23 April 2019 in IEEE Transactions on Signal Processing
In the present paper, we propose a novel sparse signal recovery algorithm called the Trainable Iterative Soft Thresholding Algorithm (TISTA). The proposed algorithm consists of two estimation units: a linear estimation unit and a minimum mean squared error (MMSE) estimator-based shrinkage unit. The error variance required in the MMSE shrinkage unit is precisely estimated from a tentative estimate of the original signal. The remarkable feature of the proposed scheme is that TISTA includes adjustable variables that control step size and the error variance for the MMSE shrinkage function. The variables are adjusted by standard deep learning techniques. The number of trainable variables of TISTA is nearly equal to the number of iteration rounds and is much smaller than that of known learnable sparse signal recovery algorithms. This feature leads to highly stable and fast training processes of TISTA. Computer experiments show that TISTA is applicable to various classes of sensing matrices such as Gaussian matrices, binary matrices, and matrices with large condition numbers. Numerical results also demonstrate that, in many cases, TISTA provides significantly faster convergence than AMP and the Learned ISTA and also outperforms OAMP in the NMSE performance.
11 July 2019 in IEEE Access
This paper presents a deep learning-aided iterative detection algorithm for massive overloaded multiple-input multiple-output (MIMO) systems where the number of transmit antennas $n$ is larger than that of receive antennas $m$ . Since the proposed algorithm is based on the projected gradient descent method with trainable parameters, it is named the trainable projected gradient-detector (TPG-detector). The trainable internal parameters, such as the step-size parameter, can be optimized with standard deep learning techniques, i.e., the back propagation and stochastic gradient descent algorithms. This approach is referred to as data-driven tuning and ensures fast convergence during parameter estimation in the proposed scheme. The TPG-detector mainly consists of matrix-vector product operations whose computational cost is proportional to $m n$ for each iteration. In addition, the number of trainable parameters in the TPG-detector is independent of the number of antennas. These features of the TPG-detector result in a fast and stable training process and reasonable scalability for large systems. The numerical simulations show that the proposed detector achieves a comparable detection performance to those of the existing algorithms for the massively overloaded MIMO channels, e.g., the state-of-the-art IW-SOAV detector, with a lower computation cost.
22 April 2019 in IEEE Wireless Communications Letters
In this letter, we propose an approximation theory for connectivity of general random geometric graphs (RGGs) with stochastic node faults, which is regarded as an abstract model of ad hoc wireless sensor networks (WSNs) with unreliable nodes. In the model, each node in an RGG is independently at fault and the disconnection probability of a resulting survival network is evaluated. The proposed approximation formula for the probability is asymptotically correct and reveals the phase transition phenomena of connectivity in the model. In addition, we obtain phase transition thresholds of a practical Rayleigh single-input single-output (SISO) model and Rayleigh multiple-input multiple-output (MIMO) model, which shows the effectiveness of the MIMO system against attenuation and random node faults. The numerical results in Rayleigh MIMO model show that the approximation formula well estimates the connection probability even for finite graphs and that it provides useful information for design of WSN immune to node faults.
6 December 2023 in IEEE Wireless Communications Letters
Recently, various multiple-input multiple-output (MIMO) signal detectors based on deep learning have been proposed. Especially, deep unfolding (DU), which involves unrolling of an existing iterative algorithm and embedding of trainable parameters, has been applied with remarkable detection performance. In this study, we attempted to construct a simple DU-based trainable MIMO detector with the simplest structure without using matrix inversion. The proposed detector based on the Hubbard–Stratonovich (HS) transformation and DU is called the trainable HS (THS) detector. It requires a constant number of trainable parameters and only quadratic training and execution costs per iteration. Numerical results show that the detection performance of the THS detector is better than that of existing algorithms of the same complexity and close to that of a DU-based detector, which has higher training and execution costs than the THS detector.
15 June 2024 in Journal of the Physical Society of Japan
14 April 2020 in IEEE Transactions on Communications
Compute-and-forward (CAF) relaying is an effective way to increase bandwidth efficiency of wireless two-way relay channels. Design of error-correcting codes and their decoding algorithms suitable for CAF relaying schemes remains an important issue to be studied. In this paper, we will analyze an asymptotic behavior of LDPC codes over two-way relay channels based on density evolution (DE). Because of the asymmetric characteristics of the channel, we use the population dynamics DE combined with DE formulas for asymmetric channels to obtain the belief propagation (BP) thresholds. Additionally, we also evaluate the asymptotic performance of spatially coupled LDPC codes for two-way relay channels. The results indicate that the spatially coupled codes yield improvements in the BP threshold compared with corresponding uncoupled codes for two-way relay channels.