In this work, we implemented a short-reach real-time optical communication system using MLP for pre-distortion. Lookup table (LUT) algorithms are commonly employed for pre-distortion in intensity modulation and direct detection (IM/DD) systems. However, storage limitations typically restrict the LUT pattern length to 9, limiting its effectiveness in compensating for nonlinear effects. A multilayer perceptron (MLP) can overcome this limitation by predicting errors and generating pre-distorted signals, thus replacing the extensive storage requirements of LUTs with minimal computational resources. The MLP-based digital pre-distortion (MLP-DPD) technique enables the creation of long-pattern LUTs for improved nonlinear compensation. In this work, an MLP-DPD scheme was implemented on a field-programmable gate array (FPGA). The FPGA was used to generate a 14.7456 GBaud pre-distorted pulse amplitude modulation 4-level (PAM4) signal. This signal was then transmitted over 20 km of standard single-mode fiber (SSMF). At the receiver, the parallel constant modulus algorithm (PCMA) was applied for signal processing. The bit error rate (BER) achieved met the 2.4 × 10⁻² threshold for soft-decision forward error correction (SD-FEC), enabling a net transmission bit rate of 24.576 Gbit/s. This approach demonstrates the feasibility of using MLP-DPD for effective nonlinear compensation in high-speed optical communication systems with limited resources.

With the rise of 6G, the exponential growth of data traffic, the proliferation of emerging applications, and the ubiquity of smart devices, the demand for spectral resources is unprecedented. Terahertz communication (100 GHz–3 THz) plays a key role in alleviating spectrum scarcity through ultra-broadband transmission. In this study, terahertz optical carrier-based systems are employed, where fiber-optic components are used to generate the optical signals, and the signal is transmitted via direct detection in the receiver side, without relying on fiber-optic transmission. In these systems, deep learning-based equalization effectively compensates for nonlinear distortions, while probability shaping (PS) enhances system capacity under modulation constraints. However, the probability distribution of signals processed by PS varies with amplitude, making it challenging to extract useful information from the minority class, which in turn limits the effectiveness of nonlinear equalization. Furthermore, in IM-DD systems, optical multipath interference (MPI) noise introduces signal-dependent amplitude jitter after direct detection, degrading system performance. To address these challenges, we propose a lightweight neural network equalizer assisted by the Synthetic Minority Oversampling Technique (SMOTE) and a clustering method. Applying SMOTE prior to the equalizer mitigates training difficulties arising from class imbalance, while the low-complexity clustering algorithm after the equalizer identifies edge jitter levels for decision-making. This joint approach compensates for both nonlinear distortion and jitter-related decision errors. Based on this algorithm, we conducted a 3.75 Gbaud W-band PAM4 wireless transmission experiment over 300 m at Fudan University’s Handan campus, achieving a bit error rate of 1.32 × 10⁻³, which corresponds to a 70.7% improvement over conventional schemes. Compared to traditional equalizers, the proposed new equalizer reduces algorithm complexity by 70.6% and training sequence length by 33%, while achieving the same performance. These advantages highlight its significant potential for future optical carrier-based wireless communication systems.