SmartIA: Context-Aware Feedback-Free Beam Discovery for Efficient Initial Access

Future wireless networks leveraging millimeter-wave and sub-Terahertz spectra rely on highly directional beamforming to counter severe path loss. However, identifying the optimal beam, specially during the Initial Access (IA) procedure, remains challenging due to the lack of feedback between the Base Station (BS) and User Equipment (UE). In this paper, we present SmartIA a context-aware feedback-free hierarchical beam scanning framework for efficient IA. We design an optimized sequence of transmitted beam patterns that uniquely encode spatial directions, enabling all UEs to identify the best beam by aggregating multi-stage information. This achieves logarithmic-complexity beam discovery with reduced acquisition time while preserving high spatial resolution. We further introduce a novel analog antenna weight vector design that achieves the required spatial coding with full power amplifier utilization. We develop an adaptive Synchronization Signal Block (SSB) transmission strategy that leverages contextual information on link conditions across the user distribution to minimize BS radio resource utilization, while preserving cell coverage under power-constrained multi-directional transmissions. Simulation results show 50% lower IA acquisition time and BS energy consumption, with 3 dB gains in both coverage and link quality compared to standard approaches, validating SmartIA as an scalable solution for beam discovery and IA in future high-frequency networks.