 ICACT20230228 Slide.00
[Big Slide]
|
Chrome  Click!! |
 |
Good morning, My name is Yafei Hou from Wave Engeering Laboratory, ATR International, Kyoto, Japan. Today, I would like to give a research presentation with the title as MUSIC spectrum based interference detection and localization for mmWave RIS-MIMO dystem. |
| ICACT20230228 Slide.01
[Big Slide]
|
Chrome Click!! |
 |
Good morning, My name is Yafei Hou from Wave Engeering Laboratory, ATR International, Kyoto, Japan. Today, I would like to give a research presentation with the title as MUSIC spectrum based interference detection and localization for mmWave RIS-MIMO dystem. |
| ICACT20230228 Slide.02
[Big Slide]
|
Chrome Click!! |
 |
The main contents of this presentation include following parts. First, I will introduce the background and contributions of this paper. Then I will give some simple explanations on the mmWave RIS-MIMO system and the idea of MUSCI spectrum which related to the interference and location information. Next, the main idea of MUSIC Spectrum based interference detection and localization will be given. Finally, the simulation results and conclusions will be explained. |
| ICACT20230228 Slide.03
[Big Slide]
|
Chrome Click!! |
 |
As you know, Beyond5G/6G era is coming for next ten years and many new applications will totally change our life styles. The major purposes of Beyond5G/6G can be shown in these figures which include many targets such as higher data rate, lower power consumption, larger wireless coverage and smarter wireless devices. |
| ICACT20230228 Slide.04
[Big Slide]
|
Chrome Click!! |
 |
To achieve these purposes, many RAN technologies for beyond 5G/6G system. For examples, some technologies as advanced antenna technologies, e.g. Massive MIMO, MU-MIMO, advanced beamforming and energy efficiency over higher frequency bands. In addition, MIMO technique over mmWave bands will be one of most promising technologies for 5G/6G. |
| ICACT20230228 Slide.05
[Big Slide]
|
Chrome Click!! |
 |
On the other hand, recently, another low-cost and powerful antenna technology, which is named as reconfigurable intelligent surface, is coming and becoming hot research topic. These two figures are shown the major properties of RIS. First, RIS does not need extra RF chains for demodulation and modulation or analog beamforming. It also will not generate extra RF signal. The input signals can be reflected or transmitted at RIS. The phase shift or radiation of input signal can be controlled by PIN diodes bias voltages (ON/OFF the diode). Therefore, the programming the ON/OFF all diodes jointly realize different beamforming modes. |
| ICACT20230228 Slide.06
[Big Slide]
|
Chrome Click!! |
 |
RIS may constitute an emerging and promising software-defined architecture that can be realized at reduced cost, size, weight, and power. It can be used for many applications as shown in these figures. |
| ICACT20230228 Slide.07
[Big Slide]
|
Chrome Click!! |
 |
Using RIS can configure a smart radio environment which can improve performance of wireless system. In addition, for short wavelength of mmWave system, the larger number of cells can be integrated which largely improve the received signal power and then the capacity of mmWave MIMO system. |
| ICACT20230228 Slide.08
[Big Slide]
|
Chrome Click!! |
 |
Usually, mmWave MIMO system utilizes beamforming to achieve high capacity for multiple desired users. However, interference often occurs which largely reduces MIMO capacity and power consumption efficiency. Therefore, if interference occurrence can be detected, and the source can be localized, beamforming technology can reduce the interference component and largely improve the MIMO capacity and energy efficiency. |
| ICACT20230228 Slide.09
[Big Slide]
|
Chrome Click!! |
 |
Therefore, the major contributions of this paper will be followings. We propose a MUtiple SIgnal Classification (MUSIC) spectrum-based method to detect an interference occurrence and localize the interference source for mmWave RIS-MIMO system. In addition, the proposal only utilizes the received signal and does not require any channel status information of any links. Therefore, the proposal reduces the computational complexity and resource of channel estimation. |
| ICACT20230228 Slide.10
[Big Slide]
|
Chrome Click!! |
 |
Let me introduce the model of mmWave RIS-MIMO system using this figure. The equation shows the received signal at one receiver antenna from one transmitter antenna which includes direct path and some reflected paths. In addition, RIS with M cells reflect the signal between the transmitter-RIS path and RIS-receiver path. The phase value of each cell can be adjusted according to different purposes. In this paper, we utilize the RIS to maximize the SNR or power value of the link. |
| ICACT20230228 Slide.11
[Big Slide]
|
Chrome Click!! |
 |
In addition, RIS channel model is divided into two scenarios that is near-field model and far-field model. Both models are shown in these two kinds of equations. Far-field model usually means the distance between the RIS and transmitter/receiver is long, and system works over a low frequency band. On the contrary, near-field model is more appropriate for RIS system located in an indoor case over mmWave band. |
| ICACT20230228 Slide.12
[Big Slide]
|
Chrome Click!! |
 |
This paper proposes a method of MUSIC spectrum based interference detection and localization. To explain the main idea of the proposal, we first explain the concept of MUSIC spectrum which can be shown in these equations. At receiver side, K receive antennas can receive the signal using the RIS channel model with L transmit antennas. Then the auto-correlation process will be operated on K-sample received signal yt. Then using SVD operation on K-by-K auto-correlation Ryy, the MUSIC spectrum can be obtained. |
| ICACT20230228 Slide.13
[Big Slide]
|
Chrome Click!! |
 |
Using the eigenvalue and eigenvector of Ryy, the MUSIC Spectrum P(theta,pi) can be shown in this equation. Usually, the AoA estimation searches the MUSIC spectrums that are orthogonal to the noise subspace. Therefore, the MUSIC spectrum can be treated as the signature of space or user location. We can use this property to realize the interference detection and localization of interference. |
| ICACT20230228 Slide.14
[Big Slide]
|
Chrome Click!! |
 |
The interference detection can be simplified as following steps. Firstly, at receive side, it needs to generate N_K MUSIC spectrums without interference as fingerprint. Then these fingerprints can be used as standard of no interference, and the smallest correlation value among all MUSIC spectrums can be used as threshold for interference detection. Finally, the detection of the interference occurrence can be realized using these rules. |
| ICACT20230228 Slide.15
[Big Slide]
|
Chrome Click!! |
 |
Using the similar method, we can use MUSIC spectrum to coarsely estimate the location of the interference source. The BS can divide the space into many small grids and select some known locations as interference sources to build the known MUSIC spectrums related to these selected locations as fingerprint samples. For the localization of the interference source randomly generated from unknown location, BS just calculates the correlation values between these known MUSIC spectrums and this new MUSIC spectrum intended for localization and find the maximum one. Then the location is decided with the selected position of the grid which has the maximum correlation value. Since the interference sources are randomly from the space, the method only coarsely detects the location of interference source. |
| ICACT20230228 Slide.16
[Big Slide]
|
Chrome Click!! |
 |
The simulation scenario is shown in this figure. We assume that system includes one BS, two UEs and RIS. One interference transmitter is set with a random location in the room. BS antenna is type of uniform linear array (ULA) which parallels to the Y-axis. Two RIS patterns are used for the increase of link power. The first one is making the RIS point to UEs which makes the link power between the RIS cells and UEs be maximum. The second pattern is making the RIS point to BS and increases the link power between the BS and UEs. |
| ICACT20230228 Slide.17
[Big Slide]
|
Chrome Click!! |
 |
The simulation parameters and path loss model are shown in this table and equation. |
| ICACT20230228 Slide.18
[Big Slide]
|
Chrome Click!! |
 |
The performance of interference detection is given in these figures with two different RIS patterns. As shown in both figures, without interference, the proposed method can judge correct information near to 100% when the FP samples is over 10. When interference occurs, the accuracy of both cases is lightly reduced FP samples is increased to 10. In addition, area B has large power loss which reduces the accuracy of interference detection. |
| ICACT20230228 Slide.19
[Big Slide]
|
Chrome Click!! |
 |
The accuracy of proposed MUSIC spectrum based localization is given in these figures with two different RIS patterns. From the results, we can find that the proposal can provide the coarse location information of the interference source. In addition, the localization of Y-axis is higher than that of X-axis because ULA of BS is parallel to the Y-axis which has large spatial resolution. However, RIS pattern seems to have not large impact on the localization performance and interference detection. The main reason is that the RIS pattern only aims to change the link powers between the BS and UEs but not for the detection of interference source. |
| ICACT20230228 Slide.20
[Big Slide]
|
Chrome Click!! |
 |
Finally, let us give simple conclusions for this paper. MUSIC spectrum can be used for AoA detection and channel estimation which includes location information of transmitter. This paper utilizes this property to realize the interference detection and coarse interference localization. The proposal does not require any channel status information of any links which reduces the computational complexity and resource of channel estimation. The proposed method can be further improved using machine learning, distributed BSs and adjustable RIS patterns |
| ICACT20230228 Slide.21
[Big Slide]
|
Chrome Click!! |
 |
Thank you for your attention! |
