We propose a general framework for solving statistical mechanics of systems with finite size. The approach extends the celebrated variational mean-field approaches using autoregressive neural ...networks, which support direct sampling and exact calculation of normalized probability of configurations. It computes variational free energy, estimates physical quantities such as entropy, magnetizations and correlations, and generates uncorrelated samples all at once. Training of the network employs the policy gradient approach in reinforcement learning, which unbiasedly estimates the gradient of variational parameters. We apply our approach to several classic systems, including 2D Ising models, the Hopfield model, the Sherrington-Kirkpatrick model, and the inverse Ising model, for demonstrating its advantages over existing variational mean-field methods. Our approach sheds light on solving statistical physics problems using modern deep generative neural networks.
This paper investigates optical orthogonal frequency division multiplexing (O-OFDM) systems with index modulation (IM) based on discrete Hartley transform (DHT) for visible light communications ...(VLC). An interesting trade-off between spectral efficiency (SE), energy efficiency, peak-to-average power ratio (PAPR) and bit error rate (BER) performance can be made by using the IM technique. And the DHT-based O-OFDM-IM systems can achieve higher SE compared with the DFT (discrete Fourier transform)-based counterparts. Employing DHT instead of DFT enables the removal of the Hermitian symmetry requirement, which contributes to transmitting more index bits and results in SE improvement. To acquire the same SE, the proposed systems can significantly lower the constellation order leading to better BER performance by contrast with the DFT-based counterparts. We employ DC (direct current) biasing and AC (Asymmetrically clipped) techniques to investigate the O-OFDM-IM systems in this paper. Simulation results indicate that SNR gains can be obtained by the DHT-based DCO/ACO-IM in comparison to the DFT-based counterparts. We also propose a novel reduced-complexity ML (maximum likelihood) detector specifically applicable to VLC. The new detector exhibits the same BER performance as the ML detector and performs better than the LLR (log-likelihood ratio) detector with reduced computational complexity.
This paper investigates a point-to-point millimeter-wave (mmWave) massive MIMO system in which the transmitter and receiver consist of Kt and K, distributed antenna subarrays, respectively. All the ...subchannel matrices between each pair of subarrays at the transmitter and the receiver are assumed to be mutually independent and known at the transmitter and the receiver. To quantify the potential of spatial multiplexing in such a system, the multiplexing gain is defined and analyzed when the numbers of antennas at subarrays go to infinity while the transmit power is held fixed. In particular, assuming that all subchannels have the same average number of propagation paths L̅, the paper shows that an average maximum multiplexing gain of K r ,K t L̅ can be achieved. This result means that the multiplexing gain under a distributed array architecture can statistically increase in proportion with K r ,K t . The multiplexing gain analysis is also extended to the multiuser scenario. Simulation results obtained with the hybrid analog/digital processing finally corroborate the analytical results.
In this article, we investigate the downlink (DL) of a cell-free massive multiple-input multiple-output (MIMO) system over spatially correlated Rician fading, in which many distributed access points ...(APs) equipped with multiple antennas serve single-antenna users. The APs apply minimum mean-square error channel estimation to obtain the uplink channel state information (CSI). Furthermore, in order to obtain DL CSI at users, this article considers the use of maximum-ratio transmission to beamform DL pilots in the DL beamforming training (BT) phase. For such a system, we derive the closed-form expressions of the sum spectral efficiency (SE) and total energy efficiency (EE). Based on the obtained closed-form expressions, we develop two successive approximation algorithms to improve the sum SE and total EE by optimizing the power control coefficients of DL data and pilot. Numerical results are provided to demonstrate the superiority of the proposed algorithms in improving the sum SE and total EE. In addition, the numerical results also show that the sum SE of a cell-free massive MIMO system with exploiting the BT scheme can be significantly improved over the system without employing the BT scheme.
Direction of arrival (DOA) plays an essential role in wireless communications, which has been widely utilized in localization and intelligent navigation. Recently, with explosive growth of wireless ...communication devices, the electromagnetic environment has become more and more complicated. In such electromagnetic environment, it remains a challenge to achieve robust DOA estimation while maximizing source discrimination with limited physical sensors. Additionally, conventional DOA methods are usually based on Gaussian noise assumption, while these methods would suffer from performance degradation in impulsive noise. Aiming at addressing these issues, we propose a novel robust DOA estimation method with high degree of freedom (DOF). In the proposed method, firstly, a novel mirror coprime array (MCA) structure is proposed for DOA estimation. Relying on the properties of non-circular signals, MCA can make full use of the received signals from limited physical elements to accurately estimate more DOAs with high DOF. Next, a novel square-based correntropy operator (SCO) is defined to effectively suppress impulsive noise in complicated electromagnetic environment. Also, we demonstrate the effectiveness of SCO in impulsive noise suppression theoretically. Further, accurate and robust DOA estimation with high DOF can be realized through the proposed MCA-SCO-MUSIC method, which innovatively fuses the pseudo-covariance matrix generated by SCO into MCA based DOA estimation method. Finally, we conduct a series experiments to evaluate the performance of proposed method, and simulation results demonstrate its effectiveness in enhancing DOF without adding physical array elements. Moreover, the proposed method exhibits better performance compared to contrastive methods in impulsive noise.
Quantum computers promise to perform certain tasks that are believed to be intractable to classical computers. Boson sampling is such a task and is considered a strong candidate to demonstrate the ...quantum computational advantage. We performed Gaussian boson sampling by sending 50 indistinguishable single-mode squeezed states into a 100-mode ultralow-loss interferometer with full connectivity and random matrix-the whole optical setup is phase-locked-and sampling the output using 100 high-efficiency single-photon detectors. The obtained samples were validated against plausible hypotheses exploiting thermal states, distinguishable photons, and uniform distribution. The photonic quantum computer,
, generates up to 76 output photon clicks, which yields an output state-space dimension of 10
and a sampling rate that is faster than using the state-of-the-art simulation strategy and supercomputers by a factor of ~10
.
The reconfigurable intelligent surface (RIS) technology has generated considerable interest due to its advantages of low cost, easy deployment, and high controllability. In this article, a ...RIS-assisted space-time block code (STBC) scheme and a RIS-assisted differential space-time block code (DSTBC) scheme with a single radio frequency signal generator at the transmitter side are proposed to reduce the hardware overhead of classical STBC and the channel state information (CSI) acquisition at the receiver side. RIS is utilized to enhance symbol error probability (SEP) performance and reduce the number of antennas of the classical STBC and DSTBC schemes. After that, we analyze the SER performance of the two schemes, emphasizing the diversity gain, and find that increasing only the number of divided parts of RIS for the two schemes could enhance the diversity gain without adding extra hardware. Monte Carlo simulations are provided to show the effectiveness of our proposed schemes and it has been shown that the two proposed schemes outperform the reference schemes in terms of SER performance.
With the in-depth exploration of the ocean, the Internet of Underwater Things (IoUT) technology has attracted increasing attention. Underwater wireless optical communication (UWOC) provides the ...possibility of massive data transmission for IoUT. The underwater optical link of UWOC can be interrupted by the obstruction of marine animals and plants, seamounts, and some underwater equipment. In this paper, we present a reconfigurable intelligent surface (RIS) assisted UWOC system to solve the problem of link occlusion. The cascaded turbulence channel fading coefficients from source to destination through RIS are modeled as Gamma-Gamma distribution, and the pointing errors caused by beam jitter and RIS jitter are considered. In the RIS-assisted UWOC systems with intensity modulation and direct detection (IM-DD), we apply the approximate distribution of the sum of Gamma-Gamma random variables to derive the probability density function (PDF) of the instantaneous received signal-to-noise ratio (SNR) for the first time. Based on the PDF, the novel closed-form expression of outage probability is given, and the outage probability analysis results show that when the number of reflecting elements is very large (i.e., tends to infinity), the outage probability asymptotically approaches zero. Subsequently, the closed-form expressions of the finite-SNR diversity order, the asymptotical diversity order, and the convergence speed of finite-SNR diversity order to asymptotical diversity order are proposed. Additionally, the ergodic channel capacity expression is investigated. Finally, we give simulation results to validate our derived results and analyze the impacts of the number of reflecting elements of RIS and the pointing errors on the above system performance indicators.
We report phase-programmable Gaussian boson sampling (GBS) which produces up to 113 photon detection events out of a 144-mode photonic circuit. A new high-brightness and scalable quantum light source ...is developed, exploring the idea of stimulated emission of squeezed photons, which has simultaneously near-unity purity and efficiency. This GBS is programmable by tuning the phase of the input squeezed states. The obtained samples are efficiently validated by inferring from computationally friendly subsystems, which rules out hypotheses including distinguishable photons and thermal states. We show that our GBS experiment passes a nonclassicality test based on inequality constraints, and we reveal nontrivial genuine high-order correlations in the GBS samples, which are evidence of robustness against possible classical simulation schemes. This photonic quantum computer, Jiuzhang 2.0, yields a Hilbert space dimension up to ∼ 1043, and a sampling rate ∼ 1024 faster than using brute-force simulation on classical supercomputers.
This paper considers a downlink distributed millimeter wave massive multi-input multi-output (D-mmMIMO) system, consisting of multiple randomly distributed radio access units (RAUs), the distribution ...of which is modeled as either a Poisson point process (PPP) or a Matérn hard-core point process (MHCPP). Compared with the PPP model where RAUs are randomly distributed without restriction, the MHCPP model constructed from a PPP by dependently thinning is more realistic, since it imposes a minimum distance between two RAUs so that they are not too close to each other. When a hybrid precoding algorithm based on antenna array response vectors is adopted, the upper bounds on average spectral efficiencies of the UEs for both the PPP D-mmMIMO system and MHCPP D-mmMIMO system are derived. Finally, simulation results are provided to demonstrate the validity of the developed analysis. The presented results also show that, when the MHCPP is generated from the PPP used in the PPP D-mmMIMO system, the MHCPP D-mmMIMO system with a low hard-core distance achieves a quite close performance to the PPP D-mmMIMO system, and increasing hard-core distance has a negative impact on spectral efficiency. In contrast, when the MHCPP is constructed from a different PPP, which system performs better depends on the intensities of two PPPs as well as hard core distance.