To improve the informatization construction and intelligent decision-making level of river and lake basin management, the water quality of a digital twin basin was considered as the starting point and a water quality evaluation platform for Chuancheng River and Baihe River in Nanyang City, Henan Province was established. Based on digital twin technology, the platform establishes a virtual space city model, uses the long short-term memory algorithm to establish a water quality prediction model, draws the distribution of water pollution factors in two dimensions based on Kriging interpolation, simulates the pollutant diffusion in three dimensions based on numerical simulation, and finally builds a visual platform for evaluation and analysis. The platform combines digital twin with three models: one-dimensional (1D) water quality data processing, two-dimensional pollutant distribution, and three-dimensional (3D) pollutant diffusion simulation to achieve visual and comprehensive management of water quality assessment. Compared with the traditional 1D water quality data management platform, the proposed digital twin 3D urban river water quality evaluation platform system solves the problems of low visualization degree, single management, and incomplete analysis, as well as provides a new technical guarantee for the management of urban river water quality.

For emerging 5G fronthaul, broadband radio over fiber (RoF) transmission has been found much more cost-effective than conventional digital fiber transmission. Current digital predistortion (DPD) techniques may not be effective in linearizing broadband RoF links. This work presents a simple envelope-assisted radio frequency (RF)/intermediate frequency (IF) DPD model for linearization of broadband/multiband RoF transmission. The model uses both RF/IF signals and baseband envelopes, resulting in the short-term and long-term memory effect mitigated efficiently even for high signal bandwidth. The model complexity is lower than the existing baseband DPDs and does not increase with the increase of the signal bands. The model is experimentally evaluated in three scenarios: two 20 MHz long-term evolution (LTE) signals spaced by 100 and 40 MHz, and three 20 MHz LTE signals spaced by 50 MHz. It is shown that the proposed DPD results in comparable performance as today's baseband models.

Unsupervised salient object detection is an important task in many real-world scenarios where pixel-wise label information is of scarce availability. Despite its significance, this problem remains rarely explored, with a few works that consider unsupervised salient object detection methods based on the fused graph from the sum fusion of multiple deep feature similarity matrices. However, these methods ignore the interrelation of the low-level feature similarity matrices and the high-level semantic similarity matrice, which degrades the quality of the fused graph. In this article, we propose a semantic-consistency-guided multi-graph fusion learning algorithm for unsupervised saliency detection, where the consistency and inconsistency between multiple low-level feature similarity matrices and the high-level semantic similarity matrice are explored to promote the robustness and quality of the fused graph. In the first stage, a semantic-consistency-guided multi-graph fusion learning method is proposed to exploit consistency and inconsistency of multiple low-level deep features and the high-level semantic feature. The semantic-consistency-guided similarity matrices are computed for preliminary saliency ranking. In the following saliency refinement stage, the semantic-enhanced similarity matrices are built by the cross diffusion to fuse the multiple low-level deep features and the high semantic deep feature. Based on the semantic-enhanced similarity matrices, the refinement saliency maps are calculated in a semantic-enhanced cellular automata manner. Furthermore, the final ensemble stage of the large margin semi-supervised classification views the preliminary ranking results and refinement results as features, adopts the large margin graphs for saliency ensemble. Extensive evaluations over four benchmark datasets show that the proposed unsupervised method performs favorably against the state-of-the-art approaches and is competitive with some supervised deep learning-based methods.

To fulfill the connectivity demands in massive machine-type communications (mMTC), this paper investigates a generalized multi-user sparse superposition transmission (GMUSST) technology based on position index modulation. Due to the high computation complexity of maximum likelihood (ML) multi-user detection, a low complexity multi-path successive interference cancellation (MSIC) multi-user detector is introduced to achieve near-ML detector's block error ratio (BLER) performance. Furthermore, considering that each user is only concerned with their own transmitted signal in the downlink GMUSST system, we propose a minimum mean square error-based SIC (MMSE-SIC) detector, which can directly extract the user's transmission signal from the received superimposed signal of multiple users and is verified compared with MSIC detector. Simulation results show that the GMUSST can achieve better transmission reliability than the existing polar coded sparse code multiple access (PC-SCMA) in the short packet communication scenarios. Especially with the hybrid automatic repeat request mechanism, GMUSST requires fewer retransmissions to achieve the same BLER performance compared to PC-SCMA.