Environmental Cross-Validation of NLOS Machine Learning Classification/Mitigation with Low-Cost UWB Positioning Systems

Barral, Valentín; Escudero, Carlos J.; García-Naya, José A.; Suárez-Casal, Pedro

doi:10.3390/s19245438

Cited by 23 publications

(22 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, we note that proposals [ 27 , 28 ] do not consider the transformation of RSSI samples using quartiles, as we proposed in our article. We add as a comment that the differential of our proposal is to use data transformation using quartiles in order to create signatures that show statistical differences between the RSSI samples collected in different RPs/TPs.…”

Section: Related Workmentioning

confidence: 99%

“…The works of [ 27 , 28 ] use low-cost UWB devices in their proposals for internal location for NLOS (non-line-of-sight) environments. In particular, the work of [ 27 ] uses machine learning (ML) techniques to analyze various RSSI measurement scenarios and identify measurements that showed variations related to NLOS propagation characteristics of the signals.…”

Section: Related Workmentioning

confidence: 99%

“…In particular, the work of [ 27 ] uses machine learning (ML) techniques to analyze various RSSI measurement scenarios and identify measurements that showed variations related to NLOS propagation characteristics of the signals. The work of [ 28 ] proposes and analyzes the performance of a location system that combines location algorithms with ML techniques applied to a previous classification process to reduce problems related to the propagation of signals in indoor environments.…”

Section: Related Workmentioning

confidence: 99%

See 2 more Smart Citations

QA-kNN: Indoor Localization Based on Quartile Analysis and the kNN Classifier for Wireless Networks

Ferreira

Souza

Carvalho

2020

Sensors

View full text Add to dashboard Cite

Considering the variation of the received signal strength indicator (RSSI) in wireless networks, the objective of this study is to investigate and propose a method of indoor localization in order to improve the accuracy of localization that is compromised by RSSI variation. For this, quartile analysis is used for data pre-processing and the k-nearest neighbors (kNN) classifier is used for localization. In addition to the tests in a real environment, simulations were performed, varying many parameters related to the proposed method and the environment. In the real environment with reference points of 1.284 density per unit area (RPs/m2), the method presents zero-mean error in the localization in test points (TPs) coinciding with the RPs. In the simulated environment with a density of 0.327 RPs/m2, a mean error of 0.490 m for the localization of random TPs was achieved. These results are important contributions and allow us to conclude that the method is promising for locating objects in indoor environments.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

QA-kNN: Indoor Localization Based on Quartile Analysis and the kNN Classifier for Wireless Networks

Ferreira

Souza

Carvalho

2020

Sensors

View full text Add to dashboard Cite

show abstract

“…In contrast to the methodology proposed in this paper, which tries to exploit the deep learning capabilities of CNNs using raw CIR data, the authors of [22] and [23] focus on extracted features from grouped ranges to classify and correct NLOS signals. The authors collected their opensource dataset in two locations situated in office environments.…”

Section: Lowmentioning

confidence: 99%

Edge Inference for UWB Ranging Error Correction Using Autoencoders

et al. 2020

View full text Add to dashboard Cite

Indoor localization knows many applications, such as industry 4.0, warehouses, healthcare, drones, etc., where high accuracy becomes more critical than ever. Recent advances in ultra-wideband localization systems allow high accuracies for multiple active users in line-of-sight environments, while they still introduce errors above 300 mm in non-line-of-sight environments due to multi-path effects. Current work tries to improve the localization accuracy of ultra-wideband through offline error correction approaches using popular machine learning techniques. However, these techniques are still limited to simple environments with few multi-path effects and focus on offline correction. With the upcoming demand for high accuracy and low latency indoor localization systems, there is a need to deploy (online) efficient error correction techniques with fast response times in dynamic and complex environments. To address this, we propose (i) a novel semi-supervised autoencoder-based machine learning approach for improving ranging accuracy of ultra-wideband localization beyond the limitations of current improvements while aiming for performance improvements and a small memory footprint and (ii) an edge inference architecture for online UWB ranging error correction. As such, this paper allows the design of accurate localization systems by using machine learning for low-cost edge devices. Compared to a deep neural network (as state-of-the-art, with a baseline error of 75 mm) the proposed autoencoder achieves a 29% higher accuracy. The proposed approach leverages robust and accurate ultra-wideband localization, which reduces the errors from 214 mm without correction to 58 mm with correction. Validation of edge inference using the proposed autoencoder on a NVIDIA Jetson Nano demonstrates significant uplink bandwidth savings and allows up to 20 rapidly ranging anchors per edge GPU.

show abstract

“…Many classification and machine learning methods have been used for measurement data analyses. Recent studies include machine learning with optics, NIR and X-rays [27][28][29][30][31][32]. Electromagnetic waves using radio and microwave frequencies have been used with machine learning based classification techniques [33,34].…”

Section: Introductionmentioning

confidence: 99%

Classification of Wood Chips Using Electrical Impedance Spectroscopy and Machine Learning

Tiitta

Heikkinen

et al. 2020

Sensors

View full text Add to dashboard Cite

Wood chips are extensively utilised as raw material for the pulp and bio-fuel industry, and advanced material analyses may improve the processes in utilizing these products. Electrical impedance spectroscopy (EIS) combined with machine learning was used in order to analyse heartwood content of pine chips and bark content of birch chips. A novel electrode system integrated in a sampling container was developed for the testing using frequency range 42 Hz–5 MHz. Three electrode pairs were used to measure the samples in x-, y- and z-direction. Three machine learning methods were used: K-nearest neighbor (KNN), decision tree (DT) and support vector machines (SVM). The heartwood content of pine chips and bark content of birch chips were classified with an accuracy of 91% using EIS from pure materials combined with a k-nearest neighbour classifier. When using mixed materials and multiple classes, 73% correct classification for pine heartwood content (four groups) and 64% for birch bark content (five groups) were achieved.

show abstract

Environmental Cross-Validation of NLOS Machine Learning Classification/Mitigation with Low-Cost UWB Positioning Systems

Cited by 23 publications

References 15 publications

QA-kNN: Indoor Localization Based on Quartile Analysis and the kNN Classifier for Wireless Networks

QA-kNN: Indoor Localization Based on Quartile Analysis and the kNN Classifier for Wireless Networks

Edge Inference for UWB Ranging Error Correction Using Autoencoders

Classification of Wood Chips Using Electrical Impedance Spectroscopy and Machine Learning

Contact Info

Product

Resources

About