A semi-supervised learning approach towards automatic wireless technology recognition

Camelo, Miguel; Shahid, Adnan; Fontaine, Jaron; Figueiredo, Felipe Augusto Pereira de; Poorter, Eli De; Moerman, Ingrid; Latré, Steven

doi:10.1109/dyspan.2019.8935690

Cited by 21 publications

(25 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Our recent work, e.g. [35], train autoencoders in two phases: a) an unsupervised pre-training phase and b) a supervised training phase. This pre-trained stacked autoencoder architecture is shown in Figure 4.…”

Section: A Autoencoder Basic Principlesmentioning

confidence: 99%

Edge Inference for UWB Ranging Error Correction Using Autoencoders

et al. 2020

Self Cite

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

Section: A Autoencoder Basic Principlesmentioning

confidence: 99%

Edge Inference for UWB Ranging Error Correction Using Autoencoders

et al. 2020

Self Cite

View full text Add to dashboard Cite

show abstract

“…This use case is similar to the one described by Figure 6. Mainly aligned with this use case, DARPA, the Defense Advanced Research Projects Agency from the United States, has started the Spectrum Collaboration Challenge with the aim to encourage research and development of smarter/more intelligent coexistence and collaboration techniques of heterogeneous networks in the same wireless spectrum bands [31,32]. One of the examples they have been advocating for is the adoption of such spectrum-sharing technologies in the CBRS band [33,34].…”

Section: A/b B/a C/a A/d B/dmentioning

confidence: 99%

A Baseband Wireless Spectrum Hypervisor for Multiplexing Concurrent OFDM Signals

Figueiredo

Mennes

Jabandzic

et al. 2020

Sensors

Self Cite

View full text Add to dashboard Cite

The next generation of wireless and mobile networks will have to handle a significant increase in traffic load compared to the current ones. This situation calls for novel ways to increase the spectral efficiency. Therefore, in this paper, we propose a wireless spectrum hypervisor architecture that abstracts a radio frequency (RF) front-end into a configurable number of virtual RF front ends. The proposed architecture has the ability to enable flexible spectrum access in existing wireless and mobile networks, which is a challenging task due to the limited spectrum programmability, i.e., the capability a system has to change the spectral properties of a given signal to fit an arbitrary frequency allocation. The proposed architecture is a non-intrusive and highly optimized wireless hypervisor that multiplexes the signals of several different and concurrent multi-carrier-based radio access technologies with numerologies that are multiple integers of one another, which are also referred in our work as radio access technologies with correlated numerology. For example, the proposed architecture can multiplex the signals of several Wi-Fi access points, several LTE base stations, several WiMAX base stations, etc. As it able to multiplex the signals of radio access technologies with correlated numerology, it can, for instance, multiplex the signals of LTE, 5G-NR and NB-IoT base stations. It abstracts a radio frequency front-end into a configurable number of virtual RF front ends, making it possible for such different technologies to share the same RF front-end and consequently reduce the costs and increasing the spectral efficiency by employing densification, once several networks share the same infrastructure or by dynamically accessing free chunks of spectrum. Therefore, the main goal of the proposed approach is to improve spectral efficiency by efficiently using vacant gaps in congested spectrum bandwidths or adopting network densification through infrastructure sharing. We demonstrate mathematically how our proposed approach works and present several simulation results proving its functionality and efficiency. Additionally, we designed and implemented an open-source and free proof of concept prototype of the proposed architecture, which can be used by researchers and developers to run experiments or extend the concept to other applications. We present several experimental results used to validate the proposed prototype. We demonstrate that the prototype can easily handle up to 12 concurrent physical layers.

show abstract

“…In the work presented in [10], the authors dealt with the problem involving the recognition of different Radio Access Technologies (RATs) employing Deep Autoencoders (DAEs). Their DAE model was used in a Semi-Supervised Learning (SSL) approach for the recognition of wireless technologies using raw IQ samples.…”

Section: Related Workmentioning

confidence: 99%

SCATTER PHY: An Open Source Physical Layer for the DARPA Spectrum Collaboration Challenge

et al. 2019

Self Cite

View full text Add to dashboard Cite

DARPA, the Defense Advanced Research Projects Agency from the United States, has started the Spectrum Collaboration Challenge with the aim to encourage research and development of coexistence and collaboration techniques of heterogeneous networks in the same wireless spectrum bands. Team SCATTER has been participating in the challenge since its beginning, back in 2016. SCATTER’s open-source software defined physical layer (SCATTER PHY) has been developed as a standalone application, with the ability to communicate with higher layers through a set of well defined messages (created with Google’s Protocol buffers) and that exchanged over a ZeroMQ bus. This approach allows upper layers to access it remotely or locally and change all parameters in real time through the control messages. SCATTER PHY runs on top of USRP based software defined radio devices (i.e., devices from Ettus or National Instruments) to send and receive wireless signals. It is a highly optimized and real-time configurable SDR based PHY layer that can be used for the research and development of novel intelligent spectrum sharing schemes and algorithms. The main objective of making SCATTER PHY available to the research and development community is to provide a solution that can be used out of the box to devise disruptive algorithms and techniques to optimize the sub-optimal use of the radio spectrum that exists today. This way, researchers and developers can mainly focus their attention on the development of smarter (i.e., intelligent algorithms and techniques) spectrum sharing approaches. Therefore, in this paper, we describe the design and main features of SCATTER PHY and showcase several experiments performed to assess the effectiveness and performance of the proposed PHY layer.

show abstract

A semi-supervised learning approach towards automatic wireless technology recognition

Cited by 21 publications

References 16 publications

Edge Inference for UWB Ranging Error Correction Using Autoencoders

Edge Inference for UWB Ranging Error Correction Using Autoencoders

A Baseband Wireless Spectrum Hypervisor for Multiplexing Concurrent OFDM Signals

SCATTER PHY: An Open Source Physical Layer for the DARPA Spectrum Collaboration Challenge

Contact Info

Product

Resources

About