We propose a directional pinching-antenna system (DiPASS), a comprehensive framework that transitions PASS modeling from idealized abstraction to physical consistency. DiPASS introduces the first channel model that accurately captures the directional, pencil-like radiation of pinching antennas, incorporates a practical waveguide attenuation of 1.3 dB/m, and accounts for stochastic line-of-sight blockage. A key enabler of DiPASS is our new "equal quota division" power allocation strategy, which guarantees predetermined coupling lengths independent of antenna positions, thereby overcoming a critical barrier to practical deployment. Our analysis yields foundational insights: we derive closed-form solutions for optimal antenna placement and orientation in single-PA scenarios, quantifying the core trade-off between waveguide and free-space losses. For multi-PA systems, we develop a scalable optimization framework that leverages directional sparsity, revealing that waveguide diversity surpasses antenna density in enhancing system capacity. Extensive simulations validate our analysis and demonstrate that DiPASS provides a realistic performance benchmark, fundamentally reshaping the understanding and design principles for future PASS-enabled 6G networks.

翻译：我们提出了一种定向夹持天线系统（DiPASS），这是一个将PASS建模从理想化抽象过渡到物理一致性的综合框架。DiPASS引入了首个准确捕捉夹持天线定向、铅笔状辐射的信道模型，纳入了1.3 dB/m的实用波导衰减，并考虑了随机视距阻塞。DiPASS的关键推动力是我们新的“等额分配”功率分配策略，该策略保证了与天线位置无关的预定耦合长度，从而克服了实际部署中的关键障碍。我们的分析得出了基础性见解：在单PA场景中，我们推导了最优天线放置和朝向的闭式解，量化了波导损耗与自由空间损耗之间的核心权衡。对于多PA系统，我们开发了一个利用定向稀疏性的可扩展优化框架，揭示了波导分集在提升系统容量方面优于天线密度。大量仿真验证了我们的分析，并证明DiPASS提供了现实的性能基准，从根本上重塑了对未来支持PASS的6G网络的理解和设计原则。