DeepAoANet: Learning angle of arrival from software defined radios with deep neural networks
Direction finding and positioning systems based on RF signals are significantly impacted by multipath propagation, particularly in indoor environments. Existing algorithms (e.g MUSIC) perform poorly in resolving Angle of Arrival (AoA) in the presence of multipath or when operating in a weak signal r...
| Main Authors: | , , , , |
|---|---|
| Format: | Journal article |
| Language: | English |
| Published: |
IEEE
2022
|
| _version_ | 1797107075840475136 |
|---|---|
| author | Dai, Z He, Y Tran, V Trigoni, N Markham, A |
| author_facet | Dai, Z He, Y Tran, V Trigoni, N Markham, A |
| author_sort | Dai, Z |
| collection | OXFORD |
| description | Direction finding and positioning systems based on RF signals are significantly impacted by multipath propagation, particularly in indoor environments. Existing algorithms (e.g MUSIC) perform poorly in resolving Angle of Arrival (AoA) in the presence of multipath or when operating in a weak signal regime. We note that digitally sampled RF frontends allow for the easy analysis of signals, and their delayed components. Low-cost Software-Defined Radio (SDR) modules enable Channel State Information (CSI) extraction across a wide spectrum, motivating the design of an enhanced AoA solution. We propose a Deep Learning approach for deriving AoA from a single snapshot of the SDR multichannel data. We compare and contrast deep-learning based angle classification and regression models, to estimate up to two AoAs accurately. We have implemented the inference engines on different platforms to extract AoAs in real-time, demonstrating the computational tractability of our approach. To demonstrate the utility of our approach we have collected IQ (In-phase and Quadrature components) samples from a four-element Universal Linear Array (ULA) in various Line-of-Sight (LOS) and Non-Line-of-Sight (NLOS) environments, and published the dataset. Our proposed method demonstrates excellent reliability in determining number of impinging signals and realized mean absolute AoA errors less than 2°. |
| first_indexed | 2024-03-07T07:11:22Z |
| format | Journal article |
| id | oxford-uuid:cc046da8-c7ed-4c8e-942b-c46f76aaa9ac |
| institution | University of Oxford |
| language | English |
| last_indexed | 2024-03-07T07:11:22Z |
| publishDate | 2022 |
| publisher | IEEE |
| record_format | dspace |
| spelling | oxford-uuid:cc046da8-c7ed-4c8e-942b-c46f76aaa9ac2022-06-22T13:05:14ZDeepAoANet: Learning angle of arrival from software defined radios with deep neural networksJournal articlehttp://purl.org/coar/resource_type/c_dcae04bcuuid:cc046da8-c7ed-4c8e-942b-c46f76aaa9acEnglishSymplectic ElementsIEEE2022Dai, ZHe, YTran, VTrigoni, NMarkham, ADirection finding and positioning systems based on RF signals are significantly impacted by multipath propagation, particularly in indoor environments. Existing algorithms (e.g MUSIC) perform poorly in resolving Angle of Arrival (AoA) in the presence of multipath or when operating in a weak signal regime. We note that digitally sampled RF frontends allow for the easy analysis of signals, and their delayed components. Low-cost Software-Defined Radio (SDR) modules enable Channel State Information (CSI) extraction across a wide spectrum, motivating the design of an enhanced AoA solution. We propose a Deep Learning approach for deriving AoA from a single snapshot of the SDR multichannel data. We compare and contrast deep-learning based angle classification and regression models, to estimate up to two AoAs accurately. We have implemented the inference engines on different platforms to extract AoAs in real-time, demonstrating the computational tractability of our approach. To demonstrate the utility of our approach we have collected IQ (In-phase and Quadrature components) samples from a four-element Universal Linear Array (ULA) in various Line-of-Sight (LOS) and Non-Line-of-Sight (NLOS) environments, and published the dataset. Our proposed method demonstrates excellent reliability in determining number of impinging signals and realized mean absolute AoA errors less than 2°. |
| spellingShingle | Dai, Z He, Y Tran, V Trigoni, N Markham, A DeepAoANet: Learning angle of arrival from software defined radios with deep neural networks |
| title | DeepAoANet: Learning angle of arrival from software defined radios with deep neural networks |
| title_full | DeepAoANet: Learning angle of arrival from software defined radios with deep neural networks |
| title_fullStr | DeepAoANet: Learning angle of arrival from software defined radios with deep neural networks |
| title_full_unstemmed | DeepAoANet: Learning angle of arrival from software defined radios with deep neural networks |
| title_short | DeepAoANet: Learning angle of arrival from software defined radios with deep neural networks |
| title_sort | deepaoanet learning angle of arrival from software defined radios with deep neural networks |
| work_keys_str_mv | AT daiz deepaoanetlearningangleofarrivalfromsoftwaredefinedradioswithdeepneuralnetworks AT hey deepaoanetlearningangleofarrivalfromsoftwaredefinedradioswithdeepneuralnetworks AT tranv deepaoanetlearningangleofarrivalfromsoftwaredefinedradioswithdeepneuralnetworks AT trigonin deepaoanetlearningangleofarrivalfromsoftwaredefinedradioswithdeepneuralnetworks AT markhama deepaoanetlearningangleofarrivalfromsoftwaredefinedradioswithdeepneuralnetworks |