Deep Convolutional Neural Networks for Indoor Localization with CSI Images

Wang, Xuyu; Wang, Xiangyu; Mao, Shiwen

doi:10.1109/tnse.2018.2871165

Cited by 184 publications

(128 citation statements)

References 41 publications

Supporting

Mentioning

102

Contrasting

Order By: Relevance

“…Several recent studies tackling indoor positioning have employed such machine learning models: a four layer deep neural network, combined with a denoising autoencoder and a hidden Markov model is employed in [51] for indoor and outdoor localization; a recurrent neural network which uses WiFi signals for an indoor positioning system is presented in [21]; another recurrent neural network, more specifically a long short-term memory network is used by Urano et al [43] with BLE signal strength data for indoor localization; Mittal et al [29] propose a convolutional neural networks based framework for indoor localization, in which the networks use images created from WiFi signatures. Another convolutional deep neural network starting from phase data of channel state information, which is transformed into images based on estimated angles of arrival is presented in [46]. In our model, the neural network input is a triplet that contains the three RSSI values recorded by the intelligent luminaires, while the output is the estimated location, represented as a coordinate pair.…”

Section: Neural Network Based Techniquementioning

confidence: 99%

Indoor Localization Techniques Within a Home Monitoring Platform

Marin¹,

Bocicor²,

Molnar³

2020

Communications in Computer and Information Science

View full text Add to dashboard Cite

This paper details a number of indoor localization techniques developed for real-time monitoring of older adults. These were developed within the framework of the i-Light research project that was funded by the European Union. The project targeted the development and initial evaluation of a configurable and cost-effective cyber-physical system for monitoring the safety of older adults who are living in their own homes. Localization hardware consists of a number of custom-developed devices that replace existing luminaires. In addition to lighting capabilities, they measure the strength of a Bluetooth Low Energy signal emitted by a wearable device on the user. Readings are recorded in real time and sent to a software server for analysis. We present a comparative evaluation of the accuracy achieved by several server-side algorithms, including Kalman filtering, a look-back heuristic as well as a neural networkbased approach. It is known that approaches based on measuring signal strength are sensitive to the placement of walls, construction materials used, the presence of doors as well as existing furniture. As such, we evaluate the proposed approaches in two separate locations having distinct building characteristics. We show that the proposed techniques improve the accuracy of localization. As the final step, we evaluate our results against comparable existing approaches.

show abstract

Section: Neural Network Based Techniquementioning

confidence: 99%

Indoor Localization Techniques Within a Home Monitoring Platform

Marin¹,

Bocicor²,

Molnar³

2020

Communications in Computer and Information Science

View full text Add to dashboard Cite

show abstract

“…A segmented signal frame was used to determine the pedestrian speed for CNN and RNN, and the walking distance was estimated by calculating the speed and moving time. Wang et al [21] extracted channel state information (CSI) data from WiFi as the input for deep CNN to predict the position of mobile devices. Mittal et al [22] also used WiFi data.…”

Section: Convolutional Neural Networkmentioning

confidence: 99%

Integration of Convolutional Neural Network and Error Correction for Indoor Positioning

Ciou

2020

IJGI

View full text Add to dashboard Cite

With the rapid development of surveying and spatial information technologies, more and more attention has been given to positioning. In outdoor environments, people can easily obtain positioning services through global navigation satellite systems (GNSS). In indoor environments, the GNSS signal is often lost, while other positioning problems, such as dead reckoning and wireless signals, will face accumulated errors and signal interference. Therefore, this research uses images to realize a positioning service. The main concept of this work is to establish a model for an indoor field image and its coordinate information and to judge its position by image eigenvalue matching. Based on the architecture of PoseNet, the image is input into a 23-layer convolutional neural network according to various sizes to train end-to-end location identification tasks, and the three-dimensional position vector of the camera is regressed. The experimental data are taken from the underground parking lot and the Palace Museum. The preliminary experimental results show that this new method designed by us can effectively improve the accuracy of indoor positioning by about 20% to 30%. In addition, this paper also discusses other architectures, field sizes, camera parameters, and error corrections for this neural network system. The preliminary experimental results show that the angle error correction method designed by us can effectively improve positioning by about 20%.

show abstract

“…For indoor localization problems, CiFi [33,36] system leverages generated images with estimated AOA values with commodity 5GHz WiFi as the input of CNN for indoor localization, which can be trained by backpropagation (BP) algorithm. This system demonstrates that the performance of the localization has outperformed existing schemes, like FIFS and Horus.…”

Section: Convolutional Neural Networkmentioning

confidence: 99%

Deep learning for indoor localization based on bimodal CSI data

Wang

Mao

2019

Applications of Machine Learning in Wireless Communications

View full text Add to dashboard Cite

In this chapter, we incorporate deep learning for indoor localization based on channel state information with commodity 5GHz WiFi. We first introduce the state-of-the-art deep learning techniques including deep autoencoder network, convolutional neural network, and recurrent neural network. We then present a deep learning based algorithm to leverage bi-modal CSI data, i.e., average amplitudes and estimated angle of arrivals (AOA), in both offline and online stages of fingerprinting. The proposed scheme is validated with extensive experiments. Finally, we examine several open research problems for indoor localization based on deep learning techniques.

show abstract

Deep Convolutional Neural Networks for Indoor Localization with CSI Images

Cited by 184 publications

References 41 publications

Indoor Localization Techniques Within a Home Monitoring Platform

Indoor Localization Techniques Within a Home Monitoring Platform

Integration of Convolutional Neural Network and Error Correction for Indoor Positioning

Deep learning for indoor localization based on bimodal CSI data

Contact Info

Product

Resources

About