Passive Inter-modulation Cancellation in FDD System

Abstract: Passive inter-modulation (PIM) shows up when two or more RF signals are mixed together via the nonlinear passive devices. With the rapid development and expansion of the wireless com-munication networks, PIM is recognized as one of the most critical issues. Especially in the base station, where several networks systems have to share the same infrastruc-ture, a few transmitted signals are physically close to each other, even they could be sharing the same antenna. So it is highly possible that several transmitted signals hit a same nonlinear junc-tion and generate PIM distortion, then the PIM distortion falls in some bands of the receivers nearby. Even more critically, since some technologies are developed with multi-carrier or multi-band, such as Carrier Aggregation and MIMO, the generated PIM could spread in a wide range of the spectrum and overlap with the receiver bands. Additionally the continuous increased power of the transmitted signals and the hyper sensitivity of the received signals make PIM issue particularly critical in the base station. Even the fifth and seventh order products of PIM could cause the performance degradation on the received signals. Isolation and shielding have been used to suppress the effect of PIM distortion, but it is costly and space-consuming solution. In this thesis, a digital self-adaptive cancellation technique is in-troduced. In the proposed technique, an estimated PIM is created based on the known baseband transmitted signal which is highly correlated to the PIM distortion present in the receiver chain, then the received signal is extracted by subtracting the estimated PIM signal from the distorted received signal in the digital domain. Time-dependency is one of the natures of PIM, since the nonlinearity of the passive devices al-ways varies with temperature, humidity and elements age. To realize an accurate estimation of PIM distortion in real time, the adaptive filter with NLMS (Normalized Least Mean Square) and RLS (Recursive Least Squares) algorithms is used to adjust the coefficients of the filter in the cancellation block. To ensure the expected performance of the technique to be reached, many simulations under dif-ferent situations are implemented and analyzed in Matlab. Additionally some proper simplifica-tions in the cancellation block are proceeded so that this technique can be realized in the hard-ware at a low cost in future.