Search

Acquiring accurate channel state information (CSI) at an access point (AP) is challenging for wideband millimeter wave (mmWave) ultra-massive multiple-input and multiple-output (UM-MIMO) systems, due to the high-dimensional channel matrices, hybrid near- and far- field channel behavior, beam squint effects, and imperfect hardware constraints, such as low-resolution analog-to-digital converters, and in-phase and quadrature imbalance. To overcome these challenges, this paper proposes an efficient downlink channel estimation (CE) and CSI feedback approach based on knowledge and data dual-driven deep learning (DL) networks. Specifically, we first propose a data-driven residual neural network de-quantizer (ResNet-DQ) to pre-process received pilot signals at user equipment (UEs), where the noise and distortion brought by imperfect hardware can be mitigated. A knowledge-driven generalized multiple measurement vector learned approximate message passing (GMMV-LAMP) network is then developed to jointly estimate the channels by exploiting the approximately same physical angle shared by different subcarriers. In particular, two wideband redundant dictionaries (WRDs) are proposed such that the measurement matrices of the GMMV-LAMP network can accommodate the far-field and near-field beam squint effect, respectively. Finally, we propose an encoder at the UEs and a decoder at the AP by a data-driven CSI residual network (CSI-ResNet) to compress the CSI matrix into a low-dimensional quantized bit vector for feedback, thereby reducing the feedback overhead substantially. Simulation results show that the proposed knowledge and data dual-driven approach outperforms conventional downlink CE and CSI feedback methods, especially in the case of low signal-to-noise ratios.

The neocortex contains a vast collection of diverse neurons organized into distinct layers. While nearly all neocortical neurons are generated by radial glial progenitors (RGPs), it remains largely unclear how a complex yet organized neocortex is constructed reliably and robustly. Here, we show that the division behavior and neuronal output of RGPs are highly constrained with patterned variabilities to support the reliable and robust construction of the mouse neocortex. The neurogenic process of RGPs can be well-approximated by a consistent Poisson-like process unfolding over time, producing deep to superficial layer neurons progressively. The exact neuronal outputs regarding layer occupation are variable; yet, this variability is constrained systematically to support all layer formation, largely reflecting the variable intermediate progenitor generation and RGP neurogenic entry and exit timing differences. Together, these results define the fundamental features of neocortical neurogenesis with a balanced reliability and variability for the construction of the complex neocortex.This study conducted a systematic quantitative analysis of the number and laminar distribution of the neuronal output by individual progenitors, revealing an intricate balance between reliability and variability for the effective construction of the complex neocortex.