Receiver's Multi-Antenna Coupling Cancellation Based on Guided Multi-level Search

Abstract

In order to achieve higher bit rates and gain diversity, trends in mobile communication industry include the use of recent trends in communication industry include the use of multiple antennas in transmitter and at the receiver [3,4,5]. multiple antennas in transmitter and at the receiver. At the At the same time, the mobile terminals minimization is being same time the terminals are getting smaller incrementing the emphasized leaving less space to increased number of level of electronic interferences in their interior. In this electronic elements. The closeness of data metallic wires and contribution, the signal distortion known as RF front-end antenna coupling is analyzed. The coupling between antennas is m anylecroice c ts leatoiferen distortincs ta n modeled with polynomials limited to third order nonlinear significantly reduce the entire system performances [6].behavior. Nonlinear software decoupling module, located at the Nevertheless these imperfections are still not analyzed output of analog-to-digital (AD) converter, is proposed for thoroughly by scientific research community. coupling cancellation. Inside the module, signal enhancement is In this contribution a proposal how to model and overcome achieved with the approximation of inverse nonlinear coupling the multi antenna RF front end coupling distortions is function with the polynomial of gth degree. The inverse function presented and analyzed. The coupling is modeled with is obtained using guided multilevel search approximation polynomials limited to third order nonlinear behavior.method. The proposed coupling cancellation algorithm is

RF- Front-End Non-Linear coupling cancellation = Cancelación de acoplo no lineal en frontales de RF

“…Since decoupling module used in practice is based on IFSA or GML approach, the calibration size can be selected without large pressure of processing load requirements. Taking into account that the generalized system, like ours, should be designed for the worst case scenario, the overall calibration size is set to N = 512 as in [Ara07c]. This calibration size is chosen in order to preserve system precision and robust behaviour demonstrated in previous simulations throughout the thesis based on 2000 calibration samples.…”