Abstract
In this thesis, the multiband (MB)-orthogonal frequency division multiplexing (OFDM) ultra-wide band (UWB) system is investigated. The symbol of each user is spread by a spreading code to all subcarriers. Thus, the system model employed is equivalent to that of a multicarrier (MC)-code division multiple access (CDMA). Many researchers have paid attention to the problem of narrowband interference (NBI) cancellation on UWB system. Several techniques are employed at the receiver to do NBI suppression or cancellation. The NBI suppression technique is robust mini-max M-estimator. The NBI cancellation techniques are consecutive mean excision (CME), forward consecutive mean excision (FCME), least median of square (LMedS) and reweighted least square (RLS). We improve the methods of CME, FCME, LMedS and RLS to from easyCME, easyLMedS, easyRLS, and Maximum subcarrier power (MSP). We also compare the complexities of the estimators mentioned above. The outlier detection probability and false alarm probability of easyLMedS and easyCME are analyzed according to the noise model. Computer simulations are run to verify the analytical. We can also use outlier detection methods to detect if NBI exists. From computer simulation we can find the appropriate parameters of CME and FCME and prove that the performance of easyCME is the upper bound of those of CME and FCME. From the bit error rate of computer simulations, the following conclusions are made. First, easyCME and MSP are better than other estimators. Secondly, EasyLMedS is batter than LMedS. Thirdly, easyCME is better than CME and FCME. Those estimators on BER curves sequenced in the descending order are LMedS, FCME, CME, easyLMedS, easyCME and MSP. When NBI power is not too large, robust mini-max M-estimatorstill can suppression NBI. But when NBI power is too large, robust mini-max M-estimator can not control NBI well.