Estimation of radio-over-fiber uplink in a multiuser CDMA environment using PN spreading codes

Abstract

Two of the major issues with radio-over-fiber (ROF) technology for wireless communications are the nonlinear distortion of the optical link and the multipath dispersion of the wireless channel. In order to limit the effect of these distortions, estimation, and subsequently equalization of the concatenated fiber-wireless channel should be done. The estimation of the fiber-wireless uplink was done in a single user CDMA environment in [1]. This paper focuses on estimation in a multiuser CDMA environment by utilizing the properties of pseudonoise (PN) spreading codes. We also consider both wireless channel and optical receiver noise. Furthermore, we propose a new iterative technique to mitigate the adverse effect of multiple access interference (MAI) on the identification process. Numerical evaluations of the developed algorithm show a good estimation of both the linear and nonlinear systems with 10 users, an SNR of 25 dB between the mobile user and RAP, and an optical noise power equalling the wireless noise power.

Tensor based receivers for nonlinear Radio Over Fiber uplinks in multiuser CDMA systems

“…A MIMO Wiener model of an uplink channel of a ROF multiuser communication system [1], [5] is considered for the simulations. The flat fading wireless link is characterized by a memoryless R x 2 linear mixer, with T == 2 users (Q == 4) and a half-wavelength spaced array of R antennas.…”

“…The E/O conversion for each channel is modeled by the following polynomial CIX + c3lxl 2x, with Cl == -0.35 and C3 == 1 [1], [20]. Table II, where a short training sequence is used to obtain an initial estimate of the channel A (0) -Q matrix H .…”

“…Substituting (8) (1) due to the fact that the term IUti, n,p 1 2 reduces to a multiplicative constant that can be absorbed by the associated channel coefficient. As a consequence, some nonlinear terms degenerate in terms of smaller order, leading to a smaller value of Q.…”

“…be expressed as a MIMO Wiener filter [1], [2], [3]. That and means that the discrete-time baseband received signals after E/O conversion can be expressed as the output of a MIMO Volterra system: is the spread signal transmitted by user t and associated with chip p of time instant n, T is the number of users and vr,n,p is the Additive White Gaussian Noise (AWGN).…”

“…In such systems, the uplink transmission is done from a mobile station towards a Radio Access Point (RAP), the transmitted signals being converted in optical frequencies by a laser diode and then retransmitted through optical fibers. Important nonlinear distortions are introduced by the electrical-optical (E/O) conversion [1], [2], [3]. When the length of the optical fiber is short (few kilometers) and the radio frequency has an order of GHz, the dispersion of the fiber is negligible [4].…”