6G OAM-THz Channel Dataset: ITU-R IMT-2030 Compliant Physics Simulation with 33 Parameters for Deep Reinforcement Learning

The 6G OAM THz Dataset is the first comprehensive, physics-based library dedicated to Orbital Angular Momentum (OAM) beam communications at sub-Terahertz frequencies (300-600 GHz) for 6G wireless systems. This dataset includes 270,000 high-fidelity samples with perfect validation quality, as well as 33 comprehensive physics parameters covering eight critical domains: atmospheric turbulence, weather environment, fading and channel characteristics, pointing and alignment, OAM beam physics, hardware impairments, propagation physics, and performance metrics.The dataset, created using rigorous physics-based simulatio ns, covers four deployment scenarios: lab controlled (70K samples), indoor realistic (80K samples), outdoor urban (70K samples), and high mobility (50K samples), providing comprehensive coverage from controlled environments to wide-area applications. Each sample includes realistic atmospheric modeling, hardware impairments, and ambient circumstances, resulting in throughput ranges of 0.1-1000 Gbps, latency performance of 0.011-2.841 ms, and SINR coverage ranging from -30 to +50 dB.This dataset is completely consistent with DOCOMO 6G specifications and ITU-R IMT-2030 frameworks, allowing for sophisticated Deep Reinforcement Learning applications in next-generation wireless communications. Multi-agent network optimization, dynamic beam steering control, adaptive power allocation, and interference mitigation procedures are some of the primary applications. The repository includes comprehensive documentation, visualization tools, and code samples to facilitate reproducible research in 6G system design, THz propagation modeling, and OAM-based wireless communications.Keywords: 6G wireless communications, Terahertz frequency, Orbital Angular Momentum, Deep Reinforcement Learning, Physics-based simulation, DOCOMO compliance, ITU-R IMT-2030, Channel modeling, Atmospheric turbulence, Wireless dataset