Performance Analysis of User-Centric Interference Cooperation in Multi-Antenna UAV Networks

Multi-antenna unmanned aerial vehicle (UAV) networks can improve throughput by spatial multiplexing and UAVs dominant line-of-sight (LOS) connections, which will intensify interference and exacerbate user fluctuation. This paper proposes user-centric interference cooperation in multi-antenna UAV networks, where multi-antenna beamforming and joint transmission are performed among cooperative UAVs, and derives a semi-closed expression of the ergodic rate using stochastic geometry. Specifically, a dynamic UAV group with adjustable radius <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">${R}$</tex-math></inline-formula> is organized for each user that forms a 3D conical region to guarantee seamless service, which provides spatial multiplexing freedom via cooperative multi-antenna beamforming to achieve interference mitigation. In addition, both signal power and inter-cluster interference power are approximated by Gamma distributions with appropriate parameters experiencing generic Nakagami-<inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$m$</tex-math></inline-formula> fading in air-to-ground (AtG) channel. Furthermore, zero-forcing (ZF) beamforming is adopted with equal power allocation among beams, which is across the distributed set of UAVs within the cluster and considering different antenna loading. Numerical results show that the average ergodic rate can achieve 86.01% gain compared to the uncoordinated single-cell multi-antenna UAV network when UAV density is <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">${100\mathrm{/km^{2}}}$</tex-math></inline-formula> and UAV group radius <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">${R}$</tex-math></inline-formula> is <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">${20\mathrm{m}}$</tex-math></inline-formula>. The optimal network parameters are investigated to maximize ergodic rate, thus providing guidelines for practical network configuration. For instance, when UAV density is <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$50\mathrm{/km^{2}}$</tex-math></inline-formula> and loading factor is <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$K/M=2/16$</tex-math></inline-formula>, the optimal deployment altitude for UAVs is <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$ h=20m$</tex-math></inline-formula>, which can achieve a rate gain of 144% compared to the case of <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$h=0\mathrm{m}$</tex-math></inline-formula>.