Off-grid direction of arrival (OGDOA) estimation methods deal with the situations where true direction of arrivals (DOAs) are not on the discretized sampling grid. However, existing OGDOA estimation methods are faced with a tradeoff between density of initial grid and computational workload. Furthermore, these methods fail if more than one true DOA is located in a same grid interval. In order to speed up the OGDOA estimation methods and solve the problem of more than one DOA in one initial grid interval, a grid evolution direction of arrival (GEDOA) estimation method is proposed. Based on the combination of OGDOA estimation methods and grid refinement, this new approach makes the grid nonuniformly evolve from coarse to dense. The proposed method contains two subprocesses, i.e., learning process and fission process. The learning process aims to update locations of grid points and estimate DOAs. The fission process aims to generate new grid points and guarantees that there is only one DOA in a grid interval. The two processes iterate alternatively. Finally, an adaptive and nonuniform grid and an estimated spatial spectrum based on this grid are achieved. Compared with the previous methods, GEDOA has better computational efficiency because fewer grid points are used at each iteration. Furthermore, GEDOA has better resolution and lower MSE at relative high SNR. This is because that the evolved grid obtained by this new approach is information adaptive. Numerical simulations validate the effectiveness of the proposed method.

Orthogonal time-frequency space (OTFS) is a promising technique for high mobility wireless communications, as it approaches a sparse representation of doubly-selective fading channel and can obtain delay-Doppler (DD) diversity gain. However, in practical OTFS systems, fractional channel parameter is a great challenge and will seriously deteriorates the sparsity and the performance of channel estimation. In this paper, we propose a novel channel estimation scheme for doubly selective channel in the presence of both fractional Doppler and fractional delay. First, we analyze the input-output relationship of single-input single-output (SISO) OTFS system based on the integer sampling DD grid, then extend it to the fractional DD grid. Compared with the integer DD domain, the model in the fractional DD domain is sparser and more accurate. Based on our analysis, two kinds of off-grid models, i.e., doubly fractional model and mixed one- and two-dimensional (1&2D) fractional models are proposed and compared. The mixed 1&2D fractional model has an advantage of low complexity, but encounters a tandem off-grid distortion. To tackle this distortion problem, as well as to balance the accuracy and the computational workload, a novel two dimensional off-grid decomposition and combination scheme is proposed based on the doubly and mixed 1&2D fractional model. Simulation results demonstrate the effectiveness and superiority of the proposed channel estimation schemes, compared with the existing work for OTFS systems under doubly selective channels.

Orthogonal time-frequency space (OTFS) modulation is a promising modulation technology, which can provide superior performance under high-mobility radio scenarios. However, fractional Doppler shift brings inter-Doppler interference (IDI), leading to performance deterioration of channel estimation. This paper proposes a two-stage algorithm for estimating the fractional Doppler channels. In the first stage, a rough position of each nonzero Doppler shift is estimated through the correlation of the basis function in the delay-Doppler (DD) domain. In the second stage, a quasi-Newton method is employed to iteratively re-estimate the fractional Doppler channel based on the prior estimation, thus resulting in significantly reduced estimation error caused by the IDI. Simulation results demonstrate that the proposed algorithm for channel estimation can achieve a much lower normalized mean square error (NMSE) than the baseline algorithms. Moreover, the achieved bit error rate (BER) is very close to that of perfect channel estimation.

Indoor fires pose significant threats in terms of casualties and economic losses globally. Thus, it is vital to accurately detect indoor fires at an early stage. To improve the accuracy of indoor fire detection for the resource-constrained embedded platform, an indoor fire detection method based on multi-sensor fusion and a lightweight convolutional neural network (CNN) is proposed. Firstly, the Savitzky–Golay (SG) filter is used to clean the three types of heterogeneous sensor data, then the cleaned sensor data are transformed by means of the Gramian Angular Field (GAF) method into matrices, which are finally integrated into a three-dimensional matrix. This preprocessing stage will preserve temporal dependency and enlarge the characteristics of time-series data. Therefore, we could reduce the number of blocks, channels and layers in the network, leading to a lightweight CNN for indoor fire detection. Furthermore, we use the Fire Dynamic Simulator (FDS) to simulate data for the training stage, enhancing the robustness of the network. The fire detection performance of the proposed method is verified through an experiment. It was found that the proposed method achieved an impressive accuracy of 99.1%, while the number of CNN parameters and the amount of computation is still small, which is more suitable for the resource-constrained embedded platform of an indoor fire detection system.

Orthogonal time frequency space (OTFS) is a novel modulation scheme to handle the high Doppler effect under time-varying channels. In this paper, in order to improve channel estimation accuracy and to reduce pilot overhead, two types of two-dimensional (2D) pilots and the corresponding matched filters are designed for multi-antenna OTFS systems. Our 2D pilots are formed using perfect arrays (such as Frank array, Chu array, etc) or perfect-sequence based Kronecker array (PKA). Different from the previous multi-antenna OTFS pilots, these 2D pilots are placed on the same area in the delay-Doppler domain, using code division multiplexing to deal with the interference between pilots of different antennas, known as pilot pollution. To improve the channel estimation performance, matched filters are also designed to better compensate the phase shift of the 2D pilot response in the delay-Doppler domain. Compared with the conventional multi-antenna OTFS schemes, the proposed scheme achieves significantly better NMSE performance, while having lower pilot overhead under multi-antenna scenarios.

Recently, atomic norm minimization (ANM) has been applied in the reconfigurable intelligent surface (RIS) aided direction of arrival (DOA) estimation problem in the non-line-of-sight (NLOS) propagation. However, most existing RIS-aided ANM (RIS-ANM) methods assume a uniform linear array (ULA) RIS which can only handle the one-dimensional (1D) DOA estimation. To this end, a novel RIS-aided gridless two-dimensional (2D) DOA estimation method is developed for a uniform planar array (UPA) RIS passive sensing system, where a general single-input multi-output (SIMO) transmission model is adopted. Specifically, a RIS-aided decoupled ANM (RIS-D-ANM) is proposed for gridless 2D-DOA retrieval. We show that the proposed RIS-D-ANM provides a decoupled formulation of the atomic norm instead of reshaping the 2D information via vectorization in the classic RIS-ANM method. Thus, the computational complexity of the original high-dimensional SDP problem is remarkably reduced without performance degradation. Furthermore, the computational cost is further reduced by integrating the alternating direction method of multiplier (ADMM), resulting in the RIS-aided decoupled ADMM (RIS-D-ADMM) method. Simulation results demonstrate the effectiveness of the proposed method.