An adaptive beamwidth selection method for multi-targets in cf-mmimo systems

Abstract

The invention is related to an adaptive beamwidth selection technique that allows Cell-Free Massive multiple-input and multiple-output (CF-mMIMO) systems to dynamically adjust the beamwidths of their APs (Access Point), focusing on specific angles to improve angular resolution and detect multiple closely spaced targets more effectively

Description

[0001] DESCRIPTION

[0002] AN ADAPTIVE BEAMWIDTH SELECTION METHOD FOR MULTI-TARGETS IN CF-MMIMO SYSTEMS

[0003] Technical Field

[0004] The invention is related to an adaptive beamwidth selection technique that allows Cell-Free Massive multiple-input and multiple-output (CF-mMIMO) systems to dynamically adjust the beamwidths of their APs (Access Point), focusing on specific angles to improve angular resolution and detect multiple closely spaced targets more effectively

[0005] Prior Art

[0006] In traditional omnidirectional sensing, a wider beam is transmitted across all directions to cover a large surveillance area, which offers broad coverage but at the cost of signal strength. This approach results in a weaker Signal-to-Noise Ratio (SNR) and Signal-to-Clutter-plus-Noise Ratio (SCNR) for each target, leading to weaker target echoes that are more susceptible to interference and clutter. Furthermore, the fixed wider angular spread provides low angular resolution, making it difficult to accurately resolve multiple targets. This method is also energy inefficient, as it transmits energy uniformly across all directions, including areas with no targets. The lack of adaptability to the environment further limits its effectiveness, as it cannot dynamically adjust its beamwidth based on target locations. While it ensures broad coverage, it sacrifices precision and resolution due to the absence of directional focus, making it particularly challenging to detect closely spaced or multiple targets because of overlapping echoes and poor angular resolution.

[0007] Authors in [1] directed the beams only towards those angular directions where targets are present, but did not consider beamwidth adaptivity for finer resolution. Moreover in [2], waveform optimization at each mobile AP is performed to maximize Signal-to-Clutter-and- Noise Ratios (SCNRs). The optimization process is performed in real time but does not mention adaptive beamwidth control to enhance resolution or coverage. In [3], the authors proposed a novel solution for detecting multiple targets using distributed MIMO radar. It employs a gridbased data matching algorithm to associate the target responses to potential target locations solving the resulting data puzzle that evaluates the various cells under test (CUTs) in the surveillance area resulting from the intertwined range cells across all transmit-receive channels. Authors in [4] proposes a multisite MIMO radar system that enhances detection of fluctuating targets using digital beamforming and spatial diversity, deriving an likelihood ratio test detector and solving optimization problems related to detection performance, system configuration, and degrees of freedom, with numerical results demonstrating its effectiveness.

[0008] All the mentioned works are focused on optimizing the detection of multi targets and they did not focus on the angular resolution aspect of multi target detection while simultaneously addressing the coverage.

[0009] As a result, all of the problem mentioned above has made it necessary to provide a novelty in the related field.

[0010] Brief Description and Objects of the Invention

[0011] In accordance with embodiments, a computer-implemented adaptive beamwidth selection method is provided for multitarget detection in Cell-Free Massive MIMO systems. The method involves an initial scanning of a surveillance area using wide beams from multiple access points. Angles of target regions, which include at least one target, are detected using a direction of arrival estimation algorithm. The beamwidth of the target regions is progressively narrowed to increase angular resolution until the number of targets in the target region does not increase during a predetermined threshold of bandwidth change.

[0012] In accordance with other embodiments, the method further specifies that the direction of arrival estimation algorithm used can be either ESPIRIT or MUSIC. This provides flexibility in the choice of algorithm for detecting the angles of target regions.

[0013] In yet other embodiments, the method includes deactivating beams directed toward non-target regions, which do not include any targets, based on the direction of arrival estimation algorithms. This step ensures that resources are focused on areas with detected targets, improving the efficiency of the system.

[0014] In further embodiments, a data processing device is provided, comprising means for carrying out the steps of the method as described in any of the preceding claims. This device is equipped to implement the adaptive beamwidth selection method, ensuring that the system can effectively detect and focus on multiple targets in a Cell-Free Massive MIMO system.

[0015] The purpose of this invention is to enhance the detection accuracy, resolution, and coverage in Cell-Free Massive MIMO (CF-mMIMO) systems, specifically for multitarget detection in environments with closely spaced targets or varying clutter levels. By introducing an adaptive beamwidth selection technique, the system can dynamically adjust the beamwidth of each access point (AP), improving both angular resolution and target detection efficiency. First, the APs detects the targets using DOA (Direction of Arrival) algorithms to estimate their positions. After identifying the target regions, the system then narrows the beams in those specific directions, focusing only on the targets of interest, and avoiding energy use in areas without targets. This method ensures efficient resource usage while optimizing energy efficiency and target localization in real-time applications, making it ideal for scenarios requiring both broad area surveillance and precise target differentiation.

[0016] Thus, the method enhances both resolution and coverage in multitarget detection by leveraging the spatial diversity and distributed nature of CF-mMIMO systems. This method enables distributed APs to scan a broader geographic area, significantly improving coverage while dynamically adjusting beamwidths. The system can narrow or widen beams as needed, allowing it to focus on specific regions of interest without sacrificing overall coverage, making it especially effective for target detection in dense environments. This approach addresses the limitations of traditional fixed-width beamforming techniques, offering a scalable, high- resolution solution for multitarget detection with enhanced coverage capabilities

[0017] The method of present invention provides superior angular resolution, allowing the system to clearly resolve closely spaced targets by dynamically adjusting the beam angle and adaptively narrowing the beamwidth based on target locations. This is particularly useful in applications such as radar sensing, autonomous vehicles, and military surveillance, where accurate target distinction is critical.

[0018] The invention leverages the distributed nature of CF-mMIMO systems, where multiple access points (APs) cover a large geographical area. The adaptive beamwidth selection ensures wide area coverage without sacrificing resolution. This is crucial in scenarios where both broad surveillance and precise target detection are required, making the system more flexible and efficient in various operational environments.

[0019] The method of present invention is highly scalable, making it suitable for dense environments where many closely spaced targets are present. It optimizes the detection process by allowing APs to focus their energy on specific target regions, thus reducing interference and improving overall system performance. Furthermore, the ability to dynamically adjust the beamwidth and focus energy based on the detected targets makes the system highly adaptable to changing environments, such as those with moving targets or varying levels of clutter. This adaptability ensures consistent performance across a range of conditions.

[0020] The present invention also provides energy efficiency by concentrating energy only in such directions where potential targets are present.

[0021] In a brief summary, the invention offers an adaptive beamwidth selection for CF-mMIMO systems, allowing each access point (AP) to dynamically adjust its angle and beamwidth. This enhances target detection accuracy by focusing on specific areas of interest, achieving high angular resolution and effectively separating closely spaced targets. The system also provides wide-area coverage without compromising detection quality, as multiple APs collaborate to monitor relevant regions. By concentrating energy only on targeted areas, the system efficiently uses resources, reducing energy consumption and speeding up scanning. It’s designed for realtime adaptability in dynamic conditions and is highly scalable, making it suitable for dense environments with numerous targets.

[0022] Description of the Figures of the Invention

[0023] The figures and related descriptions necessary for the subject matter of the invention to be understood better are given below.

[0024] Figure la. A schematic view of the system during initial wide beam scanning and system during target detection and beam elimination and during narrowing beamwidth for high- resolution target detection, respectively.

[0025] Detailed Description of the Invention

[0026] The invention is related to an adaptive beamwidth selection technique that allows Cell-Free Massive multiple-input and multiple-output (CF-mMIMO) systems to dynamically adjust the beamwidths of their APs (Access Point), focusing on specific angles to improve angular resolution and detect multiple closely spaced targets more effectively

[0027] Referring to Figure 1; The first box shows APs using wide beams to cover large areas. Each AP has a broad beamwidth, indicated by the large, rounded shaded regions extending from the antennas. In this phase, the APs scan the entire region using wide beams to provide initial coverage of the surveillance area. The large shaded areas suggest that the system is in a coarse scanning phase to detect potential targets (represented by the star symbols). This step ensure broad coverage of the area, even though the resolution is relatively low. The system can detect targets but cannot yet resolve closely spaced objects.

[0028] In the second box, some of the beams are eliminated or turned off. The "X" symbols indicate regions where no targets were detected, and the corresponding beams from those APs are deactivated. The remaining beams continue to focus on areas where targets have been detected (represented by the star symbols). At this stage, the system narrows its focus by eliminating the beams in areas where no targets were found, improving efficiency and reducing unnecessary resource use. The APs now concentrate on scanning the regions of interest where the initial wide-beam scan detected potential targets. This step refines the scanning process by reducing the active areas to only those with detected targets, leading to more target-specific focus.

[0029] The final box of Fig. 1 shows each AP using narrow beams to focus on the regions where targets were previously detected. The beams are significantly more focused, indicated by the smaller, thinner shaded areas directed toward the stars (targets). The system now uses narrower beamwidths to increase the angular resolution. This narrowing of beams allows the APs to sharply focus on the detected targets, improving the system’s ability to resolve closely spaced targets and enhancing detection accuracy. This phase represents the high-resolution stage, where the beams are fine-tuned to pinpoint the exact locations of the targets with high accuracy. By progressively narrowing the beamwidth, the system ensures that each target is detected and localized with precision.

[0030] According to the method of the present invention, the first step involves the initial scanning of the entire surveillance area using wide beams. Each AP in the Cell-Free Massive MIMO (CF- mMIMO) system uses a broader beam to cover a larger portion of the angular space. This ensures that the system performs a coarse scan of the environment, allowing it to detect multiple targets across a large geographic area, even though the initial resolution is relatively low. At this stage, the system sacrifices angular resolution for coverage, ensuring that all potential target regions are covered before narrowing the beams for higher resolution detection.

[0031] Once the wide beams have scanned the environment, the system utilizes DOA (Direction of Arrival) estimation algorithms. Preferably, DOA estimation algorithm is selected between MUSIC [5], [6], and [7] or ESPRIT [8] and [9] to estimate the locations of the targets based on the signals received by the APs. DOA algorithms provide an initial estimate of the angles at which the targets are located. This allows the system to identify the general direction of the targets, which is crucial for subsequent beamwidth adjustment. At this stage, the system processes the signals received by each AP and calculates the spatial pseudospectrum, where peaks indicate the presence of targets. The algorithm uses this data to determine the angular bins where the targets may be present.

[0032] Preferably, after DOA estimation, the beams in the direction where there is no target are deactivated. This means that only the beams directed towards angular regions with potential targets are retained. This also increase energy efficiency. This step helps to reduce the scanning space and focus the system's energy and resources on the regions of interest, rather than continuing to scan areas where no targets were detected. The angular bins where no significant peaks were observed in the DOA estimation are eliminated. This process ensures more efficient use of resources by discarding unnecessary scanning in empty regions.

[0033] Once the angular bins corresponding to the target(s) are estimated, the beams directed towards other angular sectors can be eliminated. Mathematically, let ©target represent the set of estimated target angles obtained from the DOA algorithm. The scanning region 0 = [0, 7t] can then be reduced to the subset ©target 0, where:

[0034] ©target—{ 0i G 0 | P(0i) > y }

[0035] Here, P(0i) is the pseudospectrum value at angle 0t, and y is the detection threshold. The beams directed towards angles outside ©target can be eliminated, reducing the set of scanning angles to only those that correspond to potential targets:

[0036] ©new=©target

[0037] This step ensures that the radar system focuses its energy on the angles where targets are likely to be located, eliminating the need to scan other directions.

[0038] After that beamwidth are narrowed for higher resolution. In this step, the beamwidth for each AP is progressively narrowed down for the angular regions where targets were detected. The goal is to increase the resolution by focusing more energy on the specific angular bins where the targets are located. Narrowing the beamwidth allows the system to sharpen the resolution, enabling the detection of closely spaced targets that may have been indistinguishable during the initial wide beam scanning phase. The system dynamically adjusts the beamwidth, ensuring that the beams are narrow enough to detect and resolve multiple targets accurately. The more precise the beam, the better the system can distinguish between targets located close to each other in angular space.

[0039] The final step involves selecting the optimal beamwidth for each target based on the convergence of the number of detected targets. As the beamwidth is narrowed, the system monitors the number of targets detected. Once the number remains constant over a range of beamwidths, the optimal beamwidth is selected for that target. In clearer way, the system progressively narrows the beam until it reaches a point where further narrowing does not reveal additional targets. This convergence point is considered the optimal beamwidth, providing the system with the highest resolution for target detection while maintaining efficiency. This ensures that the system achieves the best possible resolution for each detected target without over-narrowing the beam, which could lead to unnecessary energy consumption or reduced coverage.

[0040] To increase the resolution for each detected target, the beamwidth is progressively narrowed down to focus on the angular bins where the targets are located.

[0041] The beamwidth narrowing continues until the number of detected targets remains constant over a range of beamwidths, i.e., the detection no longer improves as the beamwidth is narrowed. Mathematically, we set a threshold c for convergence such that:

[0042] | Aft+i - Aft | < e

[0043] Where e is a small threshold indicating that the change in beamwidth is sufficiently small, meaning no further targets are detected or the resolution no longer improves.

[0044] At convergence, the beamwidth is considered optimal, denoted as 0opt. The optimal beamwidth satisfies the condition: dopt ~ Omin REFERENCES

[0045] [1] Ahmed, Aya Mostafa, et al. "A reinforcement learning based approach for multitarget detection in massive MIMO radar." IEEE Transactions on Aerospace and Electronic Systems 57.5 (2021): 2622-2636.

[0046] [2] P. Chen, L. Zheng, X. Wang, H. Li and L. Wu, "Moving Target Detection Using Colocated MIMO Radar on Multiple Distributed Moving Platforms," in IEEE Transactions on Signal Processing, vol. 65, no. 17, pp. 4670-4683, 1 Sept. l, 2017, doi: 10.1109 / TSP.2017.2714999

[0047] [3] S. Yang, W. Yi and A. Jakobsson, "Multitarget Detection Strategy for Distributed MIMO Radar With Widely Separated Antennas," in IEEE Transactions on Geoscience and Remote Sensing, vol. 60, pp. 1-16, 2022, Art no. 5113516, doi: 10.1109 / TGRS.2022.3175046.

[0048] [4] J. Xu, X. -Z. Dai, X. -G. Xia, L. -B. Wang, J. Yu and Y. -N. Peng, "Optimizations of Multisite Radar System with MIMO Radars for Target Detection," in IEEE Transactions on Aerospace and Electronic Systems, vol. 47, no. 4, pp. 2329-2343, OCTOBER 2011, doi: 10.1109 / TAES.2011.6034636.

[0049] [5] Ahmad, Mushtaq, et al. "Low-Complexity 2D-DOD and 2D-DOA Estimation in Bistatic MIMO Radar Systems: A Reduced-Dimension MUSIC Algorithm Approach." Sensors 24.9 (2024): 2801.

[0050] [6] Deng, Zipeng, et al. "TO A and DO A Estimation of Mimo OFDM Signals for Uniformlinear Array Radar Based on Music." IGARSS 2024-2024 IEEE International Geoscience and Remote Sensing Symposium. IEEE, 2024.

[0051] [7] Ahmad, Mushtaq, et al. "Enhanced angle estimation in MIMO radar: Combine RD-MUSIC and SDP optimization." AEU-Intemational Journal of Electronics and Communications 178 (2024): 155235.

[0052] [8] Lv, Liujie, et al. "Dual-Stream Reconstruction Network-aided ESPRIT Algorithm for DoA Estimation in Coherent Scenarios." IEEE Transactions on Vehicular Technology (2024).

[0053] [9] Li, Shujie, et al. "Multi-Target Pairing Method Based on PM-ESPRIT-like DOA Estimation for T / RR HFSWR." Remote Sensing 16.17 (2024): 3128.

Claims

CLAIMS1. A computer-implemented adaptive beamwidth selection method for multitarget detection in Cell-Free Massive MIMO systems characterized by; a) Initial scanning of a surveillance area using wide beams from multiple access points, b) Detecting angles of target regions, where include at least target, by a direction of arrival estimation algorithm, c) Progressively narrowing of beamwidth of target regions to increase angular resolution until number of the target in the target region don’t increase during the predetermined threshold of bandwidth change.

2. A method according to Claim 1, wherein the direction of arrival estimation algorithm is ESPIRIT or MUSIC.

3. A method according to Claim 1, characterized by deactivating beams directed toward non-target regions, where does not include any target, according to direction of arrival estimation algorithms.

4. A data processing device comprising means for carrying out the steps of the method of any of preceding claims.

5. A computer program comprising instructions which, when the program is executed by the data processing device of claim 4, cause the data processing device to carry out the method of any of claim 1 to 3.

6. A computer-readable medium comprising instructions which, when executed by the data processing device of claim 4, cause the data processing device of claim 4 to carry out the method of any of claim 1 to 3.

Images