Communications using ubiquitous antennas: Free-space propagation

Abstract

The inefficiency of the cellular-network architecture has prevented the promising theoretic gains of communication technologies such as network MIMO, massive MIMO and distributed antennas from fully materializing in practice. The revolutionary cell-less cloud radio access networks (C-RANs) are under active development to overcome the drawbacks of cellular networks. In C-RANs, centralized cloud signal processing and minimum onsite hardware make it possible to deploy ubiquitous distributed antennas and coordinate them to form a gigantic array, called the ubiquitous array (UA). This paper focuses on designing techniques for UA communications and characterizing their performance. To this end, the UA is modeled as a continuous circular array enclosing target mobiles and free-space propagation is assumed, which allows the use of mathematical tools including Fourier series and Bessel functions in the analysis. First, exploiting the UA's large circular structure, a novel scheme for multiuser channel estimation is proposed to support noiseless channel estimation using only single pilot symbols. Channel estimation errors due to interference are proved to be Bessel functions of inter-user distances normalized by the wavelength. Besides increasing the distances, it is shown that the errors can be also suppressed by using pilot sequences and eliminated if the sequence length is longer than the number of mobiles. Next, for data communication, we first consider channel conjugate transmission that compensates the phase shift in propagation and thereby allows receive coherent combining. The multiuser interference powers are derived as Bessel functions of normalized inter-user distances. Last, we propose the design of multiuser precoders in the form of Fourier series whose coefficients excite different phase modes of the UA. Under the zero-forcing constraints, the precoder coefficients are proved to lie in the null space of a derived matrix with elements being Bessel functions of normalized inter-user distances.