Authors: Prateek Anand, Assistant Professor Rishi Sharma, Assistant Professor Gaurav Morghare
Abstract: Low Earth Orbit (LEO) satellite communication has emerged as a key enabler of sixth-generation (6G) Non-Terrestrial Networks (NTNs), offering global coverage, low propagation delay, and high-capacity broadband connectivity. However, the high orbital velocity of LEO satellites introduces significant challenges, including severe Doppler frequency shifts, rapidly varying channel conditions, and frequent handovers, which adversely affect communication reliability, throughput, and Quality of Service (QoS). Existing Doppler compensation and handover mechanisms are generally treated as independent processes and often rely on static threshold-based strategies or computationally intensive artificial intelligence (AI) models, limiting their adaptability in highly dynamic satellite communication environments. This paper proposes an Adaptive AI-Assisted Doppler Compensation and Predictive Handover Optimization (AIDCPHO) framework for LEO satellite communication in 6G Non-Terrestrial Networks. The proposed framework integrates real-time Doppler estimation, AI-assisted predictive handover decision-making, adaptive beam selection, and dynamic link quality assessment into a unified optimization model. A predictive mobility module estimates future satellite-user link conditions using orbital dynamics and user mobility information, while an adaptive Doppler compensation module minimizes frequency estimation errors before communication degradation occurs. Furthermore, a multi-parameter handover decision algorithm utilizes Signal-to-Noise Ratio (SNR), Doppler shift, elevation angle, received signal strength, and predicted link quality to proactively initiate seamless handovers, thereby reducing service interruption and packet loss. The proposed framework is implemented and evaluated using MATLAB-based simulations that model realistic LEO satellite orbital movement, time-varying communication channels, and user mobility scenarios.
