scale Indoor Environments

Application of the TDR Sensor and the Parameters of Injection Irrigation for the Estimation of Soil Evaporation Intensity

Contact Info

Product

Resources

About