In fiber-terahertz integrated communication systems, nonlinear distortion and intersymbol interference (ISI) will degrade transmission performance. Pre-compensation is an efficient method to handle the channel distortion as it can avoid noise boosting during channel compensation and reduce receiver side signal processing algorithmic complexity at user-end (UE) considering the asymmetric access scenario. In this paper, we propose and experimentally demonstrate a neural-network (NN)-based carrier-less amplitude phase (CAP) modulated signal generation and end-to-end optimization method for a fiber-terahertz integrated communication system. The CAP signal is generated directly from quadrature amplitude modulation symbols and pre-compensated through a transmitter NN, which allows the receiver to demodulate the signal with simple linear digital signal process (DSP). In generating the CAP signal, the NN based transmitter learns a group of filters, which can generate, up-convert, and pre-compensate the signals. Based on the proposed method, a fiber-terahertz integration access system at 220 GHz is demonstrated and a sensitivity gain of 1.2 dB is achieved at a transmission speed of 50 Gbps and the forward error correction (FEC) bit error rate (BER) threshold of 1 x 10-2 compared with the baseline after 10-km fiber transmission and 1-m wireless delivering.

The integration of fiber and millimeter-wave (MMW) technology offers a promising solution for next-generation (6G) communication. However, the experimental investigation of multi-user fiber-MMW integrated communication with diverse channels is yet to be explored, as the complex and dynamic nature of this system makes it challenging to overcome the channel distortion. In this paper, we propose and demonstrate an adaptive multi-user end-to-end framework (AMEF) for symbol-to-symbol optimization of a multi-user fiber-MMW integrated system. This framework leverages a two-tributary heterogeneous neural network (TTHnet) based multi-channel model (MCM) and a multi-user transceiver (MUT) composed of a multi-user encoder (MUE) and a multi-user decoder (MUD). The multi-layer perceptron (MLP)-based MUE and MUD are jointly optimized as an auto-encoder, facilitated by the well-trained MCM, which enables gradient back-propagation and end-to-end optimization. The weights of different users are adaptively allocated by the AMEF to ensure a balanced transmission performance. We experimentally demonstrate the effectiveness of the AMEF through a 10 km fiber-MMW integrated system for two-user communication. With the proposed method, we realize a 66 Gbps wireless-only transmission and a 49.5 Gbps fiber-wireless transmission under a soft decision-forward error correction (SD-FEC) threshold. Compared to the conventional multiband carrierless amplitude and phase (mCAP) modulation scheme, our proposed AMEF achieves significant receiver sensitivity gains exceeding 1.1 dB and 0.6 dB for the wireless-only and integrated systems, respectively. These results prove that the AMEF can provide a robust and adaptive solution for achieving high-speed and multi-user communication in next-generation radio access networks.

A millimeter-wave (MMW) integrated sensing and communication (ISAC) system based on photonics is proposed. Its structure is simple, and the optical fiber-wireless structure can perfectly adapt to the current centralized fiber-distributed communication network. In the proposed system, the linear frequency modulated (LFM) signal and subcarrier modulated (SCM) 16-quadrature amplitude modulation (QAM) signal are combined by frequency division multiplexing (FDM). In the experiment, we verify a photonic-based ISAC W-band fiber-wireless integration system at 96.5 GHz over 10-km fiber transmission. We achieve 48 Gbit/s in communication-only mode and 8 Gbit/s in ISAC mode when the sensing resolution is 2.75 cm. By means of external calibration, the distance error of the system is less than 1 cm.

A flexible coherent passive optical network (FLCS-CPON) is a promising solution for the future access network. By offering increased speed, sensitivity, and flexibility, it enables more efficient utilization of network resources and allows for serving a larger number of users. However, the past studies often overlook the flexibility of channel equalization. In the FLCS-CPON, customized rate optimization has been achieved to cater to users in different channel conditions. However, in addition to rate optimization, further performance improvement can be achieved by providing customized equalization methods. In this work, we proposed and demonstrated an end-to-end (E2E)-optimization-based FLCS-CPON in a 50 km fiber transmission. It enables tailored signal constellation shaping and equalization, maximizing system efficiency and performance. Finally, we achieved a FLCS-CPON with the net data rate (NDR) varied from 124 to 210 Gbps and the power budget of 40 and 42.4 dB in upstream and downstream, respectively; 3.7 and 2.9 dB power budgets are improved by E2E optimization. In burst-mode, the dynamic range of probabilistic shaping 32 quadrature amplitude modulation (PS-32QAM) at a line rate of 250 Gbps improved by 6.1 to 16.8 dB. Additionally, a dynamic range and net-rate product (DRNRP) of 5779 dB · Gbps is achieved.

In fiber-millimeter-wave integration systems, nonlinear distortion and inter-symbol interference (ISI) will degrade system performance, and post-equalization is an efficient method to compensate for them. In this paper, a 60 Gbit/s 64-quadrature amplitude modulation (QAM) fiber-millimeter-wave integration system based on multi-band subcarrier modulation (MSCM) at W-band is demonstrated. We design a comparative experiment to study equalization method for multi-band SCM nonlinear distortion compensation. Volterra-series based equalizer, artificial-neural-network (ANN)-based equalizer and Bi-directional gated recurrent unit (Bi-GRU)-based equalizer are compared for their ability to compensate for nonlinear distortion. Experiment results indicate that it is necessary to compensate nonlinear distortion in waveform level and Bi-GRU equalizer can compensate the nonlinear distortion the best. With post-equalization, a 60 Gbps data rate is achieved at a forward error correction (FEC) bit error rate (BER) threshold of 1×10 -2 after 20-km fiber and 1-m wireless transmission.

Fiber-terahertz integrated communication system has emerged as a promising technology for 6G. In this paper, an end-to-end learning-based quadrature amplitude modulation (QAM) symbol-to-symbol autoencoder frame work is proposed and demonstrated to solve problems in fiber-terahertz integrated communication system. The bit error rate (BER) performance is studied at different peak-to-peak voltage (Vpp) and received optical power (ROP). We compare this method with a baseline using sub-carrier modulation (SCM) signal based on Volterra nonlinear compensation and Least Mean Square (LMS) linear equalization. Through appropriate channel modeling and end-to-end optimization, a sensitivity gain of 1.3 dB is achieved at a data rate of 40 Gbps and a forward error correction (FEC) BER threshold of $1\times 10^{-2}$ after 10-km fiber and 1-m wireless transmission.

Fiber-wireless integration has been widely studied as a key technology to support radio access networks in sixth-generation wireless communication, empowered by artificial intelligence. In this study, we propose and demonstrate a deep-learning-based end-to-end (E2E) multi-user communication framework for a fiber-mmWave (MMW) integrated system, where artificial neural networks (ANN) are trained and optimized as transmitters, ANN-based channel models (ACM), and receivers. By connecting the computation graphs of multiple transmitters and receivers, we jointly optimize the transmission of multiple users in the E2E framework to support multi-user access in one fiber-MMW channel. To ensure that the framework matches the fiber-MMW channel, we employ a two-step transfer learning technique to train the ACM. In a 46.2 Gbit/s 10-km fiber-MMW transmission experiment, compared with the single-carrier QAM, the E2E framework achieves over 3.5 dB receiver sensitivity gain in the single-user case and 1.5 dB gain in the three-user case under the 7% hard-decision forward error correction threshold.

We propose and demonstrate a hybrid communication architecture that combines millimeter-wave (MMW) in the radio frequency (RF) domain and free-space-optics (FSO) technologies using adaptive combining and hybrid automatic repeat request (HARQ) techniques. At the receiving end, we employed joint signal processing with an adaptive diversity combining technique (ADCT) based on a maximum ratio combining (MRC) algorithm. We derived closed-form expressions for the outage probability and throughput of the hybrid RF and FSO (RF/FSO) system, considering various characteristics of atmospheric turbulence in the FSO link. Experimental testing with 10-Gbaud quadrature phase shift keying (QPSK) data was conducted under different simulated atmospheric turbulence intensities, FSO and MMW speed-ratios, and forward error correction (FEC) overheads. Additionally, we validated improvements in terms of bit error ratio (BER), outage probability, and throughput performance.

We proposed and experimentally demonstrated a bit-wise end-to -end deep -learningbased autoencoder for a fiber-THz integrated DFT-S-OFDM communication system at 209 GHz. More than 5.5 -dB sensitivity gain is achieved compared with traditional DFT-S-OFDM system at 50Gbps.

We proposed and experimentally demonstrated a novel two-dimensional end-to-end deep-learning-based autoencoder for fiber-THz integrated 6G radio-access-network at 220 GHz. More than 1.8-dB sensitivity gain is achieved compared with traditional QAM modulation formats at 40Gbps.